Software Review
iListen 1.7
Developer: MacSpeech
Price: $179 with headset; $40 (upgrade); Educational discount available.
Requirements: Mac OS X 10.3.9. Mac with G4 or faster. Universal.
Recommended: Mac OS X 10.4, 512 MB of RAM.
Trial: None
I have been interested in voice recognition software for a while but never played with it. When the chance to explore iListen’s new upgrade came up, I jumped at it. I hit some snags, but I’m pleased to say the product itself is good.
There are two pieces to the iListen setup experience: one is the usual software download or CD loading of the software. No problems there. The other part of the game is making sure you have an approved microphone. I did not anticipate this issue, and it delayed my using the software for several days. If you purchase the software directly from MacSpeech, they include an approved microphone. Since my copy was not purchased this way, I had to scramble for a microphone. Do not assume the microphone or headset you already have will work—MacSpeech has only approved a limited number of units that will work well. Their instructions say that it is possible other microphones will work, but they will not promise excellent results, and they will not provide support.
Once you have the software installed and the proper microphone plugged in, you read a story. Yep. You read a story to iListen, so it can hear what you sound like and how you pronounce words. You will go through the document and make corrections, and read another story or two. Or half a dozen. It depends on how good your accuracy rate is. The FAQ page suggests that once that rate is in the 92% neighborhood, you won’t improve things much by reading any more stories.
My first story came in at 88%. I read two training panels and was using an unapproved microphone—but it is a relatively high-end model, the Plantronics GameCom Pro 1. After one more training panel, my accuracy rate was 90%. I tried to shush the cat during training, but then decided he would be commenting other times, so I might as well let iListen include him in its analysis.
I enjoyed some of iListen’s attempts at understanding the test story I was reading, about new North Carolina laws banning cell phone use by young drivers, and requiring rear-seat passengers to wear seat belts. My favorite error was when I read “Also becoming law Dec. 1 is a requirement for all back seat passengers riding in North Carolina to buckle up.” The software said I was requiring folks to “Bob aloft,” which I thought would be more fun, if you had a hot-air balloon.
I followed the suggestion on the FAQ page, and preset the microphone in the Sound pane of System Preferences. I did not look at the screen while dictating, as that messes with your head on the delay. I used SimpleText, but I would have turned off Spelling & Grammar in Microsoft Word. I turn it off when I use it anyway, and I don’t use Word by choice.
The software works by translating your spoken words into text. It starts with a profile based on your gender and the language you speak. I suppose men with high voices and women with low voices might need more tweaking, if the software expects certain inflections. My voice is low for a woman, so that may have influenced some of my misunderstandings. The requirements do say the user should be at least age 14, so possibly the software is not as good at dealing with children’s voices as it will be someday.
After my fourth training story, my accuracy rate on the driving laws story did not improve beyond 90%. Most of us have to read behind ourselves when we write, anyway. I would prefer not to have to make a correction in every paragraph, but until the software learns my voice better, I guess that’s how it works. It is still miles ahead of where we were a few years ago. Half a dozen training stories at a pop is about all I recommend, as your voice tires. Some of the training stories are a good bit longer, so pay attention to your throat. Some of the errors between my third and fourth training stories changed, making me think I had gone too long for the software to understand me.
The microphone is to be positioned very near your mouth, so it is very sensitive to your voice but not to other sounds. I experimented with having someone talk as he walked behind me while I was dictating. The microphone did pick up his voice when he was within two feet, but not much beyond that. iListen only listens to me! I may set up a profile for the other person, though. He set type for many years but his typing is iffy. Maybe it was because he worked upside down and backwards all those years. He might like to speak aloud his e-mails.
I like this program very much. It is fun to see that we are living in the world our science fiction writers dreamed for us decades ago. I will use this product on my next research paper for graduate school. I have already begun it. The course is Philosophy of Mind; I am eager to see if using this tool changes the way I write or think.
If you spend a lot of time writing, you should consider this product. It is a smidge pricy for my taste, but we get what we pay for, and there is a reason some stuff costs a lot. I could see it being helpful for someone writing a sermon or speech, who thinks best aloud rather than in print. Using this software would allow such a thinker to speak an idea rather than trying to construct it in letters. If you are a walk-around thinker, though, you will need a long cord, as iListen does not yet support any Bluetooth microphones.
I am not sure this software would be a great idea in an office or other environment with several people talking. You would need to test it with an approved microphone to know whether the background voices would be a problem.
The folks this software would be best for, though, are those who have typing issues. I have a wonderful friend with occasional finger problems due to a medical condition, and if he does not already have software like this, I will encourage him to explore it.
Copyright © 2006 Ellyn Ritterskamp, eritterskamp@atpm.com. Reviewing
in ATPM is open to anyone. If you’re interested, write to us at
reviews@atpm.com.
Reader Comments (471)
I am using the latest version and I must admit that after maybe 6-7 hours of reading to my Mac I am thinking that someday I will be able to use it productively.
Note: The (unmentionable competing Mac application) from I.B.M. has been orphaned as far as I can see.
This looks like it has plenty of potential.
It's accuracy is worse than ViaVoice [which works well in all OS Xs, but 10.4.7]; and it's worst failing is its inability to place corrections in the proper places. I wind up with orphaned text placed at random in paragraphs.
My advice is to get an Intel Mac [with appropriate software: Bootstrap/Parallels/Crossover] and use the PC version of Dragon Dictate - Naturally Speaking. It's very accurate!!!
Even Dragon NaturallySpeaking fanatic and Mac pundit David Pogue was able to achieve almost 96% accuracy after using iListen for only 30 minutes. If you aren't achieving decent accuracy, our support team is here to help. Please contact them at "[email protected]" for assistance.
To be useful, voice recognition software has to achieve way better than 90% accuracy. I am a registered user of both ViaVoice and iListen, but abandoned both years ago. I would have been interested to know how much better iListen Version 1.7 is than previous versions - if indeed it is.
We need a voice app for the Mac that is as good as Dragon Naturally Speaking allegedly is for the PC (ref. Pogue D.)
I cannot compare it to ViaVoice or Dragon, because I have never used those products. Possibly my rating was inflated because I am thrilled this kind of software even exists. But this program scored a lot of ease-of-use points with me, which is what most of us expect from Mac apps. My rating was as much based on potential as anything.
Thanks for reading and commenting.
Even David Pogue said he got almost 96% accuracy after using iListen for only 30 minutes - he said that in the same review Steve quoted above.
Accuracy levels in iListen can be achieved that are every bit as high as Dragon NaturallySpeaking and we have many users who have stated they have achieved this (we even have one who states he got better recognition out of iListen after 2 weeks than he did from NaturallySpeaking after 5 years!).
All that having been said, please keep in mind that Dragon has been around a lot longer than we have as a company, and they have not been faced with the mandate of rewriting their program completely 3 times just to keep up with changes to the platform. (We had to rewrite to go from 68K to PPC, Mac OS 9 to Mac OS X, and now PPC to Intel.)
It is our hope that Apple will leave a lot of the under the hood stuff alone for a while so we can further improve our company by much larger leaps, rather than the baby steps we have been forced to take up until now just so we can maintain compatibility with the platform.
Best Regards,
Chuck Rogers, Chief Evagelist
MacSpeech, Inc.
You sir are a nice guy and quite affabe. We spoke years ago b4 you left and returned.
First: ViaVoice is more accurate.
Second: Even Power Secretary was more accurate.
Third: Your support team s*cks.
I got "0" help !
They could not even understand my problem.
Let's say iListen did not recognize "problem? [above]
I use correction and it places the correction between "support...team" or any where else it pleases.
I can type faster. And I type poorly...
It does not help your cause when your software misspells: Evangelist!
Nice try.
Terrie
I use iListen primarily for Command and Control, although more recently I have been using it more for dictation because I seem to be developing some RSI issues with my right arm. My response (and misspelling) was typed, not dictated.
If you haven't tried our support lately, which seems to be indicated by your comment, you will be thoroughly surprised. I was put in charge of it when I came back and I will GUARANTEE you we are far better in every aspect at this point than ViaVoice ever was.
Come check us out. If you purchase the upgrade and our support team a). doesn't respond to you within 24 hours; and b). can't get you above 90% accuracy within 1 week and above 95% within 1 month or less, you can have your money back.
One last thing: our growing customer base - including many former ViaVoice AND Dragon NaturallySpeaking customers would certainly disagree with you that our software sucks - it may have at one point (and those who know me know that's why I left). I would not have returned if the software continued to suck. It is that simple.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I will buy a VXI TalkPro Xpress headset [w/Andrea USB pod],
as the combo has the highest recommendation,
and try the upgrade.
You still sound reasonable.
[You did the right thing to leave --
I admire your principles.]
Some people report the output level as being too high when used with the Andrea USB pod on a PC. However, the Andrea USB pod is recommended for use with Mac computers. [Macs like the higher Andrea gain.]
Terrie
I can state that over the past few weeks, the Support staff at MacSpeech has been tremendously helpful and responsive, even to the extent of responding to an emailed support request I made just late this past holiday Sunday afternoon in less than 2 hours, which was most unexpected, and which solved a huge problem for me.
I further find Chuck Rogers (whom I do not know and have never had the pleasure of meeting) to be a breath of fresh air in the business world. I am able to see and notice the palpable, impressive progress the company has already made in their product between version 1.6 and 1.7 - bearing in mind all the Platform/rewriting work and changes that have been required by Apple just to keep pace. Their obvious enthusiasm and determination give me sufficient confidence that they will continue to advance and improve their product and to continue real R&D work on iListen such that I chose to pick iListen as my solution rather than attempting to cobble Dragon through Boot Camp/Parallels, etc. Of course, each person must come to their own decision in that regard.
Just FYI, I have been using the Program on a Dual 2.0Ghz G5 Power Mac with 4.5 GB of ram and, with finishing merely the initial training, was getting around 94% accuracy. However, I am just re-loading the program onto a newly obtained Mac Pro 3.0Ghz, and am hopeful that the extra power, combined with some extra patience and care in training, will pay off in even higher accuracy for the Universal program.
Best regards,
Norm
See the article, Like Having a Secretary in Your PC, July 20, 2006, at:
http://www.nytimes.com/2006/07/20/technology/20pogue.html?ex=1157601600&en=42ba211266db6c46&ei=5070
Relevant excerpts:
"And now they've topped that: NatSpeak 9 requires no training at all.
I gave it a test. After a fresh installation of the software, I opened a random page in a book and read a 1,000-word passage -- without doing any training.
The software got 11 words wrong, which means it got 98.9 percent of the passage correct. Some of those errors were forgivable, like when it heard "typology" instead of "topology."
But Nuance says that you'll get even better accuracy if you do read one of the training scripts, so I tried that, too. I trained the software by reading its "Alice in Wonderland" excerpt. This time, when I read the same 1,000 words from my book, only six errors popped up. That's 99.4 percent correct."
There is no question that you can be more accurate with Dragon NaturallySpeaking more quickly than you can with iListen. We have never suggested otherwise.
But if David Pogue was getting 96% accuracy after only 30 minutes, imagine how accurate he would be after using it for only a week or so. At that point, does it really matter if Dragon is 99.6 percent accurate and iListen is 99.2 percent accurate - a difference of only 4 words per 1000?
The point is that yes - with Dragon you will be more accurate sooner. Over time, however, your accuracy should be about the same.
I would like to correct this statement as it DOES work fine and dictates very well into other applications (Mail, TextEdit, etc.). Accuaracy is quite good...over 90% I would say.
It may be orphaned but it works.
Kim Peacock
iListen works on PPC and Intel Macs. 1.7 works on any G4, G5 or Intel Mac running 10.3.9 or later - and our customers are consistently getting well above 90% accuracy. Most are well above 95%, and, over time, this can be expected to increase to above 99%.
Another advantage of NaturallySpeaking is its multilanguage ability. I work in English, French and Dutch and would need software that accepts dictation in all these languages. Any chance that iListen develops that ability?
For French and Dutch, you will need to stick with Dragon.Wedo have Spanish, German, and Italian versions of iListen, however.
Thanks,
Jim B.
We have many medical professional using iListen for dictation. One of them allowed us to share his comments with others:
"I'm getting 95-99% recognition after only a week using the system for all of my patient notes and correspondence. It is still learning the names of medications, but cycling each day's clinic note through the "Learn My Writing Style" feature has proven very useful. So far, and again after only one week of office use, iListen is saving me 20-30 minutes a day over my use of Dragon (which I have used over the past 5 years). That's a whole lot of heart beats to be spent doing something else besides talking to a computer. It's almost like having a little person typing away inside of my Mac Mini. Thanks for your dedication to this superb product."
-David J. Black MD
-Private Practice of Family Medicine
-Gainesville
You can find more customer testimonials here.
Why is then Dragon Naturally Speaking for PC considered a 'better' speech recognition application than iListen 1.7 for Mac?
Can someone explain simply so that I can show my K12 students? I will acknowledge your explanation.
Is it a PC - Mac thing? Or is the underlying programming very different? Can iListen surpass Dragon now Macs are Intel or has this got nothing to do with the programming or its operation?
Thanks.
For K12 students, neither Dragon NaturallySpeaking will do a very good job. This is because speech recognition programs require more mature voices and speech patterns that are grammatically correct. Most pre-adolescents will have a great deal of trouble with speech recognition programs, no matter who makes them. As far as iListen is concerned, MacSpeech supports anyone 14 or over.
In regards to comparing the two programs, we know from experience that both will provide you with 98-99 percent accuracy over time. Dragon will get to this point more quickly than iListen. There are two main reasons for this:
1). MacSpeech is a much younger company than Dragon. They were already several years old at the time we introduced the first version of iListen.
2). There have been very few changes in the underlying code base or the hardware instructions set for Windows and PCs since 1995. Contrast that with how MacSpeech has been tasked with completely rewriting our software three times: once to go from 68K code to PPC, again to go from Mac OS 9 to Mac OS X, and again just recently to go from PPC to Intel.
Nonetheless, we do have many customers who claim we are actually more accurate than Dragon - although we would never make that claim ourselves, and openly acknowledge that most people - at least initially - have a different experience.
The bottom line is that all speech recognition programs out there will give you about the same level of accuracy with a fully trained voice profile. At issue here is really that Dragon will get you to that point much quicker than iListen, not that it is more accurate no matter what.
If you have any further questions you may want to visit our online Knowledgebase for more information. Simply point your browser to www.macspeech.com, click the "Support" link at the top of the page, then click the link to our Knowledgebase.
Richard
You can, however, add as many medical terms to iListen as you would like all at one time using iListen's "Learn My Writing Style" feature. This allows you to add the words you use all at once without burdening your profile with thousands of other terms you do not use.
If the program is crashing, you need to contact our technical support team, as that is definitely not the experience the vast majority of our users are experiencing.
You can contact Support by pointing your browser to www.macspeech.com, then clicking the word "Support" at the top of the page. You can then submit a Ticket to get the fastest response.
Please contact our support team at [email protected] and they will assist you in getting iListen up and running. From what I can tell, it sounds like you might have FileVault enabled, which can cause problems with iListen.
You can also contact our support department directly by doing the following:
1). Point your browser to http://www.macspeech.com.
2). Click the word "Support" at the top of the page.
3). Click the "Submit a Ticket" link on the bottom right of the page that loads.
4). Follow the onscreen instructions.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I have enjoyed reading the comments and the review. I have a questions that I need answered that I did not see addressed here. My company has me doing interviews with folks for podcasts. Then we have them transcribed. From what I can tell DNS can handle duel voice transcription, how does iListen handle it? And do I understand you to say that iListen can not work on a Mac that is using FileVault?
Thanks,
Matt
iListen requires only one voice, and that the voice speak all their punctuation. We can't speak directly for Dragon NaturallySpeaking, but we have been told by those that use it that it is far less accurate when there is more than one voice on a recording and when the speaker(s) do not speak punctuation.
We have many journalists who use iListen to transcribe interviews, however. What they do is listen to the recording and then re-speak it in their own voice. This has the added benefit of allowing them to edit so the interview is transcribed correctly into written text (voices from an interview or lecture will be more conversational in nature and thus require heavy editing after transcription, negating much of the time savings).
Regarding FileVault, iListen is not compatible with it. This is due to a bug in Apple's Xcode and not iListen itself, however. (1.6.8 worked fine with FileVault, but 1.7 - the first version we have compiled using X-code - does not).
We STRONGLY recommend against using FileVault because it can cause complete loss of everything in your Home folder (this has happened to people at MacSpeech who have attempted to use it on more than one occasion), and it slows down the performance of your machine, which reduces accuracy for speech recognition.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Cauleen
I am severely dyslexic (a friend is writing this for me). I have a PM G4 with dual 1.25GHz CPUs with 1.25 GB of memory running os10.4.8
I was using viavoice however after the last few system updates it no longer works. I am thinking of buying ilisten MX. My concern is the same update problem will happen with ilisten. Can you assure me this will not be the case or must I be a good little sheep buy a PC and run dragon?
iListen 1.0 came out in 2006. The current version, 1.7 is the seventh major release of the program in that timeframe. We are already working on a version that will be compatible with Leopard.
So if you are looking for assurance that iListen will not "go by the wayside" consider yourself assured. Unlike IBM, we live and die by the success or failure of iListen, since all of the products we sell revolve around it.
I am greatly impressed with Chuck Rogers and am glad I placed the order. I will be able to give a good comparison opinion before long, so I'll let you know.
I would like to know if there is a forum for medical or chiropractic users to swap info, or how I could start one with MacSpeech.
Thank you all.
Michael
Thanks for your kind comments.
You can read a comparison of Dragon NaturallySpeaking and iListen in the following article in the MacSpeech online Knowledgebase:
http://macspeech.crmhelp.net/index.php?_a=knowledgebase&_j=questiondetails&_i=51
I don't know of any medical-specific forums specifically dedicated to those using iListen. We do have many doctors using iListen in their practice, however. One was even profiled on Apple's web pages:
http://www.apple.com/business/profiles/cleveland/
(The reference to MacSpeech is on the third page of the article.)
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I transcribe around 4,000 words daily through ilisten as I dictate all of my medical notes and correspondence. In preparation to use the system I ran 3 months of daily notes previously produced on Dragon through the "learn my writing style". The program picked up the medical vocabulary very well. Phrases I frequently use are transcribed flawlessly. After processing prior notes and going through the voice training, I consistently obtained 95% plus recognition. I added repeated phrases (such as segments of physical exams) to the macro's that can be brought up and transcribed on command. The cost of this program, plus a mac mini, was less than the cost of a Dragon alone offering the same capability.
After 5 years of using Dragon, and now watching my colleagues continue to suffer with Dragon, I am pleased that I changed to this product. It offers me better recognition, and therefor greater office efficiency, than I ever acheived with my retired pc and Dragon. I have been using ilisten over a year, dicatating into Appleworks. I have done very little training beyond what I described above, with recongition consistantly greater than 95%. My office notes are cleaner than what I get back from hospital transcription!
Dave Black MD
Gainesville, Fl
All speech recognition programs pretty much require the speaker be able to "write with their voice." This means they must have a certain command of the language that most children do not yet have. For this reason, we can not provide technical support for users under 14 years of age.
This is not a limitation of just iListen, btw. This is a limitation of all speech recognition programs.
Regarding training, we do have people who simply speak the text on the training panel and have the user repeat it to train. Once again, however, we do not recommend nor can we support the use of iListen with anyone under the age of 14.
Thanks, Laurie
Speech recognition does require a certain clarity to get the best results. We do have people with very mild speech impediments who are able to use iListen, but there is really no way to tell how well your father would do outside of him just using it.
Due to the very nature of speech recognition, it isn't possible for us to provide demo versions. For some people it can literally take weeks or even months to achieve satisfactory accuracy - especially if they have issues with their speech.
If your father's speech is not slurred very badly and he has quite a bit of patience to work with the software, iListen may be a very good solution for him. However, there is no way to tell in advance.
Chuck Rogers, Chief Evagelist
MacSpeech, Inc.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
TYPED BY HAND
I make digital voice recordings at construction sites and type up reports upon my return to the office. Will iListen function effectively if I play back my recordings through a microphone? Alternatively, if I directly connect the recorder to my Mac through the usb port and transfer my notes to the hard drive, can iListen then reproduce my notes?
The only answer I can give you is - it depends! A lot depends on how noisy the surroundings are, and if you are able to use a noise-canceling microphone when you make the recordings. Outside of that, you would also need to make sure you speak your punctuation and that yours was the only voice on the recording.
My gut tells me that the noise level on a construction site would be too unpredictable. Not necessarily always noisy, but sometimes noisy and sometimes not - which is going to cause poor accuracy as the software cannot adjust to different environmental conditions on the fly.
If you feel the noise level is pretty constant in the environment in which you would be using it, it may work OK for you, however.
We give you everything you need when you purchase with a microphone: you get the mic, a USB adapter, and a CD with the iListen software and complete documentation.
Like others, Mr Rogers' comments make me feel assured about dev and backup commitment. I've been an evangelist myself for a couple of products, both PC based. One of these, YeahWrite, I use the heck out of in ways I'm sure the developers never envisaged, and Parallels enables me to paste into my Mac apps. Sure, it's a bind, but not a big one.
So... Dragon could work the same way.
The advantage to me of purchasing iListen over Dragon, would therefore appear to be using it in all my mac apps directly, saving all that pasting. But (assuming accuracy to be about the same after a short while) is that the only benefit? And will it work with all my Mac apps, or only some of them?
Because here's the biter bit, guys: I can buy the standard version of iListen at the UK Apple Store for £129, or the standard version of Dragon for £59. Big difference for me! so.... want to convince me?!
Speech recognition is very subjective, since the user's experience is dependent upon so many factors that are external to the computer itself. With that as a sort of disclaimer, I will tell you this: we have customers who have used Dragon for many years and switched to iListen and swear iListen is better. To be sure, there are also those who use Dragon and tell us they'll come back "when we are as good," (although they do leave us scratching our heads a bit, since we already know we are as good since we have multiple customers who say so).
The short version is that Dragon is initially more accurate than iListen - in other words, iListen requires more time to adjust to your voice. But eventually it can be every bit as accurate as Dragon for most people.
In regards to what you can do with iListen - well, it works virtually anywhere you would normally type. We say "virtually" because it is quite impossible for us to test it with each of the 30,000+ Macintosh applications out there - but we have yet to run into one in which it does not work for dictation.
Our Command and Control capabilities leave Dragon in the dust unless you spring for their most expensive version, and even then - thanks to our implementation of AppleScript in the Mac OS - there are things you can do with iListen you can never do with Dragon.
Oh - one other thing. You mentioned you use Parallels. Did you know you can use iListen to dictate into your Windows applications while using Parallels as well? We haven't had time to completely test this (since obviously, that would not be our primary focus) but we do know you can dictate. You probably can't employ Correction, but as I said, we haven't had the opportunity to test extensively as yet.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I'm considering IListen. My question is if a mistake is made translating my voice through IListen in Word, can I make the correction by highlighting the text and then retyping the incorrect text like I would normally in Word or can you only make corrections by voice command.
The short answer is yes and no. There is nothing to keep you from editing your document either by hand or by voice should iListen make a mistake, or should you just want to change the way something was said. But if you edit by hand, you remove the ability for iListen to perform Corrections and learn from its mistakes.Here's why:
The best way to understand how iListen does its thing is to think of it as a "blind typist." Imagine you had an assistant that was the best typist in the world, but was blind. So you put the cursor where you want your blind typist to start typing, then you start typing and away it goes. That's how iListen works.
When iListen misunderstands a word, you use its Correction mode to correct the mistake. This presents a screen in which iListen DOES know exactly where all the words are, and when you make corrections in the Correction window, iListen then makes the same corrections in your document. It does this by literally counting back the number of characters to where the mistake was made, and then doing the edit.
The reason it has to do it this way is because like our blind typist in the example above, iListen doesn't know anything about the application into which it is typing. That's due to the way the Mac OS works. We can hold down the right arrow key until we get back to a certain point in the document, but we can't randomly go to the word a particular word. That level of interapplication interaction would require specific access individually to each application by its developer.
Another few dumb questions!
Firstly, I note there is a UK version of iListen available for download from your US site, and the price differential is reduced, due to the pound's strength against the dollar presently. Is the UK version geared to work with an English accent/British English usage? Being English, is this the version I should purchase?
Secondly; the download version doesn't come with headphones, naturally. I have a pretty reasonable Logitech headphone mic specifically recommended for Skype, and I find the gain is fine for Logic and Garageband, but they don't have noise reduction tech. How critical is the mic pairing?
If I were to end up ordering a boxed version of iListen replete with approved headphone mic, and i wanted to make sure it was at least better than my present ones, for quality of mic and monitoring in audio software packages (so I can just use one good set for everything, and not feel i'm spending money merely on duplication) which would be your recommedation for the best headphone mic?
The use of iListen into windows via Parallels sounds interesting. I'm an occasional published writer, and use YeahWrite for all my composition and organising - If I get iListen soon, I'll let you know how it works out.
Let's take your questions one at a time:
Yes, there is a UK version available for download from the MacSpeech site, and yes, if you speak with any variant of a UK accent, you should buy the UK version. The UK version of our software is more expensive than the US version because it has more command sets installed per our agreement with our English distributor. But I would encourage you to purchase a version that comes with a microphone from the UK (see below).
Microphone pairing is THE most important thing to getting acceptable accuracy. Your Logitech microphone was made for Windows computers and not designed with Mac OS X in mind, which prefers a signal gain about 10db stronger than Windows. This is why many Skype users on Macs report the people they are talking to have trouble hearing them. This is also why the Logitech microphone is not suitable for speech recognition on Macintosh. Also, if you are not using a MacSpeech certified microphone, then our tech support people would be unable to help you should you have trouble with accuracy.
Any and all of the boxed versions of iListen that come with microphones will work fine and will be better than the Logictech microphone you currently own. The ones we provide will also work better for Skype, Garageband, etc.
The very best microphone we have tested is the TalkPro Xpress from VXi. But the standard microphone we supply in the box works extremely well. The basic advantage of the TalkPro Xpress is that it is more comfortable and is slightly more accurate in noisier environments. Thanks - extensive testing with Parallels and CrossOver are on are extensive "to do" list, but it is not a high priority since our emphasis is on working with Mac apps.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
thank you very much for the helpful discussion. I am really impressed with Mr. Roger's attentiveness and his responses to questions. His answers are honest and free from unnecessary spin. I have a question: I am not a native English speaker. Would I be able to have success with IListen?
Thanks
As to whether you will have success with iListen if you are not a native English speaker depends, but not necessarily on what you might think.
We have several customers for whom English is not their primary language. Almost all of them have reported success with iListen, although in every case of which we are aware, the customer did have to work harder at it than a native English speaker.
A lot depends on your "pain threshold," which is probably a bad way to put it, since there is actually no pain involved. Nonetheless, even native speakers can be frustrated with accuracy with ANY speech recognition program due to a variety of factors, most of which have nothing to do with the computer.
What I can tell you is the following: I know of a physician who is also a customer of ours who had a very thick east Indian accent that seemed to me to also have some distinguishable traces of French. He came up to me at Macworld Expo a couple of years a go and started going on about how difficult it was to train the software. I could barely understand him, between his accent and the noise on the show floor.
Anyway, I was convinced he was going to complain about our software and how inaccurate it was. I was already formulating my response about how his accent just wasn't that well suited to any speech recognition program. But then, he totally shocked me.
In his thick accent, which, in telling this story verbally, I have never been able to even come close to reproducing, he told me that he stuck with the program and after nine months - (yes, nine months!) - he was finally getting 99% accuracy.
So that's what I mean by "pain threshold." We know for a fact from this customer and many other, less extreme cases that iListen continues to adapt the more you use it. So if you have the patience, yes, it will eventually work for you.
How quickly is a matter of variables too numerous to mention here, not the least of which is the "severity" of your accent. Also, keep in mind that the profile for someone who uses the software 4-6 hours a day will adapt much more quickly than that of the person who uses it only 1-2 hours per week.
I do have one final bit of advice for you: don't try to "trick" the software by mimicking a US or UK (or Australian or New Zealand, for our friends down under) accent. It won't work. Just talk the way you normally do and spend the time allowing iListen to adapt to you.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thanks for answering my previous question. I think I have a pretty good idea of how the software works.
Here's another question: Can I dictate into IListen using Word and then, after awhile, save the work, quit and then go back into Word (now not using IListen) and then edit as I regularly would do, by typing, in Word. Then after saving and quitting Word, jump back into Word, now incorporating IListen, to again dictate. The reason I'd like to do it this way is because if there are some mistakes I think I'll be able to correct them more easily by hand, but still be able to use the dictation feature in your product and the utilization of a hand-held Olympus message recorder. Thanks.
Sure. You use iListen just like you use your keyboard. Once you type into a document, you can save it and come back a day, week, or month later and then decide to type some more, you just do it.
iListen places no restrictions whatsoever on how you use a particular document. Once the text is in the document, it is just text - you can type, or dictate, or a combination of both.
The one caveat is that if you come back to dictate later, iListen won't know anything about the previous text - but you will still be able to employ Correction for what you just dictated (you just won't see the older stuff).
Using iListen couldn't be easier. You just talk and the words appear wherever they would if you had typed them instead - anywhere!
Again, the only thing you have to think about is whether or not you want to employ Correction - but even then, the rules apply only to what you just dictated. As soon as you say "Commit Corrections" it's like starting all over again as far as iListen is concerned.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I'm looking into purchasing iListen to use w/a new MacPro and have a couple of questions:
1) Web Browsing: With Dragon 8, I can click on links on a webpage (non-flash of course) by simply saying the name of that link. Does iListen work similarly, or better yet can iListen right-click links? I browse/read online a lot, and so this is a very important feature. I just don't want to have to keep switching back to Windows in order to use Dragon Naturally Speaking everytime I need to go online.
2) MouseAnywhere: Does MacSpeech have a videocast of this feature to show how it works? I am concerned about the speed/accuracy of such a feature (the back-and-forth of guessing incorrectly the number of pixels to tell the mouse cursor to move). I found I like Dragon Naturally Speaking's Mouse-Grid feature and this seems more like IBM ViaVoice's interface (which is less intuitive).
3) Dictating Code: I hope? This would be so awesome!!! I am hoping that between the tight integration with the OS and the text macros that this would be doable.
Thanks!
PS: Any chance your roadshow's coming to Denver, CO? Or if not, do you have any recommendations on reps or places that I can go see the product being used in person? I've tried looking for user groups but haven't had a lot of success so far.
First, Happy New Year to everyone!
Yes, you can browse web pages and choose links with iListen, but you can't choose the link by saying its name. Instead, you move through the links on the page by saying "Next Link" or "Previous Link." This is because iListen can't easily gather the names of the links. Dragon can because it has been especially programmed to do so (and I don't believe it works with any browser, but I could be wrong about that since I don't use NaturallySpeaking). We do hope to add that feature at some point, but due to hardware and software changes from Apple, we have had to completely rewrite iListen 3 times since it came out (once to go from 68K code to PPC, again to go from Mac OS 9 to Mac OS X, and then again this year to go from PPC to Intel).
We don't have a video of MouseAnywhere, but we do have several people who are familiar with Dragon's grid system who say they like our system better.
We also have several programmers, including some who work for Apple, who use iListen to write code on a daily basis.
We do have a road show tentatively scheduled for Denver in May. We do not have exact dates yet, however. We should be scheduling those dates in March. I don't know of anywhere else you can go in the Denver area to see iListen, but I am working on editing a video taken in Chicago this past fall. I hope to have a series of clips from that presentation on YouTube by the end of January.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
1) Web Browsing: Hm. That's a big deal IMO. I can't imagine having to say "Next Link" 50 times to get to a link on a web page (like a link in a blog comment for example). Dragon's solution is IE-only, but I've found a decent workaround in a FireFox plugin (see below). All that is required is that the voice recognition software be able to parse speech into number as if they were typed using the numeric keypad. Would iListen be able to do that?
The add-on is called Mouseless Browsing. For some reason this site is not letting me post the url so you'll have to google for it.
2) MouseAnywhere: Would you be able to point me somewhere to get more info on this? Is very difficult to get a sense of how well something does/doesn't work with "I like this better" and the likes.
3) Dictating Code: Sweet!! Are they using the BBEdit ScriptPak? or something else?
Thx Again.
In your answer to my question (#59) you said. "you just won't be able to see the older stuff), does that mean that I can't keep adding to a novel and have to start a new document each time I use IListen. What I'm getting at is that I just want to be sure I can keep jumping back and forth from Word, with and without the use of IListen, when I want to and see all the text I've worked on in the order I've worked on it in Word, with and without IListen. When you said I wouldn't be able to see the older stuff I'm thinking it's in a separate document and no longer visible in the novel I'm working on. If I can't see it, to me that means that it isn't in my novel. I don't really want to keep jumping from document to document copying and pasting. Have I misunderstood you? Thank you for being so patient with all of us. I can tell you want to be honest and forthright in your answers. I hope you had a great holiday and a fun New Year.
Mark
Yes, you should be able to use iListen to navigate links that way - but it will require a little bit of work on your part to create commands, but that is actually quite easy and our support team can help you out if you need help.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
This is going to sound a bit abrupt, and I don't mean it to be, but text is very sterile - so I apologize in advance if this comes off a bit harsh:
iListen allows you to enter text EXACTLY the same way you would enter it in an application. Let's say you start a new document and you dictate 5 pages and Correct it. Next week you open that same document and you put your cursor at the end of the first 5 pages you dictated the previous week and you JUST START TALKING! iListen will start typing at the place you put your cursor, in that exact same document in which you dictated last week. No new document is necessary.
When you employ CORRECTION in that document for the SECOND dictation session, only the words for that dictation session will appear in the CORRECTION WINDOW. This does not impact your Word document at all (other than to replace the text you dictated in the SECOND DICTATION SESSION ONLY with the corrected text).
I cannot over-stress this: you use iListen EXACTLY the way you would use the keyboard - anywhere, anytime, in ANY DOCUMENT WHATSOEVER. You can open the same document 1000 times and dictate into it 1000 different, separate times, and your dictation will appear in the ONE document, not 1000 different documents.
In order to understand how Correction works in iListen, you really need to use it - or, if you want to see some screen shots and read about it, you can download the User Guide and Tutorial PDFs from the downloads section of our support site (http://support.macspeech.com).
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Listening back, I'm not surprised. There is some huge electrical hum coming from the headset (Plantronics .85 Audio), over which my voice is barely audible. I've just checked the headphones on playback from iTunes, and they're functioning fine, so it would seem to be the microphone itself. Any suggestions? Have I missed something? Or did I just get a duff set? It's a real bummer, as I have a few days away from town, and was really looking forward to getting into this, but if it's all got to go back to the Applestore, this could take a while.
As I am sure you realize, a defective microphone is out of our control. We do not manufacture them and are prohibited by law from testing them before shipping, since components of the headset come in contact with the user's body.
It does sound like you received a defective microphone. You should return it to the place of purchase for a refund.
In the future, issues like this would be better sent directly to our support department rather than clutter the comments section of a public forum - especially since this is an isolated incident and not the typical experience of our new users.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
MacMall is selling the ilisten with a "free speech 2000" headset by phillips. Is that on the approved list...how does it stack up to others. what would you recommend for me with my mac mini (1gig ram memory and 1.42 speed running 10.3.9)
There is no such thing as a "free speech 2000 headset" from Philips. What MacMall is selling is the exact same package we sell for $149 from our web site. What comes in the box is a MacSpeech-certified microphone.
iListen will work great on your Mac Mini.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I am having trouble training iListen to "get" my dictation of C1, C2...T1...T12...L4/5, etc. for my chiropractic notes, even after using Learn My Writing Style. Do you have a suggestion? I wanted to try to fix it with phonetics, but have no idea how to spell phonetically.
Thanks,
Michael
Thanks for responding. Another question: Comp usa is selling the ilisten for $99 without the headset. Can you recommend a headset for me that comp usa is likely to carry. My credit is with them.
Thanks Again
Julian
Michael:
For requests like this, your best bet is to contact our support department by going to http://support.macspeech.com and submitting a ticket. Our technical support manager is very familiar with using the Phonetic Editor (I am not) and he should be able to help you.
Julian (and everyone else):
I have no idea what microphones CompUSA carries, but I doubt they have any that we have certified. You can find a list of those we have certified on our web site or by contacting our support team.
Here are a couple of brands to stay away from, btw: Labtec and Logitech.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Again thank you for responding to my message. I will be buying the ilisten....but can you fill me in on this issue; will Macspeech be updating the ilisten anymore for the "NOT INTEL" older mac system? Or will this be it? As I stated earlier, I have a non-intel mac mini.
Thank
JS
Version 1.7 of iListen is "Universal Binary" which means it works natively on both PPC and Intel computers. At the present time, our intention is to continue to produce Universal Binary versions of our software.
At some point in the future, you can assume we will discontinue to do this, but we have no intention of doing this any time soon. For the foreseeable future, upgrades to iListen will continue to work on pre-Intel Macs with a G4 or better processor.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Well I finally bought ilisten..the version that "Micro Center" had was 1.6.8 only. They told me that this version would run faster on my mac than the "Universal Binary" version. Have not used it yet..won't until next week..but were they right about this version for me? I have the pre-intel mac mini. I will buy the upgrade if it is worth it for me or better in some significant way of course.
Thanks Again
Thanks for your email.
What you were told by MicroCenter is absolutely NOT true. Version 1.7 is native on BOTH PPC and Intel Macs, and has many enhancements, including faster training and enhanced accuracy. It will, in fact, run faster on PPC Macs than 1.6.8 because it is a newer version.
The good news is that all you need to do is fax us your purchase receipt and we will get you a free upgrade. For instructions on doing this, please go to http://www.macspeech.com/support and click the link to submit a ticket. Follow the on-screen instructions to contact our sales team.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
We just came out with version 1.7 last June. Typically we go a year or two between major releases. We do anticipate a minor upgrade in the near future (I literally can't say when, since I don't know), but this will be a free upgrade to all registered 1.7 users.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I checked our support system and every request you have made was responded to within hours of when you submitted it. We have been getting reports recently from some customer who have told us they are not getting the automatic or human-generated replies from our support system. We are looking into the problem, but we don't have the answer yet.
I am going through all the open tickets in our system and personally emailing people today with instructions on how to check the status of their requests, so you should be getting something from me soon.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I have started using ilisten 1.7 again recently after becoming disillusioned with 1.6.
I must admit it is better than before and on an intel mac there is much less latency. I suspect I will keep on using it.
That's great news!
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I am considering iListen for the dictation of a non-fiction book. As I will be using both my desktop machine at home as well as an Ibook to 'write', my question is, how easy is it to keep the two systems synched so I don't have to teach both machines many of the jargon words not likely to be in your dictionary, as well as for dealing with my ridiculous Pittsburgh accent :)
thanks much!
k
It is extremely easy to keep your profile synced between two or more computers. All voice files are kept in something we call a "Voice Package File," which, by default, lives in the Documents folder of your Home folder. Simply copy this file to any external drive, such as a USB thumb drive, and move it into the Documents folder in your Home folder on the second computer and you are done!
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Came across this forum after a VERY disappointing start with ilisten. After reading through about half of the training dictations, I am getting laughable results - but, of course, I am not laughing. I cannot even get it to recognize the simplest of words, such as "can" and "can't".
Notably, even though I live in the US, I speak with a mostly English accent. From this forum, I've learned that there are differing US and UK versions of the software - there was no indication of such on the boxed program I bought from the apple store, and it is certainly not obvious from ilisten's home page.
It sounds from comments above that I should be using a UK version instead of the US version? If so, it would have been nice to have learned this before spending a couple hundred dollars on the program and headset. Is it likely to make much difference?
YES! To a speech recognition program, the UK language model is as different from the US language model as Spanish is from German. So you will get DRAMATICALLY better results by switching to the UK version. All you need to do is contact our support team and we will provide you with the UK version for free (as long as you can provide proof of purchase for the US version).To contact support, point your browser to http://www.macspeech.com/support and click the "Submit Ticket" link.
You also mentioned a headset. We require the use of a MacSpeech-certified microphone in order for you to qualify to receive technical support. The reason is simple: all speech recognition programs are designed to translate every noise they hear to words. Most microphones either are not noise-canceling, or they are designed with only Windows computers in mind - so they produce poor results when used on Mac OS X. (This is not iListen's fault - it is more of a Garbage In-Garbage Out problem.) So you really need a microphone we have certified to get the best - or in some cases, even acceptable - results.
BTW, if you look on our home page, the first menu item on the left side of the screen is "Products." If you click on that, you will see the second menu from the top is "UK Software."
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I started off using the mic bundled with the program, but, thinking that might be some of the problem, I went out and bought a plantronics .audio 510. How can I find if this is "certified" even if not sold through your website?
BTW, regarding the mic, ilisten's "Set up my microphone" clearly prefers the plantronics .audio 510 over the bundled mic, for what thay may be worth...
I need to buy voice recognition software due to RSI and have couple of questions.
Will iListen definitely not work on my G3 iBook?
Can you use iListen for UNIX commands in Terminal?
Can you use iListen to replace the mouse (i.e. open and shut programs, switch panes, save documents etc?)
Many thanks,
Edmund
Version 1.6.8 (which is included on the version 1.7 CD or disk image), will work on a MAc with a G3 processor, but version 1.7 will not. We would also like to point out that we can't provide much technical support if you are running on a G3, since you will get much better accuracy with a faster processor.
Yes, you can use iListen for UNIX commands in Terminal. We have many customers who use iListen for that.
You don't need to "replace the mouse" in order to open and shut programs, switch panes, save documents - all those commands are built-in to iListen. If you actually want to control the movement of the mouse, you would need our optional "MouseAnywhere" ScriptPak, which adds movement commands for the mouse. (Most people get a long fine without this. My advice would be to try iListen first, and if you decide you want to be able to move the mouse around, then add MouseAnywhere later.)
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
have been talking to support but thought I would open it up. In UK. Want to use iListen on MacBook Pro to control a CAD app. More specificaly we have extensive Quickeys shortcuts. Since Quickeys supports voice activated triggers I figure maybe we can turn off Apple speech recognition and use iListen instead. Since Quickeys prefs has facility to use a speech prefix (that alerts Quickeys that what is coming is a trigger) and since the actual voice trigger can be a natural word, e.g. "select line tool" it seems to me that iListen can be set up to accept voice input "select line tool" that will trigger Quickeys. Or am I making a lot of work for myself? In short have any readers had any experience of using iListen to control Quickeys?
What is the very best microphone to use regardless of cost. Two criteria are speed of execution of commands and obviously voice recognition.
Great forum thanks to Chuck Rogers who sets a great tone. Many thanks.
It is extremely easy to add commands to iListen. You can certainly do a command where you say "Select Line Tool" and have the same thing happen as happens now in QuickKeys. You can also have commands to operate any menu command, as long as the target application conforms to Apple's programming guidelines.
The very best microphone for use with iListen is the VXi TalkPro Xpress.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Currently I still use a portable microcassette recorder, and I dictate primarily in my exam room (which has a Powerbook G4), but sometimes I dictate in my back office (which has an intel Mac), and my assistant transcribes the tapes at her desk (also has an intel Mac). I am running a Mac-based electronic medical record.
Firstly, if I purchase iListen, can I install it on all 3 computers, or do I have to buy 3 separate licenses?
If I would have to buy separate licenses, would it then make more sense to use a digital recorder to dictate into, and install iListen only on my assistant's computer?
Finally, can iListen transcribe directly into my patient's medical record? Or will it transcribe into Word or Appleworks and then my assistant would have to cut and paste into the EMR. The EMR program is NetMedical from Practice Solutions Software.
Thanks!
From Quickeys support:
if iListen can do an AppleScript like this taken from Appendix A of the QuicKeys manual then it should work. I'm saying it should work because I have never seen iListen.
tell application "QuicKeys"
play shortcut named "Open TextEdit"
end tell
You would have to create a iListen script or whatever they call them with whatever voice trigger you want and then have it execute the above AppleScript to open the QuicKey shortcut named "Open TextEdit".
So it looks like I can use iListen to act as trigger for Quickeys. I assume I can use a natural language command to trigger an applescript, if so time to play.
Only thing is download screwed up and cannot see a way of logging in and attempting to download again.
I am about to buy ilisten v1.7 (with microphone) when I came across this comment on the web:
...Unfortunately, not listed among all these changes until after all the initial flurry of "update" emails, and buried in the list of changes now on the iListen 1.7 page, is the fact that MacSpeech REMOVED A CORE ABILITY FROM 1.6.8.
...Transcription, a core ability for years, is now an additional $75 add-on pack...The removal of Transcription as a core function in the iListen product makes it much less worthy of their base price...
Would you please comment.
Sayon
Now, about iListen -- if Chuck or anyone who has used the product can answer these questions, I will thank you:
1. Can I dictate into my video iPod, convert it to an AIFF file and will iListen be able to handle that?
2. If I can dictate into the iPod, does the quality of the microphone come into play?
3. If the iPod "sounds different" to iListen, can I create a different user/account to train iListen to understand "the other guy" on the iPod?
4. I assume it will work okay with vehicle noise, but how about radio in the background? Are other industrial noises particularly problematic?
Thanks!
Chronic RA has forced me to look into learning how to use voice recognition SW or give up what I love to do. I do own an intel based MacBook Pro and run parallels only when I have to use Apple OS 9 programs that we still rely on for teaching purposes.
These reviews have convinced me to try iListen, and I am counting on that great support staff as mentioned above.
It is pricy, but it if works it will be my salvation. It just means I will have to save up longer for my iPhone!
Thanks to all who took the time to write the comments that encouraged me to go forward.
I was so impressed with your responses and other users comments, I visited the Apple store in London. I noticed they only had the US version in stock. Following your comments on the differences between shall we say US and UK English, I explained to an Apple rep that, in London, the UK version might be more appropriate. He was unaware of the differences and promised to look into matters, as their policy is apparently to stock regional products wherever possible.
In the meantime, I ordered direct from your site in the US. Hope to get delivery tomorrow, and will let you know how I get on.
Kind regards, and keep up the good work.
Thanks
Michael
It seems atpm has stopped notifying me of new entries on this page - sorry about that. I am going to make several entries, addressing each of the comments above, one at a time.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
" if I purchase iListen, can I install it on all 3 computers, or do I have to buy 3 separate licenses?"
You can install iListen on as many computers as you would like, as long as only one of them are used at a time. If you use more than one copy at a time, separate licenses should be purchased.
KT further said...
"If I would have to buy separate licenses, would it then make more sense to use a digital recorder to dictate into, and install iListen only on my assistant's computer?"
Whether or not it would make more sense to use a digital recorder would be more conditional to your individual needs than the licensing requirement.
KT then said...
"Finally, can iListen transcribe directly into my patient's medical record? Or will it transcribe into Word or Appleworks and then my assistant would have to cut and paste into the EMR. The EMR program is NetMedical from Practice Solutions Software."
The answer is "probably." We know iListen dictates directly into AppleWorks and Word and thousands of other applications. We have no one that has ever reported that it would not dictate into NetMedical, so we have no reason to believe that it would not - but since we haven't tested it ourselves, we would not be able to guarantee that.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
While you can use QuickKeys to do things like open TextEdit, this would be redundant, since iListen already knows how to open TextEdit and many other applications just by saying "Open TextEdit." Again, I we would recommend simply using iListen whenever you can instead of using QuickKeys to do what iListen already does. While we don't have any reports of conflicts, we can easily see how a conflict could arise should you program iListen to type keystrokes that are already assigned a macro in QuickKeys.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Graham, I don't know exactly why the first tickets you submitted did not reach the team, but I see the one that did get to them got a favorable response. While I can't respond directly to most support issues, I am happy to look into things should anyone not be getting a timely response from our support or sales teams.
As you can see from my posts here, I also can't guarantee I will get notified by atpm when there is a new post, so let's all please be tolerant of the technology and forgiving of the people on both sides of it should it not perform up to our expectations.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Savon (and everyone else):
First, transcription is hardly a "core" functionality of iListen. Less than 10% of those who purchase actually use that feature. We did make it an optional product starting with 1.7 rather than increase the price of iListen. This allows us to pursue different development strategies for the "core" set of functionality in iListen, without taxing everyone - including the 90% of our customers who don't use transcription - for it's continued development.
Making this split allowed us to keep the price of iListen the same as it has been since the day it was introduced in 2000, AND allows us to collect additional revenue from those who use Transcription to pay for things like support for Video iPods and Nanos (2nd gen.)
I hope this makes sense for most people.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
George (and everyone else):
I don't know about a program called "iVoice." MacSpeech makes software called iListen. There is only one version of iListen. Our bundles refer to different combinations of microphones and functionality, in the form of transcription or command sets for different applications.
All of the differences are described on the individual product pages on our web site (www.macspeech.com).
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
"I was considering this software for radiology dictation. In that case, I would be looking at images in a foreground application while wanting the transcription to occur in a word processor in the background. Is this possible, i.e., can you choose the application into which transcription will occur, or does it have to to be the foreground application?"
iListen does its thing by fooling the Mac OS into thinking there is another input device (such as a keyboard) sending in keystrokes. As such, those keystrokes will always be sent to the front-most application.
The only way you could create notes while looking at images would be to have a small window using TextEdit either along side or at the bottom or top of the image. This would allow you to dictate without obscuring the image.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
"1. Can I dictate into my video iPod, convert it to an AIFF file and will iListen be able to handle that?"
Yes. We include complete instructions for how to do this in the documentation included with the TranscriptionPak.
"2. If I can dictate into the iPod, does the quality of the microphone come into play?"
The quality of the microphone ALWAYS comes into play. If at all possible, we recommend using one of the microphones we have certified with the iPod and a certified iPod microphone attachment (all of them have a jack for plugging in an external microphone). If this is impractical, we recommend the XtremeMac MicroMemo because it has a "boom" style microphone that allows the iPod to be held in a more natural position in relation to your mouth.
"3. If the iPod "sounds different" to iListen, can I create a different user/account to train iListen to understand "the other guy" on the iPod?"
Every sound source you use to create text from speech has the potential of effecting accuracy. We strongly recommend you create a separate profile for transcription vs. dictation.
"4. I assume it will work okay with vehicle noise, but how about radio in the background? Are other industrial noises particularly problematic?"
Doing transcription from an iPod does not allow any forgiveness of the rules that apply to normal "live" dictation. If anything, those rules are more in play because you are not speaking live. The more "in tune" your profile is with the environment in which you are speaking, the better accuracy you will achieve.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
First, thanks for your kind comments.
Regarding the price of iListen, I'd just like to point out that we haven't increased the price of iListen since it came out in 2000. In terms of the features it offers, it compares favorably with programs costing as much as $800 on the PC!
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
"I was so impressed with your responses and other users comments, I visited the Apple store in London. I noticed they only had the US version in stock. Following your comments on the differences between shall we say US and UK English, I explained to an Apple rep that, in London, the UK version might be more appropriate. He was unaware of the differences and promised to look into matters, as their policy is apparently to stock regional products wherever possible."
In the meantime, I ordered direct from your site in the US. Hope to get delivery tomorrow, and will let you know how I get on."
Yikes! I really wish I had seen this 2 days ago.
First, the Apple stores in London ONLY stock the UK version. They have never received - and never will receive - the US version (nor will anyone else in the UK).
Second, shipping and customs fees are really high when shipping from the US. Unfortunately we have no control over this as shippers charge what they charge us, and the UK imposes additional fees for importing goods.
I am sorry you were misinformed by an Apple employee in the UK. Please let them know they do NOT have the US version.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Michael (and everyone else):
We have tested the DSP-400 and it did not meet our certification standards, which are admittedly very conservative to insure the majority of our customers will not have problems with the microphones we have certified.
We do have customers who use the DSP-400 and are happy with its performance. All I can tell you is that if you already have one, try it and see what kind of accuracy you get. But if you have any problems, the first thing our support team will tell you will be to acquire a MacSpeech-certified microphone.
Regarding the DSP-200, it just barely pased our certification process, but it did pass. Plantronics insists that the two use virtually identical electronics, but despite that, our tests produced slightly better results - which were just above the threshold for certification - with the DSP-200. (That word "virtually" may be the issue.)
As I said above, our certification process is VERY conservative to insure the best possible experience using iListen with as wide a variety of microphones possible. Even those that "barely" passed certification will provide extremely good results for the vast majority of our customers. Naturally, this would mean that those microphones that barely missed certification would also work well for most people (and there is a small fraction of people for whom those that were "barely" certified would not work well).
The biggest thing to keep in mind here is that every person's voice is going to react differently to different microphones. When we test, 3 out of the 5 people who test must find it acceptable, and of those 3, one must be female (we have 3 male testers and 2 females). So just because something is not certified to work well with iListen, doesn't mean it absolutely will not work for you - it means if you complain about accuracy, we are going to tell you to get a certified microphone, that's all.
As an FYI, I have a voice that carries well (as those of you who have heard me speak will attest), so if a microphone can't be used by me, it doesn't even go to the other testers - experience has shown if I can't get it to work, neither will they.
(Yes, I was probably a bit redundant in all that, but better to over-explain than to be misunderstood on this particular topic.)
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Unfortunately after using ViaVoice on a PC for several years, I have found the current iListen correction system to be absymal. I am not the only one who does not like the system. In an otherwise positive June 2006 iListen 1.7 review Charles Moore at AppleLinks notes: "Frankly, I find that while the voice correction function has improved a lot from earlier versions of iListen, it is still clunky and annoying to use, and I just don't bother with it much."
I'll be honest. I really have no clue why the correction system is so bad. I have a bit of a linguistics background. I know about phonemes (minimal sound unit of speech), etc. I know that VR is better at phrases than short words, which is why someone above noted problems with "can" and "can't". There is not a lot of phonemic difference between these two short words. I also know from experience, no matter what IBM would have said on the issue, that ViaVoice clearly used a Variable Control Voice Actuator (VCVA) type process (threshold volume recording activation) since the beginning and ends of my phrases would often be clipped. For example, "the" at the beginning of a sentence was rarely transcribed by ViaVoice.
If you are still with me, here is my main point. Each ViaVoice VCVA segment could be corrected individually by voice for future recognition improvement or keyboard for speed. Each segment was effectively treated independently by ViaVoice. I have no clue why iListen cannot do the same thing while in Dictation mode, which would surely allow a better correction system. (In the current state of technology my suggestion would probably require giving up total voice control of my Mac while in Dictation mode, but I would gladly give up total voice control for better dictation results.) Unfortunately until the iListen correction system improves, I will be dictating in Dragon under Windows. :(
I have a very simple answer for you.
1). iListen actually does exactly that in the Correction window.
2). The Correction window is necessary because we dictate anywhere.
Let me elaborate:
In ViaVoice, you have something called the "SpeakPad." It is basically a version of TextEdit that has been modified to work intimately and directly with ViaVoice. So the changes "in dictation mode" in Via Voice are only available to you IF you use the SpeakPad - those same features are unavailable to you if you dictate directly into a different application's document window using ViaVoice.
In order to preserve the ability to dictate into virtually any application, iListen keeps track of what you say in a file that contains audio for everything you said, and it also keeps track of everything it types. Most of the time it keeps this information in sync. The catch is iListen cannot directly manipulate text in a document other than one that it owns, just like ViaVoice.
Dragon has the same issue on Windows, except that they have written support for SOME (but certainly not all) Windows applications so they can work directly within those apps. (We'd love to have the time to do that as well, but we keep having to rewrite major portions of the application to maintain compatibility with Mac OS X.)
OK - so when you are in the Correction window, you see everything you typed since the last time you said "Commit Corrections." If you change something (such as changing "can" to "can't") you just change it in the Correction window. Typically, no manipulation of the phonemes using the phoneme editor is necessary. It just works - just like ViaVoice's feature. In fact, it does exactly the same thing. It applies the correction, then examines the words on either side of where the correction was made, and makes adjustments to your profile to increase the odds of the correct word or phrase will be typed next time you say the same (or similar) phrase.
The other thing I can tell you is that the Correction window is one of the more polarizing features in iListen. You either love it or you hate it. Believe it or not, for each person who has complained about it, there are an equal number of people who tell us to improve it, but not remove it - since it lets them concentrate on the Corrections in much the same way the Story Editor in QuarkXPress or InDesign allows you to concentrate on content outside of the layout. (One of the improvements we have introduced in iListen, btw is the ability to use Copy and Paste to apply Corrections. This feature makes the Correction feature work lightening fast, but requires you not use the "Commit Corrections" command until you have finished correcting the entire document - again, a restriction placed on us by Mac OS X.)
Making a useful speech recognition product that works in Mac OS X is very challenging. Keep in mind that Dragon abandoned the effort before they ever got to a beta release. So if *THEY* found it too difficult, with all of the resources of Nuance behind them, imagine how difficult it must be for us - a much smaller company who must rely on the sales of one flagship product and its add ons and accessories.
As I have mentioned before (although I can't remember if I have mentioned it here), we have had to completely rewrite iListen three times from the ground up due to changes in Mac OS X and Apple Hardware. You might think introducing new features during one of those rewrites would have been a good thing. Well, a). we did that to some extent, and b). we have to be very careful. If we take what we know works and rewrite it, and then it doesn't work - we know where to look. If we introduce new things, then we don't know if it was the new hardware or software, or something about the new feature that is the problem. I am sure you can imagine how frustrating that is for us.
With all that having been said, while I can't give you specifics, I can tell you that we have had major improvements either "blueprinted" or even partially written for some time. Last year we were able stripped out Transcription from version 1.7 (see my other comment above), which was our first Universal Binary version of iListen. This allowed us to split the development path and introduce support for Video iPods without having to do a major revision of iListen at the same time. (I mention that just so you can see we do take these things into account, and we are making progress - just not as much as you - or we - would like).
Right now we are rewriting major portions of iListen so it will be compatible with Leopard. Hopefully, once Leopard comes out, we can once again switch our focus to implementing some of the new features we have been trying to implement for so long. I can tell you there will be improvements to Correction. The exact nature and extent of those improvements will depend a lot on what Apple has in store for us post-Leopard.
Chuck Rogers, Chief Evangelist
MacSpeech. Inc.
Regarding your points:
1). "iListen actually does exactly that in the Correction window." I wondered about that, which I kind of alluded to in my last paragraph of my previous post.
2). "The Correction window is necessary because we dictate anywhere." and its related quotes: "The other thing I can tell you is that the Correction window is one of the more polarizing features in iListen. You either love it or you hate it." I fall in the latter camp.
So, we agree on many things, including some of which I guessed correctly about. :)
My iListen suggestion is really very simple then. I want "iListenPad." It is that simple. I cannot type worth a lick, but fortunately I do not have RSI or quadraplegia. I can mouse around a little as needed. Maybe your marketing research shows that there are not enough people like me, but I bet there really are a lot of people like me. BTW, I ALWAYS used SpeakPad in ViaVoice, just as I will most likely always use "DragonPad" in Dragon NaturallySpeaking.
PS I knew about the aborted port of full-blown DNS to Macintosh. I thought it occurred before the product was owned by Nuance. Hasn't DNS been owned by a few companies? BTW, I actually considered mentioning the aborted port in my original post, because I thought it was relevant. However, I did not know the details, and I did not feel like Googling (TM) the issue.
PSS You may have actually inspired me to consider trying iListen again, but I want iListenPad. :)
I can't say whether or not (or when) we might do an "iListenPad," but I can tell you that one of the reasons we didn't go down that path in the first place was to differentiate ourselves from ViaVoice - the first version of which required the user to use the SpeakPad. With ViaVoice no longer a factor, all I can tell you for sure is that we no longer have to differentiate in that particular way, so who knows?
The "whether or not" in the above paragraph really has a lot more to do with how difficult or easy Apple makes it to keep up with changes in the OS (I can tell you keeping up with the Intel switchover was not easy, and Leopard compatibility is proving to be only slightly less problematic than the original transition to Mac OS X). There are a lot of other things to consider as well. I think anyone who used iListen in the past and then tried 1.7 would agree we made some major strides in initial accuracy. But if faced with a choice between a dedicated text editor and better accuracy, I think most people would probably favor accuracy.
Regarding Dragon, I won't go into their entire history here, but I think they were actually just "Dragon" when they originally considered a Mac product, then they were acquired by Scansoft, which later changed its name to Nuance when they merged with a company of the same name. Scansoft was actually the bigger company, but they took Nuance's name.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Again, thank you for the polite exchange. I wish you and your company well.
PS I really do understand your company's issues with Apple changing hardware and software. However, it is what makes a Mac a Mac. The Windows landscape is littered with previous generation programs and hardware that kind of work, while the Apple landscape is filled with products that simply work or don't. There is much less of the annoying Windows quasi-work.
Correction adjusts the profile after the fact. What you are talking about is a change to its interface. Introducing a SpeakPad-like feature would only change how Correction is performed, not reduce having to do it.
While I do understand that a contextual menu or something similar would be sexier, and I really do understand there is quite a bit we can do to improve the look of the existing Correction window (given the time to do so), I still don't understand why people think the existing Correction window is somehow less functional. If anything, it is more functional, allowing access to not only all the text, but both the choice list and the Phonetic Editor.
This is why I have always been more of a fan of the Correction window than a dedicated notepad that requires the (presumably) corrected text be copied and pasted where you really want it. For my money (OK, admittedly - YOUR money), I find the existing Correction window metaphor a lot more useful:
1). Dictate your text - it is already where you want it.
2). Correct it. All the text that needs to be Corrected appears in a window where you can concentrate just on the task at hand, without any other interface or formatting to interfere.
3). Finish corrections and POOF! Your original is corrected as well!
I certainly don't see this as any less convenient then the alternative:
1). Dictate your text in the dedicated notepad.
2). Make the corrections.
3). Copy the text from the dedicated notepad.
4). Paste it where you want it in the first place.
It has always seemed that a dedicated notepad adds an unnecessary step.
Just my opinion, FWIW.
Chuck Rogers, Chief Evangelist
MacSpeech. Inc.
I know that touching the mouse or keyboard before employing Correction is a big no-no, but I am not paralyzed so it is going to happen. BTW, I do not understand the point of dictating text where I "want it," when the sync problem can occur "if you have any automatic formatting turned on, such as indentation, in the original document (this would throw off iListen's Correction feature by inserting characters in the document iListen did not type)." [Quote from KB 131.] If I am word processing in Word for its formatting features that means I am going to have turn off some auto formatting then go back and format text. Plus, why couldn't a simple verbal script transfer the text from "iListenPad" to the app I want the text in?
I do not currently have iListen installed, so I cannot answer this question myself. Could I just dictate straight into the Correction window to achieve my "iListenPad?"
Again, thank you for the polite dialog. I REALLY WANT to use iListen. I cannot stand using Windows, but I type worse than that. If you know what I mean. ;)
The limitations on dictation are imposed by the operating system. Technically, iListen is not out of sync. It is the original document that is out of sync, due to influences allowed by the OS itself. We have no way to programmatically prevent Word, or Pages, or FileMaker, or Excel, or QuarkXPress, or Keynote, or wherever from allowing you to do whatever you might do to throw the program out of sync.
But you really do bring up the crux of the matter. If you are comfortable with dictating into a "pad" interface, where all you would see is text and you would not be able to apply formatting or indenting, etc., - or where that formatting may or may not translate exactly into what you want to appear in your final document (which was the inherent problem with ViaVoice), then why would you have a problem dictating without applying automatic formatting where you want the text to appear? (For those of you who are having trouble following this, it isn't that you can't *ever* use formatting or indenting - you can. You just can't apply them until after you have employed Correction.)
To answer your question directly, no, you can't dictate directly into the Correction window, but if you really must dictate somewhere and cut and paste, you can always use TextEdit, which, due to its sparse interface, is much less prone to allow the document to get out of sync with iListen.
Just for the record, the "out of sync" problem to which you refer is EXTREMELY rare and only affects a very, very small fraction of our customers - and even those seem to be satisfied with the fact that it is quite easy to avoid the problem - just don't touch the mouse or keyboard until after you say "Correct That!"
Finally, the argument becomes moot when you have adapted your profile well enough to achieve 98-99 percent accuracy. At that point, it becomes easier to just edit the mis-recognitions manually (no matter what speech recognition program you are using).
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I have a Telex microphone with both a left and right earpiece (which I assume makes it the stereo version). I can't find a model number anywhere on it, though it looks like a Telex H 851 USB Stereo headset/microphone pictured on the MacSpeech site.
The adapter part that goes between the headset and the usb cable says Rev A 0250 on it and above that in small type the numbers 301167-000. But no other model-like numbers anywhere.
Anyone know if the H851 was the only stereo headset Telex made? If so, this is probably that and would work with iListen 1.7, right?
Thanks,
TW
I checked, and your query never made it into our system. Like many other support sites, our site suggests articles from our KnowledgeBase that may address your problem. At the bottom of that list is a second button you must press in order to actually submit a request. This gives you the opportunity to cancel the submission if one of the articles addressed your issue.
What usually happens is that many people glance at this list and do not scroll down to see the buttons at the bottom of the list. So the ticket never gets submitted. You can also know if your ticket was submitted because you will receive an automatic confirmation. If you do not receive an emailed confirmation that your ticket was submitted, then it wasn't.
Regarding your issue, we have no way of telling you, based on the revision or part number or the USB adapter, whether you have a H-851 or not. Telex did make multiple models of microphones. All I can suggest is that you try it, and, if you don't get good results, acquire one of the newer microphones that are certified.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I also cannot find the difference clearly spelled out between the regular and MX versions of the software. It appears it's the addition of scriptpaks, right? If so, why not state clearly the MX bundle is different because it includes extra ScriptPaks rather than making the consumer flip back and forth to compare the two pages. Maybe it's somewhere else and I missed it. I clicked on the Products tab and saw a long list of graphics and links for all the possible bundles. Have you ever seen one of those commerce sites that have a comparsion chart of the similar products but with a feature list and little checkmarks next to the ones that are in each product or bundle? They're very user-friendly and yummy.
Not nitpicking, just suggestions. OK, OK, it's nitpicking. ; )
TW
The support software is from another company. The buttons are that way because they actually want you to read through the list of articles to see if there is one that answers your question. If you read through all the articles, you end up at the bottom, where you see the buttons. Pretty user friendly if you think about it!
Regarding the MX software, I'm not sure how much more clear we can make it. On the iListen MX product pages it lists the following:
What's in the box
VXi TalkPro Xpress USB headset/microphone.
MouseAnywhere ScriptPak.
iLife ScriptPak Bundle.
iDVD ScriptPak.
iMovie ScriptPak.
iPhoto ScriptPak.
iTunes ScriptPak.
Garageband ScriptPak.
There is also an article in our KnowledgeBase that describes the differences at http://www.macspeech.com/support/index.php?action=kb&article=80.
The matrix idea would be OK for other lines of products, but we only sell one version of iListen. The bundles all have various combinations of other stuff, all of which are described at the top of each of their respective web pages.
In our particular case, a matrix would be pretty boring, since we only have two "branded" models, iListen and iListen MX, and the only difference is you get a better mic and more command sets with MX - which, as I said, is clearly listed on the MX product page.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
In order to use Dragon NaturallySpeaking on a Mac you either need to boot up in Windows using Boot Camp (which means you won't have access to your Mac programs), or you can use a program called Parallels to run Windows and the Mac OS simultaneously. Either way, you can only use Dragon NaturallySpeaking to dictate into Windows applications - it cannot be used to dictate into Macintosh applications.
Our application, iListen, can dictate into both Windows apps (when running under Parallels) and virtually all Macintosh applications. It takes a bit longer to train than Dragon, but will eventually give you about the same accuracy.
Chuck Rogers, Chief Evangelist
MacSpeech. Inc.
Fast forward to now, where the dictation landscape for the Mac has changed considerably. I find myself in the peremptory phase of coming up with a good argument to convince my boss not to go with Dragon running on Parallels. I simply do not want to support more instances of Windows, no matter how "contained" they are.
His main argument seems to be a (hard-headed) assertion that Dragon is "just far better" but says he would love to be proven wrong.
Certainly I can point out the ability to dictate in just about anything, the low price of iListen, and perhaps some other general points. I even heard implied there's a money-back guarantee... is that true?
As I am not a current user of either modern speech product, so have to rely on my own research. He's had a lot of experience with Dragon.
I suspect he's sunk a lot of money into Dragon and feels he needs to save face there. Perhaps he doesn't want to learn something new, or wants the immediate accuracy of Dragon. Perhaps he just wants a reason to noodle in Vista, even if it's still on a Mac. I can't tell what's really behind his assertion.
So what's the gold argument here? The "switcher" argument? Would love some advice from the pro evangelist.
Cheers!
We try to not get drawn into a Dragon vs. iListen argument if we can help it. The simple fact is that ours runs on Macs and theirs doesn't. If I had to use very broad strokes to characterize the positive aspects of one over the other, it would be this:
- With iListen you can dictate anywhere, including Windows applications when using Parallels. There is probably some limitation as to how or if you can employ Correction when within the Parallels environment, but we really haven't had the time (or to be honest, the inclination) to test extensively. Even on the Windows platform, Dragon cannot dictate into any application and still preserve the ability to employ Correction.
- We are open and honest about the fact that it will take longer with iListen to achieve the same accuracy level as Dragon, although we know from many, many customers it is entirely possible - it just takes more time.
Because we know iListen can, over time, be as accurate as Dragon (98-99 percent is what our customers are claiming), we no longer offer a money back guarantee if the sole reason for wanting a refund is due to accuracy issues. This is because, as I said above, we know that iListen can, over time, be as accurate as anything else out there. Offering a refund because the customer hasn't worked with the program enough to achieve good accuracy would sort of be like a keyboard manufacturer offering a refund when the user doesn't know how to type.
The "gold" argument is simply that iListen, at $149 with microphone, is comparable feature-wise to the $900 Dragon NaturallySpeaking Professional version - with complete scripting capabilities AND the ability to dictate into any virtually application. The increased training time amounts to a fraction of the time you will spend using the program, and to not have to boot up Windows (regardless of any anti-Windows biased one may have), means more resources available for the other programs on your Mac.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
iListen has embedded commands in dictation mode to do many things you would need to do. For instance, while in dictation mode you can execute a text macro, move around in your document, employ Correction, and many other things.
Command mode makes iListen more accurate by not requiring you speak in any special ways to have your commands recognized vs. typed out. For instance, suppose you wanted to dictate the phrase " I am able to copy by voice by saying 'copy selection' and I can paste by saying 'paste from clipboard.'" If those commands were available all the time, you would rarely get them to type out if you wanted to dictate them.
So you use Command mode to issue commands such as those above. iListen also has a command mode shortcut. All you do is say "One Shot Command" and iListen switches to Command Mode, waits for you to utter a command, and then immediately switches back to Dictation Mode. Very handy.
iListen also has a spelling mode. This allows you to spell out a word that you would use infrequently and have no need of teaching iListen to recognize.
If you need further clarification, please take a look through our KnowledgeBase, or download our User Guide from our support site (www.macspeech.com/support). You can also contact our support staff through the support site with any questions you may have.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I'll let folks know how it goes for this first-time user of voice-recognition software. And I'll be sure to let my friend know if his info is out-of-date.
Nan
I'm contemplating switching platforms at the moment, and have 2 immediate questions.
1. I'm curious why there is not more discussion on the speed versus accuracy trade-off. Or to put it bluntly, there is only so much the software can do before you need to reach for more computing horsepower. What hardware and amount of RAM which you recommend for iListen for
a. acceptable, and
b. fantastic performance?
(It has only been in the last 6 months after my most recent upgrade to a fast dual core processor and 3 GB of RAM that the performance of the existing voice recognition software I use started to approach the marketing hype)
2. The only wireless microphone you have recommended is the Samson Airline 77. Which looks great. But not practical for a quadriplegic who would require assistance to change the battery. Can you recommend any other rechargeable wireless headsets? (along the lines of the GN Netcomm 9120 or the Plantronics CS60-USB)
thanks very much for your help,
cheers,
Allan H.
All speech recognition programs (ours included) benefit from faster processors and more RAM. But iListen works very well on G4 PowerBooks. 10.4.8 or the recently released 10.4.9 update are fine.
What you may find is that it will take you longer to achieve high accuracy (98-99 percent) due to the slower processing speed of the G4, but if you are persistent there is no reason why it can't be achieved.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
You can get acceptable performance with a G4 and 512MB of RAM. The more processing power and RAM you give it, the better it will perform. I hesitate to comment on what "fantastic" might be, because your perception and mine may differ. Also, there are to many other variables to consider, such as voice quality, environment, microphone used and it's position, as well as how closely one's voice matches the basic profile for one's gender with which everyone starts as they develop their own unique profile.
I can say this: one's experience with a G5 with 1gb of RAM will be better than one's experience with a G4 with 512mb of RAM. The same holds true for someone with an Intel-based computer over a G5, as well as an Intel with quad-core and 8GB of RAM over someone with a Core 2 Duo and 2GB of RAM.
Regarding wireless headsets, we have tested the Necomm and the Plantronics CS60 and found they do not provide enough signal strength to Mac OS X for speech recognition. Bluetooth is also out because it does not transmit enough audio data to iListen to be used at all.
Of the wireless headsets you mentioned, we could not get satisfactory performance out of the Netcomm no matter what we tried. I used the predecessor to the CS60, the CS50 for several months, and finally achieved acceptable accuracy. My intent was to start using it in my presentations so I could roam around the audience while using iListen. Ultimately, we scrapped the idea because we did not want to invite the questions about the wireless microphone since it took so long to get the profile to the point it could be used reliably.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
My PowerBook G4 has 1.67 GHz and 1GB RAM so it sounds like I should be all set with a bit of patience. My young son got a new MacBook with the Core 2 Duo chip for x-mas though. Assuming he allows me to use his computer (smile), would it be worth my while to do my "training" on his computer then copy the relevant iListen files to my PowerBook?
Thanks again. Several writer friends now are interested in hearing of my experience with iListen. I put in my order today!
Nan
I used a 17" PowerBook with similar specs for almost 3 years, including for the first 3 months after 1.7 was introduced and it worked fine for me. When I got my MacBook Pro, I just transferred everything over and immediately noticed a bump in accuracy.
I can't say with absolute certainty that everyone will have that experience, but there really isn't any reason why you shouldn't.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
thank you very much for your comments. More than 15 years after I last used an Apple computer The combination of OS X, iListen, and a fast Core 2 Duo is starting to look attractive.
But it will have to wait until the budget is looking a bit healthier.
Cheers,
Allan H.
Thanks.
Unfortunately, you have two problems: the MP3 file format and speeches.
iListen is unable to use the MP3 data file format by design. This is because that format is not particularly well-suited for speech recognition due to the way sound is compressed. Dragon NaturallySpeaking supports this format, but people who use it report that it is less accurate than WAV or AIFF.
The bigger problem, however, is the fact that computers are not powerful enough to translate speeches and lectures. You have the following requirements in order to get accurate transcription from a recording:
1). The speaker must have an established voice profile in iListen.
2). The speaker must speak his or her punctuation. Punctuation is key to achieving good accuracy. It provides aural landmarks the software can use to establish both what words to type as well as when to capitalize the first word in a sentence.
3). The environment can have a dramatic effect on accuracy. Many environments where speeches are given produce echos or feedback from PA systems, which will really inhibit the software's ability to produce accurate results.
It has been conservatively estimated that we are probably 15-20 years away from computers that can take speech recorded in "hostile" environments (i.e. without the use of a high-quality, noise canceling microphone) and reliably convert the speech to text without significant errors.
Regarding accuracy, we do have customers who claim they get 80% accuracy recording a speech or a lecture. First, while 8 out 10 words sounds like it is good, keep in mind that speech recognition never mis-spells a word, so finding all the errors is going to be difficult. Also consider you would have to go in and add all the punctuation and sentence capitalization manually - not fun.
Realistically, however, you won't get even 80% accuracy. 60-70 percent would be more like it. In many cases, accuracy could be as low as 30% if the speaker was not using a head-worn microphone and there was lot of ambient noise in the environment in which the recording was made.
I would like to stress one thing, however: people hear these limitations and blame the speech recognition technology, which they shouldn't do. The problem is that computers are not yet powerful enough to do the near real-time analysis that would be needed for speech recognition in these cases.
All speech recognition programs use a method called Hidden Markov Modeling to determine patterns. My personal belief is that we won't make huge leaps in the speech recognition area of technology - we won't come close to achieving that "Star Trek" like experience - until we have Quantum computers on the desktop. At that point, the software can substitute Quantum probability computing for the Hidden Markov Modeling method.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Here's my question: I also have a G-4 iBook (1.33GHz/1.25GB). If I train iListen on the iMac, can I transfer the profile to the slower iBook and will it get the benefits of the training already done?
Mike S.
Absolutely. You can move your profile to as many computers that have iListen installed as you would like, and you can create as many profiles as you would like.
See the following article in our online KnowledgeBase for more information:
http://www.macspeech.com/support/index.php?action=kb&article=59
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thanks for all your helpful responses.
Quick question: I read that iListen has a vocabulary of 30,000 words, and a "background vocabulary" of 300,000 words. What does this mean? I ask because I tend to be rather sesquipedalian :) and when using the "learn my writing style" feature iListen tends to not recognise many words that I use frequently which are nonetheless to be found in any reputable dictionary. Where are these 300,000 words??
Many thanks in advance,
Will.
There are 30,000 words that are kept in RAM all the time. These 30,000 words are in a dynamic list that keeps changing based on your usage, so it always has the 30,000 words you use most frequently. For most people, btw, this list is actually between 8,000 and 10,000, so about 20,000 of these words never change.
The "background lexicon" refers to the list of words iListen refers to when it can't find a word in the 30,000 it has in RAM. This list resides on the hard drive, inside the iListen application package. When iListen can't find the word it thinks you said from the 30,000 in RAM, it searches the 300,000 on the disk (which also include the 30,000 in RAM, btw.) If it can't find it there, it asks you to add the new word.
The 30,000 words in RAM always consist of the words you have added, plus any of the words you have used most recently, whether you use them frequently or not. So your 30,000 words is going to be different from my 30,000 words. (As newer words are added into RAM, words that have never been used drop off.) This is what helps iListen become more accurate and faster over time.
This technique, btw, was far more useful when both computers and hard drives were a lot slower. When we first came out with iListen, you could very much tell when it was searching the hard drive because there would be a noticeable pause and then it would type very fast to catch up with you. Today, you have to watch very carefully to notice this as the pause is much smaller due to faster processing times and disk access.
As is no secret, we license the core speech engine from Philips Speech Processing. Although the engine has undergone significant modification in order to work on a Mac (so much so that Philips insists we call it the "MacSpeech engine," as a matter of fact), there are portions of it that are like a black box to which we have no access. The list of words it has is one of those things.
Regarding words you would find in "any reputable dictionary," all I can tell you is that the choice of words for those 300,000 were selected in partnership with Oxford University Press, publishers of the Oxford English Dictionary. What I can tell you is that iListen will seem to not recognize plurals often, giving the illusion that it does not recognize that particular word, when, in fact, it does recognize the singular version of the word. Plurals are handled differently in order to maximize the number of words the program can recognize.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Michael
Thanks for confirming publicly what we have been preaching for years - when it comes to sound input, Macs are just more picky. So not every microphone out there - not even some of those $400 professional mics - will work well with iListen.
For those who are interested, the issue is that Mac OS X prefers an audio signal that, in general, is about 10db higher than your average Windows computer. For purposes of Voice Over IP (VoIP) like Skype or Vonage, or for gaming, it doesn't make much difference - you can always turn up the volume.
But when you "turn up the volume" for purposes of speech recognition, you also increase the volume of any noise - which, of course, is bad for speech recognition ( you would be surprised at how many sounds can impact speech recognition in even the quietest of rooms). This, btw, is why many Skype for Windows users complain that Mac users are harder to hear - the Logitech microphone that works well on Windows computers doesn't have the gain Mac OS X prefers, which results in a quieter output on the other end, making the Skype for Mac user harder to hear.
The issue is Mac OS X, not iListen itself - but that is why we certify microphones - to insure users get the best possible experience with iListen, or, if they are not, to at least be able to eliminate the microphone as the source of the problem.
So thanks again, Michael - I really appreciate the validation! BTW, of you have a lot of text macros in Dragon NaturallySpeaking, you should be able to transfer those by cut and paste into iListen if you are using Parallels.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I am considering ilisten as I am gradually moving from PC and nat spk to a mac environment. the description on the uk site mentions an optional transcription module - is this available in teh uk and how much is it please - I didnt find it on the apple or amazon sites. also will the uk version work with the olympus vn-2100pc dig voice recorder?
thanks
S
The TranscriptionPak is download only from our web site. It is $75 US.
It will not work with the Olympus VN2100 because Olympus does not provide a way for these files to get transferred to a Macintosh in a format iListen can use.
Chuck Rogers, Chief Evangelist
MacSpeech. Inc.
There are specific requirements for successful transcription from an audio file, both technical and practical. From a technical standpoint, it must be a 16-bit, 16kHz (or greater) file, mono is preferred and will sometimes render better accuracy. Also, not all WAV or AIFF files are created equal. The TranscriptionPak documentation includes both SUPPORTED and UNSUPPORTED instructions. "Supported" instructions means that if you have trouble getting it to work when you follow the instructions, our support staff is trained to assist you. "Unsupported" means we have tested the solution described, but cannot guarantee it will work for all people in all circumstances, and, if it doesn't, our support staff is not trained to figure out why it isn't working for you.
For instance, many people want to transcribe from an unsupported recording device. There are tons of devices out there. We only test those that provide both a way to move the file easily onto a Mac AND after doing so provide a means for converting the file to a format iListen can understand - either WAV or AIFF.
Speaking of which, not all WAV or AIFF files are created equal. This is particularly true of WAV files. We provide instructions for using Audacity to convert a file if you have problems. (This is an example of an unsupported procedure - all of the devices we support have well-documented processes for moving and converting the file, and our support team can help out if there is a problem.)
Also, all the same restrictions that apply to live dictation apply to transcription from an audio file. The speaker MUST have a trained profile with iListen, and MUST speak his or her punctuation. In addition, it is strongly recommended those using a recording device for deferred transcription use a MacSpeech-certiifed microphone if at all possible. Many a time we have had complaints of poor accuracy from people using transcription and we have them send us a sample, only to hear a room full of background noise and/or a pretty substantial echo as the sound of their voice bounces off hard surfaces, such as the table the recording device is sitting on. We have even had them send in recordings with multiple voices on the recording.
Successful transcription from a recording is actually very simple: if you hear something other than your voice on the recording, so will iListen - and since iListen cannot parse out other sounds the way your brain can, it is going to convert those sounds into words that come as close as possible to matching what it thinks it hears.
We always seem to get the same questions whenever transcription comes up, so I thought I would try to address them while I had a few minutes.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I am Belgian, native Flemish, hence a non native English speaker. I have bought a month ago iListen at the amazon uk shop. Since I always buy books at amazon uk, because of import taxes for amazon us, it is also amazon uk that was my choice for buying iListen. I have been trying out iListen for a week now, quite intensively, but with rather poor results in acuracy. Of course, probably this is also due to my being non native English. But, reading the very interesting responses of Chuck on this forum, I would like to continue trying to make it better. So now my questions.
I was not aware of the difference (and most of all the importance of this difference) between the UK version and the US version. So, most of all since I am used to buy at amazon uk (and had some books going to be shipped anyhow, etc...), I bought iListen there. So, this means that I have the UK version now. I think however that my English is muchy closer to US English than it is to UK English. Could this mean that I would get better results starting with the US-version?
My next question. If I would like to get the US version. If I add the US language pack, downloadable at the macspeech site to my UK version, does this gives me a US-version? Or is a US-version different from a UK-version with an added US-langauge pack?
Just a mention (because I saw it mentioned in one of the earlier messages). Also in my case, the planetronics headset included in the iListen package I bought gave me 'unsuitable for speach recognition' as test outcome. Also in my case there is a big humming produced by the microphone, and hence probably it is broken. During the week that I used iListen, I used a headset of a Logitec headset of a friend.
I meanwhile read on this forum that the logitech is not sensitive enough, and hence this explains probably also part of my very bad results till now. I will hence buy a new headset too (I suppose it will ne too difficult to get another one from amazon, since it is not easy to prove that this one works badly).
Diederik Aerts,
Brussels, Belgium
First, regarding languages: yes, the UK version of iListen is as different from the US version as Spanish is from German. All speech recognition programs (not just iListen) do their magic by analyzing phonemes, not entire words. (A phoneme is the smallest particle of speech - if it were chemistry, a phoneme would be an atom). Although a citizen from the US and the UK both speak "English," both the phoneme set and contexts differ. There are between 80 and 88 phonemes in the English language, depending on which source you look at. US uses about 40 and UK uses about 40. This includes all the different dialects (such as Midland, Liverpool, Scottish, Irish, Cockney, Welsh, etc. in the UK and Southern, Texas, Brooklyn, Midwest, Northwest, Maine, Boston, etc. in the US). While there is some overlap, it is different enough to require different language-versions of the software.
The other issue is context. In the US, if we say "skip to the lou," we probably don't mean we are merrily finding our way to the nearest restroom (loo in the UK). So the UK and US versions have contexts adapted for those cultures.
OK - enough with the language lesson. If you purchase the US LanguagePak, you will have the ability to create a profile using the US language model just as though you purchased the US version originally. When you start a new profile you will go through one additional step to select the language model you want to use.
Regarding the microphone, you need to contact our support department on that. If you are getting a hum in the headset, they will be able to give you instructions on how to minimize that (to help isolate the source of the problem), and arrange for you to get a new one, if necessary.
You can contact our support department by pointing your browser to http://www.macspeech.com/support, then click the "Submit Ticket"link.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Another question, or even perhaps proposal. Would it not be interesting to work out a kind if 'International English' version? Let me explain why I think this would be interesting. I am a professional scientists, physicist to be more specific. Hence I write most of my articles in English, which explains my interest in starting to use iListen. Now, it is well known amongst scientists that we understand in general very well the English of other non native English people, spoken (in conferences for example) as well as written (in articles), whether these other non native English people are French, German, or even Japanese or Chinese. But, it happens from time to time that an article is realy difficult to understand, or a talk difficult to follow, and each of these times this is because the writer or speaker is native English. This is not a big mystery, and most of all due to the fact that the vocabulary of the non native speakers is more reduced. But, it would mean that it makes probably sense (also commercially, if one things of the amount of non native people regularly using English) to have something like an 'international version' for non native speakers.
We do not have enough experience with English-speaking Dutch folks to address which particular language-version would be better. To be honest (and I certainly do not intend any offense by what comes next), we have never been asked the question before. In reviewing our records, we have fewer than 100 people who have ever purchased from your part of the world - and we have no way of knowing if those people might be native UK or US speakers. What I mean to say by this is that MacSpeech, being a small company, very much as to "pick its battles." To put this in economic terms, the market for a version that addresses the issue you bring up (the number of people who would benefit from the research and development for and of such a product), would not be willing, as a whole, to bear the cost of that research and development. In other words, given the cost of researching and developing a solution, and the fact that such research and development might well take months to accomplish, would you be willing to pay upwards of $1,000 for the resulting product? Perhaps you would, but we doubt there would be enough Mac users in your part of the world who would be willing to commit to that kind of an investment in order to move accuracy from 90% to 95% or better when all is said and done.
Regarding your comments about "International English," there are way too many problems involved for that to be practical. As you said, there are non-native English speakers that are French, German, Spanish, Japanese, and Chinese, and many, many more. Our existing products already work well for these speakers - although they do have to work harder on getting iListen (or any other speech recognition engine) to adapt to their voice.
The larger problem is that there is no "International" set of phonemes. The German native English speaker will have his or her own set of phonemes, as will the Spaniard, the Franco, and those from asian countries. Each will pronounce things according to the slant his or her native tongue gives it, so there really is no "International" English that can be developed.
Our best example is from a physician who came up to me at Macworld Expo a couple of years ago. He started talking to me in a very thick East Indian accent - very thick. At first, I could not even understand him so I had to ask him to step aside - away from the booth - where it was less noisy. When I started to understand him, I thought he was complaining about iListen and was getting ready to gently explain to him that it would be very difficult - if not impossible - for iListen to understand him. Before I could do so, however, he says "...but I stook wit it and afder 9 monts - 9 monts! - I am finally gitting 99 purrcent accuracy!" (cheap attempt at typing an east Indian accent added for effect).
I have no idea if he really was getting 99% accuracy, and I suspect that it was probably more like 96 or 97 percent, given his accent. But it was an enlightening experience, and one of the reasons I remain convinced of how good our product really is.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I know very little about any "foreign" languages. However, I am a native American English speaker, which enables me to detect English written by non-native speakers--grammar and words are not quite right. An "international version" of iListen for non-native English speakers makes some sense, but I am not sure that it would work. (I like your restricted vocabulary observation. It strikes me as accurate, but I would want to see some research.) The overarching issue that these personal observations address is "context". I am not sure that context is the same for all "international" English speakers since grammars may be different, which means that an "international version" of English for iListen would probably not be as helpful as you would like.
Another issue is that "international" people can pronounce English with different phonemes. For example, growing up I knew a Persian (he did not consider himself Iranian) man who spoke English with a German accent because that is the accent that his English teacher had. I cannot begin to imagine how that phoneme mix up would initially confuse iListen.
I could not resist adding my two cents, but I thought your ideas and observations were very interesting.
Having said so, I do believe that rather suddenly there might be a break through, for the scientific world at least, to start speech recognition.
Coming back to the 'International English'. There was a time (20 years ago) that scientific journals had native English editors, who would correct the English of non native speakers before the accepted articles would appear in print. This however has almost completely stopped, at least in the fields I work in. There starts to grow an acceptance of the 'International English' as a kind of new language. And, of course, if one thinks of 'how languages' have evolved in the past, this is most probably what is going to happen for the future, the birth of a kind of (probably simplified) standard International English (if Chinese does not take over of course :-).
There also remains my experience, however shared by many other non native English speakers (we discussed this matter at conferences several times), that 'all the non native speakers amongst each others' do understand each other better than they understand native English speakers (in general). Phonemes might not be so different, is my guess, and perhaps sound most of all 'very different' for native English speakers.
Well, I will certainly be interested to follow how all this evolves.
One last remark Chuck. Macintosh is very much present in European universities, all over Europe by the way. In Belgium I estimate that around 30% of scientists at universities use mac. I am heading a group of 25, and we all use mac.
The issue is simply computing power. Personal computers are not yet fast enough to do the real-time processing of the hundreds of thousands of variables involved in analyzing "everyone's" speech (meaning, non-native vs. native speakers, children, those with speech impediments, etc.). Personally, I don't think we will have the breakthrough you are talking about until Quantum computing makes it onto our desktops. That is a reality that is just now in the early stages of exploration. It will probably be 15 to 20 years before we see something like that.
In the meantime, things will still improve by baby steps and individuals such as your self will probably need to spend more time working with software such as ours to get the same results as a native speaker.
Regarding your market, we have no doubt the market is there. But it is a matter of resources. MacSpeech has been growing by leaps and bounds each year since we shipping iListen in 2000. Even so, we have barely tapped the US market for speech recognition, let alone those for overseas markets. To say we would like to tap into these markets is a colossal understatement. But we need to concentrate on the low-hanging fruit in our own backyard first. For a small company such as ours, that is simply what makes the most economic sense.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Upon viewing some of the earlier comments in this thread, it seems that Dragon Naturally Speaking is essentially easier to get up and running for a novice like myself.
I really can't agree that Dragon is easy to "get up and running." With iListen you can be dictating in as little as 5 minutes from the time it is installed. The difference comes in your initial accuracy. INITIALLY, Dragon will be more accurate by anywhere from about 5 to 7 percent (depending on a variety of factors). But over time, this advantage disappears, as many former Dragon users who have switched to iListen will tell you.
More importantly, however, is the following:
- iListen allows you to dictate into virtually any Macintosh AND Windows application (if you are using Parallels). Dragon only works with SOME (not all) Windows applications.
- At $149 with a microphone, iListen has features Dragon doesn't have until you spend over $900. Even then, iListen has some unique features that Dragon doesn't have at all - such as complete compatibility with AppleScript.
iListen is far more "plug and play" on the Mac since it does not require you purchase both Parallels AND Windows in order for you to use it. If you are a novice, as you say, and you have a Mac, iListen will be MUCH, MUCH, MUCH easier to get up and running over Dragon NaturallySpeaking.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I need to be able to transpose from a recording that I would be listening to while I speak into iListen set up in MS Word. What is the average "learning" time for the software? I'm running OS 10.3.9 on a G5 dual processor tower (purchased in 2004) with 2G RAM. I will be putting Tiger on in the next few weeks.
I am an educator and can purchase the software from my university bookstore. The iListen package contains, I believe the standard microphone. Do I need to be looking for a superior mic, and if so which one(s)? The price is $145. Is this the package that retails for $149?
I'm sorry if my questions seem a bit simple, but I just need a speech recognition package that is simple and straight forward.
Thank you again for your time.
Your questions aren't simple at all - well, most of them, at least. Let's work our way from easiest to hardest:
First, your bookstore probably sells our $149 version for $145. This is fine. All the microphones we sell will give you about the same accuracy in a quiet environment. The more expensive mics will perform better in noisier environments and are more comfortable to wear for extended periods. Think of it this way: the microphone we provide at $149 is a standard microphone: the "Chevy" or "Ford" of microphones, if you will. There is absolutely nothing wrong with a Chevy or Ford - both will get you from point A to point B with no trouble. The more expensive microphones are like a Cadillac or Lincoln. More bells and whistles, and more comfortable - but the journey from point A to point B pretty much takes the same amount of time (it just might be a bit more enjoyable, though).
Your other question - what is the "average learning time" cannot be easily answered. I recently saw an email in our support system from someone that was getting 96% accuracy after reading just one training story. On the other side of the spectrum, I know of an east Indian doctor who took nine months to achieve 99%.
The software never stops learning your voice. The more you are willing to employ Correction, the more accurate iListen will become. There are so many variables involved, that it is impossible to predict in advance what your experience might be. What I can tell you is this: you should be getting between 85 and 95 percent accuracy after reading the first training story (which takes about 5 minutes). In fact, if you are getting less than 80%, there is about a 99.9% chance that something is not set up right. If that happens, our support team is ready and willing to help you.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I'll give it a try, pick up the package today. Since I don't have the program yet, I can't really understand the concept of "Correction". Is this a preference that can be set? (Another simple question.)
Nice job of evangelizing first of all! I'm pretty much convinced to give iListen a try, but being an American living in Italy, I'm not sure what would be the quickest (and most economical) way to get up and running with the US version of iListen.
I imagine the version sold by Apple Store Italia will be the Italian version, or perhaps the UK version (or maybe even a combination of the two?). Is that right? The price there actually is a bit strange, though. It's 279 euros with a Plantronics .Audio 85 headset and microphone. Given the strength of the euro, that seems about 100 euros too high to me even with the 20% sales tax they apply here. Given this high price, I would be reluctant to buy this PLUS the US LanguagePak.
I could buy the downloadable version of iListen and then buy the microphone separately, but I suspect I may have a hard time finding the exact models you certify available here in Italy. A quick search for the Plantronics .Audio 85, for example, hasn't turned up anything so far other than bundled with iListen. And it's not even on Amazon.co.uk, not that that would matter since they don't ship electronics outside of the UK and Ireland (unless bundled with iListen software...).
What about shipping times and costs to Italy from the MacSpeech online store?
Other ideas?
On an unrelated note, I'm a pretty fast and accurate typist (perhaps at professional typist speeds when I'm "in the zone"). How does dictation with iListen (or indeed any "competing" software) typically compare with typing in terms of words per minute (considering accuracy and correction issues)?
Thanks!
Grey
Regarding purchasing the software, the Apple store in Italy will be selling either the Italian version or the UK version. I am not sure which, since Apple purchases from our distributors, I am not sure which version they are stocking in Italy.
What you could do is the following: purchase iListen in Italy (regardless of what version it is) with the headset and then either fax or PDF our support team your proof of purchase. They could then arrange a download of the US version for you.
Regarding speech recognition vs. typing, once you have an accurate profile there is no contest unless you can type over 100 words per minute. Outside of that, however, every person seems to have an individual "threshold" at which speech recognition becomes a viable choice for them personally. There really is no average, since we are all unique. Some people get 95% accuracy and think it is wonderful (regardless of what speech recognition accuracy they are using), while others will get 98 or 99 percent accuracy and complain that correcting the mistakes isn't worth their time. So that is something only you will be able to decide.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
OK, if you consider that I'll end up with Italian and US LanguagePaks for that price (and that by buying here as a business I can write off the VAT and the expense), the difference isn't that great, but still...
Anyway, thanks a million! You're a great representative of what I hope will prove to be a great company and a great product.
Cheers,
Grey
We don't have any control over the prices at which the non-English versions of iListen are sold. The distributors are responsible for all the translation work, but we have final approval on the finished product, but no control over the pricing.
At $300 US you would be better off ordering directly from our web site. Even with overseas shipping, it would probably save you $40-$50. One tip: make sure if you order from our web site you have it shipped UPS or FedEx, not US Postal Service. We have had very poor experiences with overseas shipping when the US Postal Service is involved.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
We don't have any control over the prices at which the non-English versions of iListen are sold.
Yeah, I understand that. It wasn't a criticism really, just a curiosity.
I'll check out the direct order option, too.
Thanks again!
Grey
I went through all of the iListen stories and ended up with an abyssmal accuracy rate. I then found out that reading all of the stories in one sitting is a bad idea because of voice fatigue (I wish your software would say that before I wasted all the time in that training.)
Anyway, I want to learn how to train your software but I can't find out how. Your website has no information about it, not even in the knowledge base. What I mean is this ... give me a 10 day plan, assuming I'm willing to work 1 hour every day ... to make your product accurate enough for me to dump Dragon, which I'm completely willing to do. Because, to be perfectly honest, the voice software built in to Vista is working better than iListen at the moment.
Also, I'm curious whether Apple is going to sell you guys completely down the river by bundling a voice to text product in Leopard or 10.6.
Thanks for your email.
Our intel is that we will not see dictation capabilities built-in to the Mac OS until it can achieve 99.9% accuracy for 99% of those who use it, including children. I am not at liberty to tell you who I heard this from, but it is reported to be Steve Job's attitude regarding the viability of speech recognition in general.
It is not surprising that you would get better accuracy with Vista, for many of the same reasons stated earlier in this comment section.
Regarding your accuracy issues, we have several articles in our KnowledgeBase (I know, because I wrote them!). Just go to www.macspeech.com/support and enter "Accuracy" into the "Search KnowledgeBase" field.
Make sure you read through "How Can I Get The Best Results With iListen?" and follow the instructions in "TIP: Creating A Good Voice Profile." If you still have problems, go to www.macspeech.com/support and click the "Submit Ticket" link and ask our support staff for assistance.
BTW, whenever you contact our support staff, you can speed things up a bit by ALWAYS providing the following information up front:
- What version of iListen are you using?
- What model Macintosh do you have?
- How much RAM does your Mac have installed in it?
- What version of Mac OS X are you using?
- What type microphone are you using and how is it connected to your Mac?
- Do you have any anti-virus software installed?
Those are the questions they ask back 90% of the time, pretty much no matter what the issue is.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I just stumbled on this website as I was looking for iListen information and reviews. My question is - Can IListen dictate into Final Draft?
Thank you. :)
I have started to use the US version as a native Dutch speaker instead of the UK version that I bought first. Perhaps interesting for others to know, this indeed works better from the start. Hence what I imagined intuitively, namely that the English of non native Europeans (at least Dutch) seems to be closer to the US English than to the UK English.
Apart from this positive results I however encounter a strange problem. When I say 'correct that' to start making corrections, most of the time 'nothing happens'. I see on the bottom of the Feedback window 'I heard correct that', which seems to indicate that my saying was understood, but the correction window does not appear. This is weird because I did not have this problem with the UK version. I have been looking a long time in 'trouble shooting', but could not find a solution. It is very annoying, because eager to make the corrections, to get better results step by step, I am stuck now with the mistakes.
Is there a way by using the keybord to make appear this correction window if by voice it does not work?
Diederik Aerts
This is not really the right place to address problems an individual might be having with a particular feature, and I am not the right person to be handling that anyway. Please contact our support team by pointing your browser to www.macspeech.com/support and click the "Submit Ticket" link. I am sure our support team has probably encountered this before and can help you out.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Yes, you can dictate directly into thousands of Macintosh applications including Final Draft. We also have a Final Draft ScriptPak. (ScriptPaks are not necessary for dictation, but they enhance iListen by allowing you to control a particular application's menus and keyboard shortcuts by voice).
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I decided to try it out, and I am hoping to get excellent results with it!
God Bless!
:)
I am extremely impressed with your insight and fair balance in explaining the fine details of iListen and and your competitors. I have never used a voice recognition software but I am a very patient person. I have 2 questions. First, You have a package for $299 which includes software and Olympus recorder and options for ordering the best mic(36.00) and instructional videos(36.00). I am a pharmaceutical representative and speak to numerous physicians everyday. After each sales call I need to take good call notes on what I talked to the physician about. I am currently using a small tape recorder and recording the events of the call after I get back to the car. I come home and transcribe these notes into my WORD program on my MacBook Pro (2GHz with 2G RAM). I am looking for a complete package that will take the burden of typing off of me. My question is: Is the package you offer the one I need? and,
after getting everything installed and trained, can I plug in the Digital Voice recorder in to the computer, start the transcription process and walk away and let it transcribe til completion? Remember my notes are for my use only and not for any formal business purpose so I can have the patience to build my profile. Just like college, not many people could decipher my notes but I knew exactly what every note meant.
The process for transcribing your speech from a voice file is very simple:
1). Record some speech on a MacSpeech-certified voice recorder. For the very best results, we recommend using a Video iPod or iPod Nano (2nd generation). The iPods give you better accuracy than any of the digital voice recorders on the market because they sample at a higher rate, providing iListen with more data from which to do the transcription.
2). Copy the files from the recording device to your Mac and convert them to a format iListen can use. The process for this is a bit different whether you use an Olympus recorder or an iPod, but it is very simple to do and we provide complete instructions.
3). Have iListen "listen" to the file and transcribe it to text. You need the optional TranscriptionPak to do this. The TranscriptionPak is, of course, included with all our "Transcription Solution" packages.
Keep in mind that iListen's transcription feature is for the convenience of being able to do dictation while you are away from the computer. It is not for recorded interviews, speeches, or lectures. Computers simply are not yet powerful enough to do the kind of processing necessary to do near real-time transcription for that kind of data.
The above disclaimer is mainly for the benefit for others who read these comments. From your description, you should be fine, and any of the Transcription Solutions we sell should do the trick for you. If you decide to go with a solution for an iPod or iPod Nano, keep in mind that we only sell the microphone, not the iPod itself.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
1. What is it about the VXi microphones that make the UK concerned? That is, why are they considered dangerous?
2. If I want to get a microphone that is RoHS compliant, which one is best?
3. How much better is the VXi Talk Pro Xpress USB than the Plantronics .Audio 85 USB Headset?
4. If I buy the boxed set of ilisten 1.7 with the VXi Talk Pro Xpress microphone, but I later find that ilisten works as well with another microphone, will I be able to return to the VXi microphone?
Also:
5. I have a PowerBook G4 10.3.9. Will that be enough to run ilisten 1.7? (I see that the 10.4 is recommended.)
6. I don't often use iDVD, IMovie, iPhoto, or Garageband. Is there any reason for me to buy the ilisten MX 1.7 version versus just ilisten 1.7 with the microphone, or even the software and the microphone separately?
Thanks!
Thanks for your comments. Let's take them right down the line:
[1. What is it about the VXi microphones that make the UK concerned? That is, why are they considered dangerous?]\
Not having been certified as RoHS compliant and being "dangerous" are two entirely different things. If VXi microphones were dangerous in any way, they would not be allowed to be sold in the United States. Simply put, as VXi does not have any resellers in the UK, they have not submitted their microphones for RoHS certification, so therefore they are not viewed as being RoHS compliant. BUT, there's a loophole. Since the VXi models we sell were in manufacture before RoHS regulations went into effect, they are exempt from the requirements. The RoHS compliancy requirements apply only to models introduced AFTER July 1, 2006.
[2. If I want to get a microphone that is RoHS compliant, which one is best?]
The best microphone we have tested that is RoHS compliant is the Plantronics .Audio 85.
[3. How much better is the VXi Talk Pro Xpress USB than the Plantronics .Audio 85 USB Headset?]
All of the microphones we sell, including the .Audio 85 and the TalkPro Xpress, provide about the same accuracy in a quiet environment. The TalkPro Xpress will provide VERY slightly better accuracy in a noisier environment. The .Audio 85 also has the benefit of being stereo and foldable.
[4. If I buy the boxed set of ilisten 1.7 with the VXi Talk Pro Xpress microphone, but I later find that ilisten works as well with another microphone, will I be able to return to the VXi microphone?]
I don't make those decisions, but I don't think so. That's sort of like saying if I buy a car with this great stereo but then find a better stereo, can I return the one that came with the car for a refund. Also, while we can't make guarantees due to potential differences between individual voices, the vast majority of our customers would agree that the TalkPro Xpress is the very best microphone for use with iListen. My own personal experience is that that holds true in noisy environments, but in a quiet environment (such as my home office), virtually every head-worn microphone we have ever certified works almost identically well.
[5. I have a PowerBook G4 10.3.9. Will that be enough to run ilisten 1.7? (I see that the 10.4 is recommended.)]
As long as you have at least 512MB of RAM in your machine it will work fine.
[6. I don't often use iDVD, IMovie, iPhoto, or Garageband. Is there any reason for me to buy the ilisten MX 1.7 version versus just ilisten 1.7 with the microphone, or even the software and the microphone separately?]
As stated, it is difficult to answer the question. We only sell one version of the iListen software itself. We have various bundles available, of which "MX" is one. You can purchase 1.7 with our standard mic (either a VXi P41R or Plantronics DSP-200 - our choice not yours) for $149, with a Plantronics .Audio 85 for $159, or with a VXi TalkPro XPress for $179. All of these bundles are available on our web site. If you purchase the "MX" version (at $219) you get the extra ScriptPaks in addition to the TalkPro Xpress. That is the only difference.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I think Alice's question was actually different than the question you answered:
"will I be able to return TO the VXi microphone?"
Unless that's just a typo, I think what she meant was would she be able to use different microphones for the same voice profile?
While I'm here, I'm finding iListen's dictation accuracy to be pretty good even without a great deal of training, but command recognition can be very frustrating. I find I'm having to change a lot of the common commands to use words that I naturally pronounce more clearly, which seems more important than command length. Sometimes, though, it seems that iListen misses the first syllable of my utterances, and it seems that is effecting recognition accuracy, too (both in dictation and in commands).
Cheers,
Grey
D'oh! [Slapping of head sound]
It's late and I am a mile higher in Denver from the New Orleans swamp to which I have grown so accustomed. My apologies.
Yes, of course you can return TO the VXi microphone. You can have as many profiles as you want, so even if you need to create a new profile for a different microphone, you can still return to the old profile.
The issue here is exactly how different those microphones are from one another. Using the "Set Up My Microphone" procedure should, in most cases, compensate for most differences. But it isn't out of the question that different microphones may be so different from each other that they would each require their own profile.
Regarding your particular issue, Grey, I can't say for certain, but try moving the mic just a bit closer to your mouth and do the "Set Up My Microphone" procedure again. The mic should be 2-3 fingertips from the corner of your mouth. When properly positioned, you should be able to take a drink from a normal-sized glass without touching the mic or spilling the liquid.
When iListen types "extra" words (such as to, and, a, of, the, etc.) it usually means the microphone is a bit too close. If it seems to be missing small words or parts of words, it either indicates the mic is too far away or not calibrated properly or both.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
PS - I mentioned Denver. For anyone who lives in the Denver area and is interested in iListen, I will be at the Aspen Grove Apple store 5/10 at 2PM, the Park Meadow store 5/10 at 7PM, and the Cherry Creek store 5/11 at 3PM.
Also, the microphone settings appear to have automatically put mic output at 100%, so I can't turn that up. Would playing around with the silence detection make a difference? Right now it's at 64% (again the automatic setting).
I suspect it may just be because of the way I speak. I'm a fairly soft-spoken kinda guy, but I'm afraid that trying to speak up would be unnatural and would actually lower accuracy rather than increase it. I suppose I could retrain iListen for my "speaking up voice", but I doubt that that voice would be consistent from one session to the next.
Ideas?
Cheers,
Grey
I'm pretty much at the limit of what I can provide, and individual support probably isn't in the best interest of others who follow this thread anyway.
What I would suggest is that you contact our support department by going to www.macspeech.com/support and clicking the Submit Ticket link. Let the support team know what Mac you are using, how much RAM it has, what version of Mac OS X you are using, and what microphone you are using, in addition to the other information you have provided here and see what they have to say about it.
I can tell you that whether you are soft-spoken or not, in and of itself, has nothing to do with accuracy. Our CEO is very soft-spoken and I am not - we both get exceptional accuracy from a well-trained profile.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Cheers,
Grey
Must I buy both iListen and the approved MacMice MicFlex USB Desktop microphone from MacSpeech? I've read various online comments that unless both s/w and mic are purchased from MacSpeech, there's no tech support. I am on a limited budget and found it less expensive elsewhere.
Can I use an USB extension cord so I don't have to sit directly in front of my Mac? If so, is that limited to 10 feet?
Lastly, when I have enough money, will the Desktop Array allow me to walk around while dictating? I hate sitting still while talking.
Or is the better alternative to purchase either the Olympus recorder or Video iPod/iPod nano and the mic adapter and also the transcription pack, whether using Olympus or iPod?
Ideally I don't want to be glued to computer. I am a very fast typist. But I am more inspired to tell stories when not in front of the computer and either walking, or stretching my legs out, away from computer.
Thanks in advance.
I'll try my best to answer your questions.
For best results I strongly recommend you use a head-worn microphone. You can wear the headband around your neck instead of over the top of your head and still have the microphone in the best position for speech recognition.
My preference for a MacSpeech certified head-worn microphone is purely scientific. A noise-canceling microphone whose position is consistent in relation to your mouth will give you better accuracy. In my opinion, much better (on the order of 5%, which is 5 fewer words per 100 that are wrong) over the MicFlex. It is simple physics, really.
MacSpeech has never had a requirement you purchase directly from us in order to receive support, and never will. I believe someone must be confusing our right to refuse to offer support if you are not using a microphone certified by us. We don't care where you buy it (although naturally, we appreciate the business), as long as what you buy is certified by us.
For best results using the Voice Tracker Array microphone (which is what I assume you were referring to) you should sit more or less with it directly in front of you. While it will follow you to some extent, you need to be 12-18 inches from the array in order to achieve the best results. Not something you can do while walking around.
If you are keen on walking around, the very best option would be to use a TalkPro Xpress microphone with either a Video iPod or a second generation iPod Nano. As mentioned before, you could wear the headset around your neck instead of your head. Yes, you can add a USB extension cord, but yes, you are limited to 10 feet. This is a USB limitation that has nothing to do with iListen.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thanks for your prompt response! I originally wanted to purchase:
iListen s/w only and
Olympus recorder (lower end model) and
transcription pak
However, your website states it's best to dictate directly into the program. Urgh, I want to get away from my computer. I am very willing to put in the hours needed for training.
I have a strong, clear, steady voice. My biggest concern is if, after putting in the time to train it, what if it just doesn't perform that well cos it's from a wav file, not direct input? I can't afford to eat a $300 plus investment.
Thanks for your feedback. It's hard to buy something and not try it first. I've been burnt for so many years. I know your company is working hard and stands behind its products, but I'm asking about something that is not the preferred recommendation for using your product.
I have an Olympus WS-100 digital recorder, but Olympus website doesn't state all the specs and I have a feeling it won't work. I've hardly used it, but purchased before my current Intel Mac. I hate buying another recorder, and want to be sure this will work.
thanks!
We don't provide refunds for lack of accuracy, as we now have thousands of customers and know for a fact iListen is very accurate. We realize that is not everyone's experience, but we are willing to work with anyone as much as necessary to improve their experience. We also understand that it takes longer for some people to train the software from others, and that some of those who take longer for the software to learn their voice do not have the patience. Most people are getting accuracy in the high 90's after anywhere from 2 days to 2 weeks, depending on many variables, not the least of which is how much they are using the software.
Regarding accuracy, if you use an iPod with a MacSpeech certified head-worn microphone, your accuracy will be as good as live dictation. I have used it that way with my iPod and am still amazed at the accuracy I get.
The WS-100 is not supported because it does not come with software that can copy the recordings to the Mac in a format iListen can use. It also will not be anywhere near as accurate as an iPod.
We'd love to offer a trial version, and in fact, when we first came out with iListen, we did. The problem is the old "Garbage In-Garbage Out" problem. If you don't have the right kind of microphone and have it positioned the right way, the software gets garbage and can be very inaccurate. It's hard enough to counter the bad reviews from those who do not seek help from our support staff, so for obvious reasons, we do not want the software blamed for doing the best job it can under adverse circumstances. Better to just avoid the adverse circumstances.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
When money situation improves and I have more compelling reasons for an iPod, then I'll add the transcription pak, iPod, dock and iPod mic adapter. But at least I've taken the first step!
Final thought: I probably will puchase for download some scripts after I train for mere dictation. I spend so much time in email/word processing that it could be a valuable investment.
Thanks to all of you w/your comments, again, esp Chuck.
Have fun talking to one's computer and have it respond!
I want to use iListen on a Quad G5 (PPC) and an macbook (intel). Do I need to train the software separately on each computer? Or can I copy the training results from one computer to the other?
The honest answer is "it depends."
iListen is Universal Binary, so it will run natively on both machines. You can also copy your profile back and forth from one computer to another. I would recommend you keep a backup of the profile you started with before you move it to another computer.
In theory, you should be able to make changes to your profile on either computer and thus continue to improve accuracy. There is a possibility, however, that the different processors may impact the way sound is processed, and therefore, you may notice a degradation of accuracy after the second move. Here's what I mean:
Suppose you create a profile on your Intel Mac, first back it up, then move it to the G5. So then you make changes to that profile on the G5 and move it back to the Intel Mac. If you notice accuracy is worse than it was originally, then something about the changes you made on the G5 adversely affected your profile. No problem. Just replace it with the backup.
BTW, you may not notice a significant difference right away. It may take weeks or months, simply because iListen is continually learning your voice. So don't let that first backup be your only backup. Backup often, and keep multiple copies. (We can't wait for the Time Machine feature in Leopard - it is really going to help with this!)
Chuck Rogers, Chief EVangelist
MacSpeech, Inc.
For people with mid-Atlantic accents would you recommend the US or the UK version of your product? I tried Dragon NS once, the US version, and was terribly disappointed with its lack of accuracy. That was all that was sold in Canada, where I live. My accent hovers somewhere between UK and Canadian English.
Also, if I wanted to buy just one microphone that I can use with 1. your product 2. Skype and 3. dictating to a live human transcriptionist in case all else fails, what can you recommend as the best?
thanks.
I would have no way to determine which language model would be better for you. My suggestion would be to purchase the North American English (US) version and if that does not work well for you, contact our support team and they will arrange to swap out the US version for the UK version for you to try.
Regarding a microphone, all of the microphones we sell not only work well for all three of the uses you describe, but will actually work better with Skype, since (unlike the microphones Skype sells) they will actually provide clearer transmission of your voice to the person on the other end of the conversation.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
OK I'm convinced. How would I get iListen with the VXi Talkpro Express in the UK? Al the sites I've visited appear to only sell with the Plantronics version. Can I order the UK version from the States? I have an iMac Intel with 3GB of ram and latest version of Tiger.
Regards
Graham.
Unfortunately the TalkPro Xpress is not RoHS certified, which is a requirement for shipping to the UK. I would recommend either the Plantronics .Audio 85 or the Plantronics .Audio 510 instead.
The 510 tested as good as the Xpress, and the 85 almost as good.
While you can order a version from the US, the shipping will cost you about half as much as the product itself due to the customs and brokerage fees.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Great review and feedback. Chuck I am really impressed with your honest responses.
I just switched to a Mac Book from a PC about two months ago and I am now considering purchasing iListen. From what I am gathering the core difference between iListen and Dragon is one is for a Mac and the other a PC. Is that an accurate assessment?
I am considering purchasing the software for three purposes.
1) For dictating instead of writing when working on a book and articles. Your software seems to be fairly straight forward and more than capable to handle this request.
2) I am also looking at using iListen for dictating speeches/presentations I give as well and pairing it with my iPod Video to complete this task. I'm concerned that iListen may not be as beneficial for this task. Thoughts or suggestions?
3) Finally, I am planning using it for meeting dictation. I participate in several committee's where I plan to use my iPod as a mini recorder. Then I would like to turn those recording's into notes/minutes. Based on previous comments this may not be a good task for iListen. Am I correct? And any tips on how I could use software to handle this task?
I have noted you are coming to Portland this week and I will try and connect with you at the Mac store in Pioneer Place Mall. However, I ahve a very tight schedule and I may not find the availability so if you could respond I would appreciate it.
Thank you!
Brian
First, thanks for the compliment. I hope you can make it to one of my presentations in Portland. Both are on this coming Wednesday (7/25/07).
Let's see if I can have a go at your questions...
While we do have former Dragon users who will swear iListen is faster, more accurate, and easier to use. The opposite is true as well. Usually, it comes down to two things: what people are used to, and (in some cases) how well their voice works with one program versus the other. Sometimes we encounter a person for whom iListen works better than Dragon, or vice-versa.
In general, however, the biggest difference is in training time. It will take most people more time to train iListen to achieve the same accuracy as Dragon. This isn't gospel, of course - and there is no way to tell in advance whether or not you are one of those people who naturally seem to work better with one over the other. But the biggest factor is how much you use the software. A person who uses either Dragon or iListen 5 or 6 hours a day 5 days a week will achieve much greater accuracy faster than someone who only uses it an hour or two per week.
Regarding your second question, iListen requires you speek your punctuation. Other than that requirement, it will do a great job.
Regarding your third question, no software is going to do a very good job transcribing speech from multiple people who are not speaking their punctuation. Your best option here (and also for the instance of transcribing speeches and presentations) is to listen to the recording and re-speak it in your own voice, adding punctuation as appropriate.
Let me know if you have any further questions.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Anyhow I am so impressed that I am going to take advantage of Amazon's excellent 30 day return offer and send it back. Parallels and Dragon I guess.
Have you contacted our tech support department? Speech recognition is the most complex thing you will ever ask your computer to do, and it is also much more subject to influences from other influences than anything else.
In general, you probably won't get any better results from Dragon under Parallels if there is a problem with the sound coming into your computer - something our support team can help you figure out. Also, iListen is the only speech recognition prorgam that will not only allow you to dictate into any Macintosh application, but also any Windows applications that is running under Parallels.
You can contact our support team by going to www.macspeech.com/support and click the "Submit Ticket" link.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
There appears to be a fundamental weakness should the software confuse two words and this is a source of real frustration. It is good with technical words because they are so dissimilar it nails them every time.
It has limited use and frankly it is far quicker to type. There needs to be very large increases in speech recognition if this is to replace typing.
I certainly would not buy this product again and will wait at least 3 years before venturing any cash again.
Thanks for your feedback. I can tell you your experience is not typical of our average customer. We have thousands of customers who depend on iListen every day.
Did you contact our technical support team with your issues? They can give you detailed instruction to resolve just about any issue you are having. If you have not already done so, please do. I can promise you one thing: once you have a fully trained profile, it is far faster to dictate than type. I know this to be a fact - I demonstrate iListen all over the country in Apple stores. There is no way I would subject myself to the potential ridicule of customers and Apple store employees if the software did not work, and work extremely well.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I am sorry you did not have an excellent experience with iListen. All I can tell you is that although it does take longer to train iListen over Dragon Naturally Speaking, we have many customers who insist that with enough training, iListen will be at least as accurate, if not more so.
You can read testimonials from our customers at http://www.macspeech.com/testimonials. I am sure our support team would be happy to work with you if you would like. Only giving iListen 1 week doesn't really provide enough opportunity for it to become as accurate as Dragon.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thanks for your comment.
What you might want to try is the following:
Restart your computer, then launch iListen and turn on its microphone. Say "One Shot Command" and then "Open TextEdit." Once TextEdit is open, try something for me.
Close your eyes and think of something to say, including the punctuation. Once you have ONE sentence in your head, keep your eyes closed and say it, including the punctuation. For this exercise, make sure you aren't using any proper names or technical terms iListen is unlikely to know - all we want to know is how accurate it is with words it DOES know.
Once you have dictated one sentence, keep your eyes closed and think of another. Dictate that, with the punctuation. Continue doing this for three or four more sentences, then open your eyes and take a look at the text - I'll bet it was a lot more accurate than any other text you have been dictating.
If you didn't get at least 95% accuracy, send our support team another note with the text you dictated, both corrected and uncorrected. But if you DID get 95% or better accuracy, then it demonstrates that iListen is doing its job exactly the way it was designed to.
Dictation is a skill, just like typing. Many people blame the software's lack of accuracy on their own inability to dictate. This would be like blaming a keyboard for one's inability to type. All I am trying to do here is eliminate a variable or two and see if perhaps iListen is actually working OK and you just need some practice dictating.
The only other thing I can think of would be if you happen to be using the incorrect language-version of iListen. That happens more often than we would like. The US and UK dialects are as different to iListen as Spanish is from German. So if you happen to be using the wrong language-version of iListen, that would certainly account for poor accuracy.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I appreciate your indulging me. What I have found is that people who have used Dragon are used to using speech recognition in specific ways that, for some reason or another, seem to cause them to have more issues from those who have never used Dragon.
I can't put my finger on anything specific, but I had one person who switched, had a tremendous amount of difficulty, and now claims to get better accuracy from iListen than he did from Dragon, tell me the following: He said that the problems were a lot like someone who has driven a Ford or Lincoln his entire life suddenly switching to a Chevy or Cadillac. Initially, this person is probably going to regret the purchase and might even feel that the Chevy or Cadillac is inferior in some ways, but eventually, as he gets used to the differences, this feeling goes away.
Basically, people who are used to Dragon have expectations that, in some ways aren't met at all, and in other ways are just met more slowly (like the time it takes to achieve the same level of accuracy). In addition, iListen offers several things that Dragon NaturallySpeaking does not, so these things may help balance things out for some people, while, for others, they may not matter at all.
So trying my exercise may help expose the potential accuracy you can get with iListen by removing some of the ways you approach the task of speech recognition as a result of using Dragon so successfully. In general, the more successful a person has been with Dragon, the longer it takes to get used to iListen, including achieving comparable accuracy - for the same reason a long time Ford or Lincoln owner will take more time to adjust to a new Chevy or Lincoln.
In other words, the fact that you "do this all day long" in Dragon could be the very reason you are having so much trouble with iListen. (Another analogy would be the time it takes to get use to the way Macs work compared to Windows computers.)
As an FYI, I typically get around 98% accuracy - and I use it far more for Commands (where I actually get 100% accuracy) over dictation. Believe me, if I didn't know without a shadow of a doubt iListen worked and worked extremely well, there is no way I would be traveling around the country demonstrating it to people 2-3 weeks out of every month. Indeed, if it didn't work extremely well, there is no way I even *could* demonstrate it successfully to so many people.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
For the record, I drive an Alfa Romeo.
Anyway, thanks for your thoughts.
Colin
I take your point. It could very well be you have one of those very rare voices that seems to get better results with one speech recognition system over another.
FWIW, MacLife said they achieved "Near Perfect Accuracy" in their review of iListen for their June 2007 issue.
Chuck Rogers, Chief Evangelist
MacSpeech. Inc.
Our podcast is Windows 2 Apples at
http://inetsynch.podbean.com
This is lifted from episode 13:
"I am far too spoiled by DragonDictate to continue investing more time in "training" iListen. After all, the purpose of speech to text software is to enhance productivity and I feel an additional investment of time would be wasted. I wish MacSpeech well in refining their software but after experimenting with the speech to text engine bundled with Vista I now realize I have new options."
The Vista speech recognition is so good I plan to " purchase Parallels or Fusion and install Vista on the Macintosh. This is a particularly attractive option now that I no longer feel the need to install Dragon as well."
Cheers
Sam
I think the key word here is "spoiled." For most Mac users, the additional investment time to establish a good profile - even if it were 10 times that of Dragon - would pale in comparison to the amount of time they would actually be using the software productively. Also, for most people, including many "switchers," the limitations of using Windows in Parallels or Fusion (not being able to dictate into or control Mac apps, slower response in Windows, etc.) would outweigh the disadvantage of the additional training time required for iListen.
In short, if you (or anyone else) reading my comments are happy using Dragon (or anything else) with Parallels or VM Fusion - that's great. But I would still say that for most people the benefits of just dictating anywhere - including any Windows program you are using in Parallels or VM Fusion - would far outweigh the additional training time iListen requires.
Chuck Rogers, Chief Evangelist
MacSpeech. Inc.
As you know I spent a great deal of time "training" iListen and was informed by you that MacSpeech has found "excessive" training to be counter productive in some situations. Indeed, the accuracy of iListen seems to degrade as I begin to make corrections. And as mentioned elsewhere in this thread, the placement of the corrected word can be unpredictable. Often after trying to correct words, iListen would begin adding words at the end of my next phrase!
I agree with you re. the limitations of kludging a fix with fusion and Vista but, until there is an option that is at least a close equivalent to Vista and Dragon, it is the only way to get a Mac speech to text tool that works well enough to make the investment in time worth while. Even though the idea of polluting our Mac with Vista gives me pause, I will make that investment because the software is truly revolutionary. I have been able to dictate in a much more natural way using Vista than I have even with Dragon.
By the way, I did try iListen once again last night after working with Vista and was again confronted with the clunky interface, quirky correction scheme and unusable output. I really wanted it to work for me but life is far too short.
After reading this and several other threads, it seems your role is more chief apologist than chief evangelist. I do hope MacSpeech is able to continue development with iListen and bring it up to par with your competition but I suspect that may be very difficult. Microsoft now has set a bar that is even higher than that held by Dragon and folks like Ray Kurzweil are moving the technology forward at at exponential pace. His team is now demonstrating Star Trek like speech translation systems that will finally break down the barriers imposed by our multi-lingual world.
Cheers
Sam
Windows 2 Apples
I certainly am not an apologist for MacSpeech. If iListen did not work, and work extremely well, I wouldn't be associated with the company.
People are always free to make their own decisions. I give people clear, honest options: work with iListen to improve accuracy over time, or use something else. No apologies necessary.
Our testimonials - including many "switchers" - speak for themselves. Also, to be clear, I have dictated this text with only one mistake, proving exemplary accuracy is indeed achievable. In regards to your comment that our interface is "clunky," I don't know how to respond, except to say that a). we have certain limitations imposed upon us by the operating system that require certain decisions on how the program displays its options; and b). we are working hard on making this more attractive in a future version. Right now we have our hands full with changes necessary to support Apple's upcoming "Leopard" release.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
PS - the remainder of the text was dictated with one additional mistake.
You continue to harp on training, yet, you have already gone on record with me stating that training can actually degrade performance.
After spending time with the Vista speech engine, I believe speech recognition technology has begun to move at logarithmic speed and MacSpeech will need a significant infusion of money, talent or luck to stay in the game. The potential for serious competition on the native Mac platform will increase if its popularity grows. It would be wonderful if MacSpeech could really deliver a competitive Mac product but I suspect your company does not have the resources to make that happen. Please prove me wrong.
Enough of this, back to work using Dragon and Vista.
P.S. Will post here when my video YouTube reviews are complete ... stay tuned.
I don't want to put words into Chuck's mouth, but as I was reading your blog and some of the other exchanges between you and Chuck, it occurred to me that at least a big part of what Chuck was probably saying is that long individual training sessions can be counterproductive, not so much that more training on the whole is counterproductive (although I suppose there may be some issues to consider there, as well, such as voice changes due to colds or whatever).
So when you've said that you read through almost all of the stories in your first training session, that was almost certainly too much, and your voice probably got tired and, consequently, changed slightly. On another occasion, you also mentioned training for nearly an hour, which again is probably too long. But in further defense of MacSpeech, this issue is mentioned somewhere in the iListen instructions, I believe, or certainly on their web site.
Anyway, I'm curious to see your YouTube video about Vista's voice recognition because all I've seen so far about this hasn't been overly impressive. I'd also be curious to hear more about the speech translation technology you mentioned, because as a professional translator I have to say that I'm very, VERY skeptical that something like that is going to happen anytime soon.
So do keep us posted.
Cheers,
Grey
You misunderstand something very fundamental.
Reading training stories only adapts the profile to recognize your unique way of pronouncing phonemes. For the vast majority of users, doing more than 1 or 2 stories can cause the software to over-compensate, causing the degradation you refer to. Those you have foreign accents or unique vocal characteristics may have to read more training stories (there are a total of 13).
The continued "training" I refer to is to employ Correction which further adapts your profile. It is this "training" that continually adjusts and improves the profile. Use of the built-in Correction interface steadily improves recognition accuracy as you use the software. This "training; has nothing to do with reading training stories. Sorry for the any confusion with this matter.
Our current user base simply doesn't bear out your comment that a "larger percentage" of users are "wasting money and more importantly time trying to make [our] product work for them." The amount of time one needs to spend adapting one's profile (to use a more accurate term from "training"), pales in comparison to the long-term productive use of the software.
Our best example of this is a physician in Gainesville, Florida who is a former Dragon NaturallySpeaking user who switched to iListen and says he was able to achieve 98% accuracy within 2 weeks of installing iListen. He has added over 1800 words to iListen and over 1000 text macros. Dr. Black is one of the reasons I don't need to be an "apologist" for iListen.
Regarding your comment about MacSpeech needing "a significant infusion of money," while that would be nice (wouldn't everyone like that?), MacSpeech has been growing by leaps and bounds, particularly in the last year, and we have thousands of satisfied customers who use iListen every day. In one sense, that already proves you "wrong." But we are constantly working on new, improved versions and will continue to do so.
Sam, I'd like to provide some closure to this thread by once again expressing my regret that you did not have the same positive experience the vast majority of our customers have, and that you feel the additional effort to adapt your profile in iListen is not worth the time and effort over the inconvenience of using a competing product while running Parallels or VM Fusion. MacSpeech is still a small company, but the smart money is on us, having already survived for 10 years in an ever-changing Mac market. Hopefully we can win you over at some point in the future as we continue to improve the product.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Interesting about that over-compensation thing. I'd put iListen on a back burner for a while, but it's about time for me to get back to figuring out the problems I'd been having with it, and I think at least a part of it may have been that I read one or two stories too many.
Guess the first thing I'll do is ditch my current profile and start over.
Thanks!
Grey
I have been using Dragon for many years and,yes, I understood the issues with training and rested when I became tired. I trained Vista in one setting and was amazed at the accuracy. I have just opened a new user on Dragon and will try using it without any training and report back.
This is Mr. Rogers quote re. training:
"believe it or not, the lack of accuracy is probably at least partially due to having read most of the training stories".
As you can see there is no mention of fatigue.
I can assure you, as can David Pogue and others, that Vista is a major advance in speech recognition technologies and I continue to be amazed!
Thanks for visiting our site Drane, please feel free to comment on our content or presentation as well.
Cheers
And Chuck:
As I have said good luck. As long as you are the only game in town when it comes to native Apple products you can stay a float but I suspect that picture may change. The technology is evolving very fast now and my money is on Microsoft and Nuance and ... who knows who will pop up but I suspect they will if Apple grabs more of the market.
Point taken I will not post the link. Thanks for the chance to comment.
Sam
When I move the slider to the greatest accuracy position I began to see 100% of my speech correctly converted to text. This I believe, most would agree it is a very impressive achievement. Perhaps, if the Intel based Macintosh computers grab a larger share of the market nuance will in fact create a version of DragonDictate to run in the native Apple operating system.
This entire reply was dictated with only one error using the newly created profile which has not been trained at all.
That equates to an accuracy of 99.3%.
So let's review:
With no training and DragonDictate's accuracy slider all the way to "Speed" you get about the same level of accuracy as iListen after the 5 minutes it takes to set up your microphone and read the first training story. (Most new iListen users report between 90-95 percent accuracy after reading the first story). This would also be in agreement with the review that appears over these comments.
You slid the DragonDictate accuracy slider to its "greatest" position and you got 99.3% accuracy.
Question: did you try sliding iListen's accuracy slider all the way to "accurate?" You still probably won't get 99.3% accuracy IMMEDIATELY, but your accuracy will improve. Let's make sure we are making fair comparisons here, that's all I am asking.
Sam, I don't know why you have decided to make it a personal mission to dis iListen. You keep saying you wish us luck, etc. And I keep telling you that iListen is capable of the same accuracy if you put forth the effort. You have made it clear you don't want to expend that effort, and that you also think some other company is going to come in and have us for dinner at some point.
OK, we all get that.
If you keep posting, I'll keep replying. That's my job until the people at ATPM put a stop to it. But I don't think we are seeing anything new here other than you don't know iListen as well as you know Dragon Dictate.
So let me once again try to put closure to this for you with as much clarity as possible:
- iListen will not be INITIALLY as accurate as the market leader's product on the PC, but it can become as accurate over time;
- MacSpeech is a much smaller company than Nuance, Microsoft, or Apple. We are also much smaller than IBM and we outlasted them on the Mac.
- Because we are smaller, and Apple sometimes presents more challenges for development over Windows, it can seem we make progress a bit slower. But considering that Dragon was already a well-established company when MacSpeech came into existence, the reality is we have made remarkable progress catching up. We aren't there yet, but we continue to make progress.
So I'll leave you with this (hopefully). Yes Dragon's products are INITIALLY more accurate, and they may have some features that we don't have. We also have features they don't have. If initial accuracy is the thing that is most important to you, use one of Dragon's fine products. If you are more interested in features that integrate well with the Mac and performance on the Mac that can't be beat - even by Dragon in Windows, then choose iListen.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
As I review postings from users of iListen, I see the same frustration that I experienced when trying to use your product, so I'm not alone. I believe if potential customers realize they have the option, even though it's more expensive initially, that it is an option that may be much more economical in terms of their time and productivity. I believe the common thread seen in all of the responses to your comments, even from those who have been very frustrated in trying to use iListen, is they appreciate your company's efforts and wish you well but have to get on with life and their business and can't spend the time required to "train" iListen.
I think it would be great if I listen is able to develop a product that is competitive with DragonDictate and Vista and I hope to see that appear on the market soon. However, I'm fortunate in that I have access to an XP with DragonDictate and Vista speech recognition software so I have options many of your customers don't have.
To your company's credit, it is the frustration I experienced with iListen that drove me to experiment with the Vista speech engine. And now I can share my experiences with others looking for a leading edge speech to text conversion product that can be used with their Apple computers.
Cheers
Sam
I just looked and the accuracy slider has always been all the way to right ... to greatest accuracy!
I love iListen. I have tried Dragon on several occasions with two friends who own it on their pc. It did not work very well for me none of the times. So you see, experiences are different. By the way, we strongly differ in another aspect of appreciation. I find a system like iListen, which gets better and keeps getting better the longer you use it (I just wait till it starts responding me ☺), an intrinsically better system than one that is built to perform immediately at its best. After all, most of us using speech recognition do this for particular and very specific types of texts. For example, I am writing scientific articles, with a lot of specific jargon, and hence many of the words I use are not at all common. A system that learns to use these words instead of the common ones that resemble them has a huge advantage, even if it takes time for the system to do it. My personal opinion is even that speech recognition should best evolve in this direction. Having a system that 'almost only understands you, and is trained to do its best for your specific types of text' will certainly outperform quickly a system that wants to 'work at its best for everybody right from the first time'. Hence I hope that iListen will evolve in this direction, making available a system that as much as possible can adapt to the specific user and its specific use. Hence, to put it sharply, I would like it to differ more from Dragon than it does now.
Best wishes,
Diederik
But each to their own. As you can see from the broad range of machines I use that I use what does the job. Getting the Mac to take my dictation is a challenge and I am learning more each day.
Have A Nice Weekend
Did not even think to look till I saw Chuck makes frequent presentations at Apple stores.
Seems I have many in my boat.
Yes, there are many people who complain about iListen. They are a very small fraction - much less than one percent - of our user base. Far more people have great experiences. Again, I am sorry you decided not to work with it and our support team as necessary to achieve the same excellent accuracy the vast majority of our customers experience.
Sam, for the love of pete, why have you made it your life's mission to go so postal on us? What on earth have we done to you except not meet your expectations - which I have already ceded to you many times over. We haven't lied to you. We haven't misrepresented our product in any way. We haven't told you we are better than Dragon or anyone else. It's OK to not like or want to use iListen. Move along now - nothing to see here.
For the record - if there is anyone left reading this thread that even cares - the vast majority of people who complain about iListen fall into two categories:
- Those whose expectations are not realistic. Try as we might, there are people who think that speech recognition works today like it does in the movies and on Star Trek. It doesn't. Not even Dragon is that good. These people end up disappointed and want to talk about how they wasted money instead of invest a couple of weeks to build up a good profile.
- Those who have legitimate problems and for whatever reason take the time to write a bad review instead of using that same time to contact our support team. (To Sam's credit, he DID contact support and he did work with them, but he does not want to spend a couple of weeks just working with his profile to bring it up to speed. Instead, he dabbles with iListen and then talks about how much better PC programs are.)
In the interest of fairness, I urge everyone who gives a hoot to also check out our testimonials and the many positive reviews for iListen, which can be seen at the following links:
http://www.macspeech.com/testimonials
http://www.macspeech.com/reviews
As you will see, the boat with people who like and use iListen is far bigger than the one with people who don't.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Will let the Apple Store reviews speak for me and others. Don't fret I will let go of iListen after my video reviews.
Having far too much fun creating the videos at the moment to do much more posting but think the iListen,Vista and Dragon video projects will be fun and a bit challenging to produce.
Have a good one!
Good job MacSpeech!
another major difference is that Nuance/NaturallySpeaking staff as a matter of policy do not contribute to any forums - they rely on an ex-employee to provide (unofficial?) feedback. We keep getting assured though they do read the forums.
What I find frustrating is that, unless I win the lottery, I can't afford to purchase comparable hardware and software to compare both ilisten and NaturallySpeaking.
cheers,
Allan H.
In conclusion, and I do mean in conclusion, I am transitioning to iListen VERY successfully. The recommended Parrot microphone made all the difference in the world for my voice; I think different mics work better for different voices from my years of DNS and iListen experience, so you may want to check into that issue if you have problems.
I would have dictated this, but I am at my daughter's house and she doesn't have iListen, yet, but she will after we go to the Apple store for her Birthday shopping today - even if it's only for me to have it when I visit.
Michael
Let's do some math. Assume I'm an attorney (and I am not), charging a minimum of $200 an hour and I spent three hours setting up and learning to use DragonDictate. I believe most after those three hours, if using version 9, would get accuracies close to 99% even with no training. The attorney using DragonDictate has invested $600 of his time plus the cost of DragonDictate which I believe sells for around $100 now. He becomes considerably more productive than he was before. Total: = $700 + $70(for headset) = $770
Now let's take the attorney using iListen and give him three hours to set it up and learn to use it, and at least one hour a day for 13 more days training to get close to the accuracy experienced by my first attorney using DragonDictate. That would give us $600 to set up and learn plus one hour a day at $200 per hour for 13 more days to get iListen up to speed ($2,600). That attorney has spent the cost of iListen plus $2,600 in his time to train iListen. Let's add now: $600 (setup) +$179 (for iListen) + $2,600 (training) = $3,379.
I believe the Dragon customer would be ahead of the game even if he purchased a program like parallels and a new boxed copy of Vista business and had it installed by someone from Best Buy or Circuit City
Lets say $470 for Parallels and Vista and $300 for set up which adds another $770 to our $770 yielding a total of $1,540 for Dragon verses $3,379 for IListen.
I suspect Chuck would argue that the Dragon customer will not have the same flexibility of entering his data directly into his Macintosh programs and that's true. That could be a make or break, but, I suspect for very few. My hypothetical Dragon customer spends less time getting his product to work and is enjoying enhanced productivity the many days and hours it takes the iListen attorney to see those kinds of gains.
At any rate, I believe his argument does not hold water. For me I know that was certainly the case. I invested many hours, well over one week "training", iListen and many hours trying to correct it with the quirky and frequently unpredictable corrections window.
I will credit MacSpeech, as I have in my blog, for offering to help me and add that their response time was almost immediate when I had questions or issues. However, time is money and your product was costing me a great deal of that commodity.
The MacSupport team was very responsive to my many requests for help. Thankfully, since DragonDictate just works and the interface is so intuitive and program easy to use, I have never had the chance to test their response time nor to be charged in for help.
VMware Fusion 1.0 may be a better choice but I have not seen postings from anyone using it with Dragon. My plan is to use the excellent speech engine in Vista with the VMware Fusion but will wait till the release of the updated Apple OS to do the install and test.
Cheers
http://wordofmouth.typepad.com/george_silvermans_word_of/womworthy_products/index.html
Got an Intel Mac (MacBook Pro) in May, tried Crossover and Parallels, which didn't work properly with Dragon NaturallySpeaking. However, VMWare Fusion works like a charm with Windows XP and Dragon. Parallels had all sorts of quirks with the USB headset and froze up constantly. Crossover wouldn't work at all. With VMWare Fusion, it can now see my Plantronics Audio 500 headset perfectly, and the accuracy is well over 99% and getting better all the time.
It even works perfectly with my Olympus D-30 recorder, even in noisy environments like a car. I'm about to try it in an airplane. I can dictate on my daily walk, into a tiny recorder and have a new section of my new book, a blog post, article or whatever a few minutes after I get back.
I also believe it is possible to do a scientific comparison of the accuracy of various technologies and algorithms used to process speech across a wide spectrum of accents and styles of speech. Worth doing some searches to see if such studies have been done.
Cheers
As a professional who has been using iListen in a challenging work environment for 10 years, and used Power Secretary before that, and have fully setup, trained and reviewed ViaVoice and Dragon, I know iListen has no equal.
Without the ability to dictate directly into MacOS X native applications, what's the point? Copy-and-paste is decades-old now. For the EMR forms with dozens of fields that I have to deal with every day, Dragon is totally unusable.
This is not, as Caldwell states, a problem for the "very few". Just Google "EMR" and see what comes up. And why would anyone familiar with MacOS X want to lower their expectations by burying their head in a Windows tar pit?
Caldwell is very prolific, having composed the last 3 consecutive posts. He's talking to himself.
Glad iListen works so well for you. There was no need to Google EMR ... have two physicians in the family who often express their opinion of the crazy patch work of paper work they have to deal with. And looks like to me I am stimulating excellent feedback from folks such as yourself.
I am not a fanboy for any OS including Windows ... see my Windows 2 Apples blog for confirmation of that.
This has been a learning experience for me and admit I am enjoying the adventure. Because of my experience with iListen I have discovered the Vista speech engine and Fusion. Not a bad use of my time in the scheme of things.
Cheers
As mentioned in another posting I was trained as a psychologist and then made a career change that lead me to start two companies developing physiological monitoring devices. Before that leap, I was director of Rehabilitation Engineering at an institution for disabled children. We designed environmental control aides and communications systems using the comparatively crude tools available at the time.
After experimenting with the Vista speech recognition and "control" software I realized how useful it would be those in need of special control and communication devices and systems. I would recommend you see if you can get an inexpensive Vista machine and experiment.
I had a friend who was hit by a car and lost the use of her arms and legs as a result. She used Dragon in its early days and found it useful but limited. Unfortunately, she died a few years ago but if alive today I would work to get a Vista machine for her as we did in outfitting a van so she could transported to University ... where she eventually got her MSW.
I offer the Vista solution not because I am a Windows "Evangelist" but because in my opinion is the best system available for those needing this kind of technology.
Best Wishes
Sam
The Correction window is a polarizing feature, of that there can be no doubt. We are constrained considerably by what the Mac OS allows us to do, but I can assure you we continue to explore alternatives, and when we find one we think is better than what we have now, we will certainly implement it.
In the meantime, a lot of time can be saved when using iListen by simply understanding why the Correction process is the way it is, and using the options we have built-in to iListen to suit your individual working style. To that end, I recommend you read the following article in our online KnowledgeBase:
Why Is Correction So Slow?
http://www.macspeech.com/support/index.php?action=kb&article=195
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
At this stage, the thread that was begun between Chuck Rogers and Sam Caldwell has moved into the realm that is simply not appropriate for this venue. For at least the time being, we will not be removing any of the comments that have been posted to this point, but any additional discussion on this topic won't be retained.
Both sides have been clear with their viewpoints and all anyone can do is agree to disagree. Moreover, as strictly a Macintosh venue, ATPM is uninterested in hosting discussions on using Windows as a solution—not even Windows that is installed on a Mac using Boot Camp or some form of virtualization.
To restate succinctly, further discussion on Chuck and Sam's thread is officially closed.
Sam
I've just discovered this board after looking up iListen on Google. I've read up to February 2007 but have to get back to work.
My experience with iListen has been dreadful. It doesn't understand my voice, and it takes me an hour and a half to dictate and correct what I could have typed in 15 minutes.
I was working with MacSpeech support - very helpful, very nice, but now the problem is that they have written me off. They say that iListen just doesn't like some voices and mine may be one of them. (I'm American living in the Uk for many years, and my accent is sort of mid-Atlantic. I use the UK version because of the spelling.) The choice I was given was to either soldier on or get a PC and use Dragon.
I was quite upset about this until I realised that what I was being told was "it's not our software that's at fault - it's your voice!" Anyway, I asked our Mac technician if I could run Windows on my Mac and use Dragon but was told that this would mean using Windows applications only.
What I'd like to know is: does anyone know any iListen trainers in the UK, especially in southern England, that I could contact? I don't want to give up on my Mac. I am typing this, not dictating - if I were dictating this I'd be here all night!
Thanks
Helen
Sam
I checked our support system and could find no record of anyone telling you we had given up on you. In fact, the last entry was from our support team that offered a couple of suggestions, as well as stating that some voices work better with some speech recognition systems over others.
I also replied privately with some information more specific to the problem you are experiencing and an offer to help sort things out personally with you
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Just to clarify - it was in a telephone conversation that I was told my voice may be incompatable with iListen.
Thank you,
Mike
The good news is that I have been working with Helen offline. She is getting 93% accuracy after reading just two of the training stories. I should have her up above 95% by the end of the week. It is too soon to tell, but it looks like the problem *may* have been microphone position - which is critical to get good results in any speech recognition program.
Regarding Leopard compatibility, we do not comment on non-released versions of our software. I will tell you this, however: we are working on a Leopard compatible version of iListen and will release it as soon as testing is complete.
Best Regards,
Chuck Rogers, Chief Evangelst
MacSpeech, Inc.
I wonder how many other products will lag to deliver Leopard updates. I will watch and wait before jumping. Consider it an iPhone lesson.
We are essentially ready for Leopard, but we do not release our software until Apple releases their official version. We test against the same version they ship to their customers to avoid any issues.
We will be testing against Leopard starting this Friday evening and will continue testing until we are confident there are no issues with Leopard. Once testing is complete, we will release an update.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
As far as the Parallels/Dragon route, no thanks, I'm doing my best to leave Windows behind, but that is another story.
Michael
Cheers
We are hard at work on a Leopard-compatible upgrade. Best of all, the upgrade will be free to anyone using version 1.7 or later. We will release it as soon as testing and bug fixing is complete, hopefully sometime in November.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
thanks, marlon
austin, tx
We have countless examples from people who will tell you that with a fully trained profile, iListen can be just as accurate as Dragon NaturallySpeaking. We have also been very forthright in telling people that it will take longer to get a fully trained profile in iListen. How much longer depends on a variety of factors, not the least of which is how many hours per week one uses the software, and how consistently one employs Correction so iListen can learn from its mistakes.
Unfortunately, Apple does not provide developers with a place to respond to those who post reviews. In our case this is extremely unfortunate as the simple truth is that if iListen didn't work at all two things would be very, very true right now: 1). No one would *ever* give it a good review, and 2). MacSpeech would not still be in business.
This is also why we insist on professions and locations for all those who provide us with their testimonials to the quality and accuracy of iListen. We want to make sure people understand we are not making these people up.
(You can read the testimonials at http://www.macspeech.com/testimonials.)
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Chuck and everyone else:
You continue to suggest that there are more people happy with your application than unhappy. However, when I look at the math it seems to be quite the opposite. The five star rating on at Apple is one place to start.
When I counted the number of people who rated your product less than three out of five (i.e., a 1 or 2), I get a count of 22. I'm assuming three is neutral and four and five would be considered positive. The number of comments rating your product four or better is a total of 11. So it seems to be at least 2 to 1 who do not find your product the wonderful experience you suggest it is.
You have inspired me to collect and post my findings on all listed iListen reviews.
Please check my math.
The simple fact is that bad news travels faster than good news. When someone has an unpleasant experience with a product, they are more apt to vent, whereas people who are satisfied rarely take time to praise.
iListen DOES have a much higher percentage of people who are dissatisfied with it initially. We have a very high success rate in helping those who will take the time to do what it takes to achieve accuracy over 95%. That process can typically take anywhere from 2 days to 2 weeks.
Most of those who write bad reviews are those, such as yourself, who are unwilling to put forth the effort. The simple fact is, if iListen did not work EXACTLY as I have said it does, we would never get a good review either.
The average rating on Amazon or Apple is mid-range. It is interesting to note that very few people give it 3 starts. They either give it one or two, or they give it 4 or 5. This is exactly what I am talking about: it works. If it didn't, no one would EVER give it a good review.
Please check out the new testimonials we have added at http://www.macspeech.com/testimonials. One just added is from a former Dragon NaturallySpeaking user who is extremely happy with iListen.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thanks Macspeech for you product.
David J. Black M.D.
Would a court hold the software company, be it iListen, Vista or Naturally Speaking, liable if a mistake, perhaps an arguably fatal mistake, was traceable to the translation software? I have dealt with FDA issues for almost 30 years and this is a natural question for me to ask and is totally independent of any issues I may have with iListen. Do you or others have any knowledge about product liability issues when software like this is used in medicine? I believe human medical transcribers have to be certified before they can offer their services.
While this is admittedly only my two cents worth, it's somewhat of an informed opinion since my day job involves journalism/communication/editing.
As far as I'm concerned, it doesn't make any difference what methods were used to input text—be it a keyboard, voice recognition, or telepathy. ;-) When something that is published is that critical, it must be proofread—preferably by multiple people. There's no substitute for proofreading, and that's something that can be done regardless of the method of input.
So, Sam, my take would be, no, the software company would never be liable. Likewise, a defense could not be "I used voice recognition software."
Thanks
Sam
I originally commented back last year (#10). As a litigation attorney and judge for well over twenty years, I can assure you, Sam, that there is NO way that the Software Company could, or would, be found liable under U.S. law.
The case would be summarily dismissed, assuming one could even find any attorney unprofessional enough to take such a case. But then if you are independently wealthy enough to pay a law firm on an hourly basis to take on such a case, I'm sure that they will happily take your money. :-)
Oh, and I still use iListen, by choice, with very good results.
Best regards,
Norm
Could you comment on the use of the newer generation iPod Nano(fat version)/Classic/Touch as digital recording devices for iListen? Which microphones work with which iPod? How exactly do I convert to a file format compatible with iListen? Which iListen software components do I need?
Thanks.
First, you can't use an iPod Touch or iPhone because they do not have the ability to record audio. In order to transcribe from an audio file you need to use either an iPod Nano (2nd or 3rd generation), iPod Video, or iPod Classic.
If you use a second generation iPod Nano, you would need to use the XtremeMac MicroMemo for Nano. If you use a 3rd generation Nano, iPod Video, or iPod Classic you can use either the XtremeMac MicroMemo (NOT the one for Nano 2nd generation), or the Belkin or Griffin microphone attachments. If you are using a noise canceling microphone it doesn't matter which microphone adapter you use. If you opt to use the microphone built into the adapter we recommend the XtremeMac MicroMemo because you will get slightly better accuracy with its boom microphone, and it is ergonomically more comfortable.
In order to transcribe from an audio file you will also need the optional TranscriptionPak installed into iListen. This is available for $75 from our web site. Instructions for transferring files from the iPod to the Mac and converting them using iTunes are included in the TranscriptionPak documentation.
As an FYI, we strongly recommend using an iPod over any other recording device. The sampling rate is much higher in an iPod than the digital recorders we support. A higher sampling rate means more data. More data means better accuracy.
Also, keep in mind that transcription is solely for being able to dictate when it would be inconvenient to do so directly into the computer. It cannot be used to transcribe interviews, lectures, or speeches. In general, you will almost always get better accuracy through live dictation.
Best Regards,
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
... the talkpro parrott that I ordered with Ilisten 1.7.1 is very good .... using it with NaturallySpeaking 9.5 that is! I also have a HP laptop/Vista with NS9.5 .... My macbook/Leopard has iListen. Right out of the box ... NS9 blows ilisten away! even after reading several stories .... the accuracy was poor compared to NS9 out of the box .... I wish nuance would make a version for mac .... I think macspeech is ether delusional, or just not paying attention .... side by side comparisons don't lie! ... and by the way I have been using Vista Voice ... and it is better than iLIsten dictating right out of the box ... if it was to get full features ... it could challenge NS9 because it works so well moving around the op system ... which by the way is pretty important!
Mary had a little lamb, her fleece was white as snow, and every where that Mary went, the lamb was sure to go ... I just dictated that three times with iListen and had to make corrections each time .... it should have gotten better, but it actually got worse ... NS9 and Vista Voice both got it right out of the box .... and yesterday I dictated a two page letter with NS9 with no corrections needed using the plantronics CS50 wireless .... I've got no axe to grind ... I just want it to work! If Macspeech does not catch up very soon .... they are out of the game!
also... I have the headsets that were sent with NS9, a wireless, and a radio shack usb model .... every one of them work well with NS9 and vista voice ... I can tell the parrott (which was a $35 upgrade) is solid ... but the excess wiring to make it USB is kind of stupid!
I'm going to keep playing with iListen ... but when I need to dictate something I'm going to fire up the PC ... I would be embarrassed to put out a product like this.
As I have stated repeatedly, we make no attempt to hide the fact that it takes longer to train iListen than Dragon. For some people, it takes longer than others - while for a small fraction of people, they actually get better results from iListen over Dragon NaturallySpeaking with very little training. It just goes to show you how different voices can get widely disparate results with different speech recognition engines.
Regarding your microphones, keep in mind that Macs do not provide a microphone input jack. The sound input jack on Macs are line in only. That's why you need to use a USB adapter on Macs.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I should have added one other thing: iListen is horrible at poetry. The more you correct, the worse it will get. For a proper test on the ability of the program to learn from its mistakes, use productivity oriented prose. Text from an article of the Living section of your local paper is usually a good place to get that kind of text.
iListen is designed for productivity-oriented use. The vast majority of people who use speech recognition do not use it to recite or create poetry. They mostly use it for emails, reports, etc. While Dragon may be better at poetry, iListen is optimized to do better with everyday text, as poetry breaks common grammar rules that apply to "everyday" usage.
For instance, let's take the "Mary had a little lamb" poem. If you were writing an email to a friend to recount that event, it might go like this:
"My daughter Mary has a pet lamb COMMA whose fleece is white as snow PERIOD That lamb follows Mary everywhere she goes EXCLAMATION MARK"
Try dictating the above, and correcting as necessary. It should get it right after two or three corrections.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I believe you have conceded that iListen is inferior to other solutions such as Naturally speaking on several counts. MacSpeech concedes iListen takes more time to "train". In fact you suggest it will take several weeks to get iListen accuracy to approach that of other products.
You have suggested to your customers that iListen is very sensitive to the type of microphone you use and the positioning on the microphone. This is true of all voice to text systems I have used but I and many others have excellent results using inexpensive headsets and even desk top microphones with the Nuance products.
Now, you concede that iListen is not well suited for poetry and that the Nuance product may be better for those dictating anything other than prose or business documents.
So it does seem that in several important areas you concede that Nuance bests MacSpeech.
The fact that iListen may take weeks to achieve what is possible almost instantaneously with other products is a very telling comment on the state of your technology and your companies ability to offer solutions for Apple computer users.
We concede that Dragon NaturallySpeaking is further along in its development, and that allows many people to have a more positive experience with it sooner than they will with iListen. We also acknowledge that Dragon has been around a lot longer than MacSpeech, and further, that it has not had to contend with the variety of changes to the platform on which it operates to anywhere close to the same extent we have experienced.
Nonetheless, we have made excellent strides. iListen can be as accurate as Dragon NaturallySpeaking once you have a fully trained profile, (and once again, I concede it takes longer to establish a fully trained profile in iListen - although I dispute that it takes "weeks" for the majority of our customers). On the other hand, iListen works in virtually any application and offers far more control over the Mac OS X and applications that run on it than Dragon does over Windows and its applications. Finally, with iListen, you can dictate not only into virtually any Mac application, but also any Windows or Linux application if you are running something like Parallels or VMFusion.
For those committed to using Macintosh computers, iListen is a viable choice.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Would you please stop your evangelizing against iListen here. It's gotten terribly tiresome, and I'm personally going to stop tracking these comments because of it. Which is a shame, because it's been a good source of information about iListen.
~Grey
Sam, consider yourself having been served notice. Do not make a single new reference to a Windows-based dictation application...for any reason. It has no place here.
Do not make us completely shun your ability to post comments. If you want to talk about Macintosh products, then go for it, but if you make a single reference to another Windows-based product, I will personally go back and delete every single comment you previously posted on this page.
Is that clear?
I would not spend my valuable time commenting and sharing my thoughts re. iListen if it were not offered as the best Mac solution for disabled. But MacSpeech sells to this market and I feel there are far better options for those with physical disabilities. My comments are based on my experience in working with disabled and my experience with iListen as well as other Speech to Text products.
I have promised Chuck that I will support and champion iListen if it is improved to the point I can use it in place of the Nuance option. I would love for that to be the case and await a version of iListen I can support.
I will abide by your sanction and continue to read your site postings on all things Apple. Please accept my apologies for offending you and ATPM. It was certainly never my intension. And again thank you for the opportunity to have shared my previous comments.
Take Care
And best wishes for the New Year.
Sam
I have 2 USB mics which work well with Skype etc an A-0374 logitech -
and a Telex H531
Am i really going to have to buy a third mic to use ilisten?
How is ilisten 1.8 working with Leopard?
thanks
At first I was impressed with the accuracy using my older 1.7 profile and was preparing to eat my words on iListen. As long as I never tried to use the iListen correction window it seemed to do well. When I did try to make corrections, 1.8 iListen was better at finding and replacing text but took much longer to do so. As in the older version, I noticed accuracy seemed to decline after correcting text and stray characters appeared each time I instructed iListen to start a new paragraph.
The correction option seems to be the weakest aspect of iListen. I suspect, I could use iListen as long as I only made manual corrections to my text. Unfortunately, this becomes very problematic when dictating long documents. I hope the next iteration includes a revamped correction tool. If this is feature is made to work as well as their recognition engine it could be a useful and welcome tool for Mac users.
The only person I know who had both Leopard and iListen gave up on iListen but I am anxious to see feedback from others since I intend to wait until early next year to upgrade my OS.
iLIsten 1.8 works great on Leopard! Your Telex H831 is one of our certified microphones. The Logictech will not provide you with the best accuracy, however.
As an FYI, one of the reasons ViaVoice no longer works is because they wrote their own sound driver to compensate for the fact that Macs require a higher gain level than Windows computers for sound input. Logitech mics do not provide this additional gain. This is also why users of Skype for Windows often complain that Mac Skype users are harder to hear.
Also, since the release of 1.6 most former ViaVoice users have reported iListen is more accurate than ViaVoice. If you have any problems with accuracy, please contact our support department for assistance.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Today, I decided to try iListen to dictate a short reply to forum entry and opened Word instead of Pages. I had previously tested version 1.8 while using Pages. I noticed major degradation accuracy and assume it is because Word is more of a resource hog. When I went back to Pages accuracy improved. I wonder if others have compared iListen performance across applications.
All speech recognition programs (not just iListen) are very processor intensive. Anything that uses more of the CPU's resources can result in lower accuracy. This will be more noticeable on pre-Intel Macs, but it is still noticeable on Intel Macs, particularly those with only 512MB or 1GB of RAM.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Today, I experimented with microphone positioning, USB interfaces and the pace at which I speak and it's still not clear why there are times that iListen seems to be much more accurate than others. At the risk of making another false assumption, I'm beginning to think that the pace at which you speak has more to do with accuracy that anything else however even this is still unclear.
Perhaps others who are having good success using iListen can offer tips on how to improve accuracy with greater consistency that I've been able to achieve.
thanks--if ilisten worked as well as you, it would dominate. Oddly I tried my mics @ a friend 's house who has v 1.8 on leopard and, the telex, which is H-531 not H-831 did not pass but the logitech did. I think you could sell more if you had a short trial period. I am reluctant to put $100 in based on what i see online
Did you try your Telex with iListen? If not, in what way did the Telex not "pass?"
We don't offer a trial because you would get very, very bad accuracy from the built-in microphone or any other mic whose gain level is not as high as Mac OS X prefers. As we cannot offer microphones for trial, we don't offer trial versions.
iListen works every bit as good as I say it does - but don't take my word for it - there are plenty of others who have commented here who use iListen every day, and you can read other testimonials at http://www.macspeech.com/testimonials.
As I have said before, for some it works extraordinarily well almost immediately, others have to put in quite a bit of work to get acceptable accuracy, but it is achievable.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
If iListen reports the mic is inadequate, try moving it closer to your mouth and run through the procedure again.
I did check our list of certified microphones, and the H531 is not on the list, so it could very well be that microphone doesn't provide the extra gain iLIsten wants. Even though the Logictech passed, you will get better accuracy with a microphone we have certified.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I'm also interested that there's an upgrade out. It would be so nice if MacSpeech would inform customers of these things.
Helen
MacSpeech sends out a notice to all customers who have opted to receive our newsletter whenever we update the program. If you opted out of receiving our newsletter, then you will not receive notice when we update iListen.
You can change your preferences regarding receiving the newsletter in your Account Profile in our online store.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
just checking in for the new year... how's the product working with the new OS? I can't quite sort it out from above...so maybe y'all can preach a bit... what is the "corrections" angle? Can you articulate the position of your critics and resolve?
Working up post xmas funds to get VR software... thanks marlon -austin
add to newsletter: [email protected]
Happy New Year to you as well (and to everyone else)!
Version 1.8 of iListen works great with both Tiger and Leopard, and most users are reporting increased accuracy as well.
Regarding the Correction window, it really is something people either love or hate. What most people don't realize is that there we don't have a lot of choices when it comes to Correction if we want to maintain the ability for people to dictate into any application (which we do).
Here's the thing: Mac OS X isolates each program into its own little compartment within memory. The advantage of this is that if one application crashes, your entire computer doesn't go down with it the way it did with earlier versions of the OS. But the disadvantage is that it limits the ways in which other programs, such as iListen, can communicate with those applications.
One way is through AppleScript - but unfortunately, not all applications support AppleScript, and even when they do, developers are allowed to implement support for it in any way they choose. This means there is no standard way to access a particular word or phrase in a document.
This leaves the input driver for the keyboard as the only viable method of getting text into an application. But this has limitations as well.
iListen's Correction window is very similar to the SpeakPad in ViaVoice. It contains all the text you dictated and allows it to be corrected in a window that iListen can control. This means that within iListen's Correction window, you have the ability to randomly select a word and choose the word it should have been from a list, or type the correct word in. Once you are done correcting all the text, iListen then replaces the text in the original document with the corrected text from the Correction window.
What people want is the ability to just select a word in their original document and correct it there and have iListen learn from the correction - but because of the above mentioned limitations, this is not possible.
In summary, the Correction window works great once you understand how to use it. If you truly don't like it, then my recommendation is to use it as much as possible as you are building up the accuracy in your profile and eventually, you can just correct mis-recognitions manually once your accuracy is high enough that you don't need to bother with the correction window at all.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
For those wanting to simply use iListen to dictate text rather than type, this might be acceptable. However, I have noticed that when using the correction feature, as intended, my accuracy seems to drop rather than improve. Additionally, I believe the quirky correction option makes it virtually unusable for many users unable to access their keyboard or mouse.
Sam's comments are not exactly correct. The Correction window can be entirely manipulated by voice, including the ability to spell new words that iListen would otherwise not recognize. We have many disabled users who use the Correction window in this way on a daily basis.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I do hope MacSpeech will continue to work on this feature and improve it to the extent your resources allow.
What I would like to know is will using Word for Mac be slower and more problematic than using an Apple word processor for dictating the medical reports? I know nothing of Apple software since I have only used a PC with Windows. What do you suggest?
Thanks for your time.
Have a good and safe New Year
Sam
Windows to Apples
If you have questions you think I can help you with please feel free to email me at :
[email protected]
If you have documents from your clients; convert them to simple text and let Ilisten analyze them. It will alter the analysis algorithm to their word order and vocabulary, drastically improving the recognition. Do this in addition to their initial training dictation. Also, I believe you can further train the system to their voice and word use as you overtype their dictation by using the correction window.
Dictation using language recognition is enunciation, volume, and rate dependent. Everything must be very constant to optimize the process. I hope your clients are very committed to this, or it may be faster to just transcribe and skip the language recognition aspect of the process.
Best of luck
Dave Black MD
This should be easy for you. Macro, macro, macro ( for even 2-3 sentence statements) and use similar scenarios that have been run in the "learn my document" feature.
I trained Ilisten in 30 minutes, had the system analyze 3 months of prior documents; and have never used the correction system. The recognition has been that good. I dictate all of my notes for 20-25 patients a day as well as 10-20 letters.
A basic imac with iwork would serve you well.
Dave
I will try this approach on a new profile and see what happens. I do notice that accuracy degrades as I use the correction option.
Since I cannot afford to lose a day of work I will probably use Dragon on a Mac with Boot Camp while I am learning iListen as fast as possible. Apple has made it quite convenient since they have included Boot Camp with Leopard.
Thanks again for your helpful comments. Hope you have a great 2008.
It seems that most of your questions have probably been answered. All I will do is directly answer your original question. iListen allows you to dictate directly into any application - so it doesn't matter what you use. That having been said, Microsoft Word does use more of the system's resources than TextEdit, which comes with every Mac, and this can have a very slight impact on accuracy - particularly on slower Macs. If you are switching to a newer Intel Core 2 Duo you shouldn't notice any difference, however.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Not exactly. If you install Microsoft Office for Windows in the Windows partition you use with Boot Camp, you will only be able to use Office for Windows when you are in Windows. There are 2 reasons for this: 1). Software applications that are written for Windows do not work natively under Leopard, and, 2). when using Boot Camp, the Windows version of Office would be accessible only on the hard drive partition that is available to Windows (this partition is not available to Leopard).
The best way to run Office on a Mac is to buy the Mac version, in which case you would not need to install Windows on your Mac unless you had other Windows applications you needed to run.
Apple has some great information for switchers at the following URL:
http://www.apple.com/support/switch101/
There is also an independent web site with lots of information for switchers:
http://appleswitcher.com/
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thank you again for your time.
No, that is not possible - for a couple of reasons. 1). When you are booted into Windows using Boot Camp, you are, for all intents and purposes, using a Windows machine. You cannot switch over to Leopard without rebooting (at which time, you would no longer have Windows available). 2). Dragon only allows you to dictate into Windows applications. This holds true whether you are using Bootcamp or a virtual machine such as Parallels or VMFusion.
iListen, on the other hand, allows you to dictate into any application in Leopard, Windows, or Linux, providing you are running Windows or Linux in a virtual machine environment such as Parallels or VMFusion while running Leopard or Tiger. (You can't use iListen when you are running Windows under Boot Camp because when you do so, Leopard is not available and therefore iListen cannot run.)
I hope this answers your question.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Sam
Worth reading.
To All
Have a safe New Year.
Sam
For instance, the microphone that ships inside the box works fine with Leopard. But, like any hardware, a certain small percentage of those shipped are going to be defective. The person who complained about the mic not working with Leopard should have contacted our support department, who would have arranged for a replacement immediately.
Regarding the customer who was getting 30% accuracy, that indicates something was very, very wrong with his set up. I have had total strangers set up a new profile inside a very noisy Apple store and have yet to see anyone get less than 85% accuracy. The normal accuracy after reading only one training story is over 90%, btw.
Anyone getting less than 90% accuracy with iListen after reading one training story needs to contact our support department immediately so they can determine what the issue is that is preventing them from obtaining the excellent accuracy the vast majority of our customers achieve.
Simply put, if iListen didn't work at all, we wouldn't still be in business after 10 years, we would never receive a good review, and we certainly would not have been chosen as Macworld UK Magazine's Education Product of the Year for 2007.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I stand by my original statement: all someone has to do is contact our technical support department if they are getting less than 90% accuracy from iListen, and we will get them to where they are achieving 90% or better in very short order. If these people who wrote the bad reviews had contacted us before writing their poor reviews, they would have never written the poor reviews.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Any feedback would be very welcome.
Chris
www.barrybennett.co.uk
MacSpeech does not have a certification program for iListen, as its functionality is pretty straight forward. We will, however, be starting a consultant's program next year.
If I can be of any assistance in raising your confidence in iListen, please feel free to contact me directly at [email protected].
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
First off I was frustrated to discover that the microphone didn't work even though it was one approved for use for iListen. It took took me several hours to learn (after testing it on another computer to be sure there was no physical problem and finally finding my way to the Macspeech website) that the supplied microphone (Plantronics) has an incompatibility with OS10.4, and that what is required is to visit the support part of the MacSpeech Web site where, if you think laterally (because you won't find it otherwise) you will find a "fix" which sorts out the same problem if you have upgraded to OS10.4 and discovered that your microphone no longer works. As soon as that was done, hey presto, the microphone worked.
This miracle was down to my efforts and not due in any way to macSpeech, so this is a definite black mark due to the several wasted hours and lack of professionalism shown by MacSpeech in dealing with a known problem. If the problem could not be avoided in the first place, and one appreciates that voice recognition software is very fiddly, both the problem and the solution should have been made far clearer, to save everyone a lot of aggravation and to protect the reputation of macSpeech.
The second black mark is from something the similar - in the first training text a "." Is spelled out as a "fullstop" whereas in actual practice a "fullstop"must be called a "stop" or a "period" otherwise it is produced as "full.". Clearly, nobody at macSpeech has correlated the two - again what feels like a lack of polish or professionalism.
It is very clear to me that the slower one dictates the better the recognition. Sam Caldwell suggested this in his recent post an my experience strongly confirms this. Dragon uses context sensitive recognition techniques (one can actually see it happening in practice as words in the text get changed while it makes up its mind) which work better if dictation is in longer phrases rather than the short phrases which seemed to be favoured by iListen. This is not necessarily a criticism of iListen and it can actually be easier to dictate like this in any event.
My experience has also been that dictating into TextEdit produces better recognition than is achieved in Word. Given the difference in memory requirements, this is not surprising and is similar to the difference between dictating in Word for Windows and Notepad or the Dragon Pad supplied with Dragon.
My biggest complaint, and which seems to be in common with most of the other complaints made, relates to the correction routines. Because speech recognition will never be 100 percent, it is extremely important to have fast, easy and precise correction mechanism.
The lack of ability to select words randomly is a real weakness. "Do select" is very hit and miss as to whether it finds the correct place (or simply inserts the words "Do Select ..." into the text). Remember that mostly it will be used to find words which have not been correctly recognised in the first place so having to look for that word will always be 50/50. There doesn't seem to be a way to stop it searching so if it goes past the word to be corrected, it will possibly go all the way to the beginning before it stops. A real nuisance in a long document!
I also echo one of the earlier comments that on occasions changes made in the correction window seem to come up in random places in the text itself. Chuck has tried to explain why and no doubt there is the reason for it but it is not helpful to blame the customer, for example, for having moved the cursor or having made a manual correction - very often correcting an error can be quicker and easier than having to call up the correction window, That seems to upset the whole system as making any other changes via the correction window becomes dangerous as they might appear in the wrong place.
Having read Chuck's explanation, my suggestion for the next version would be that iListen should come with its own notepad (as Dragon does) which, if I understand the explanation correctly, will allow random changes etc.,. That notepad could also be the correction window (another way of saying this is that one should be able to dictate to directly into the correction window). One could dictate directly, select and correct words at random and then copy and paste the corrected to text into other applications without the present inefficiencies of the correction window. to meet, this would be a lot more acceptable than the somewhat unpredictable present system.
If iListen can sort out its correction procedures, this will be a much better product. (I see this is a really long post - apologies, but this is a serious product which I hope will be succesful and I trust that these comments are of value)
The review I referred Michael to was written by someone with a great deal of experience with speech to text software. I to have had many years of experience with speech to text technologies and I'm not disposed to write an iListen testimonial.
Thanks for taking the time to provide all those details.
First, regarding the Plantronics headphone issue, this was due to a bug in Mac OS X version 10.4.10. Upgrading to 10.4.11 should fix the problem. There was no "fix" that either MacSpeech or Plantronics could employ to address this issue, since it was due to a bug in the operating system itself, although we did provide instructions for a workaround, which you found.
More importantly, however, is that you can always get an immediate response from our support team simply by emailing MacSpeech at [email protected]. The next time you encounter any problem at all, please do this, or enter a support ticket directly on our support site. Doing so will save you many hours of frustration.
Regarding Correction, I continue to be mystified by those who insist having a notepad would be a better solution. It would not, it would be exactly the same as having the Correction window with one detrimental exception - unlike the current Correction window, you would have to cut and paste whatever you dictated into the notepad into whatever document you wanted the text to appear. This seems to me to totally eliminate any advantage of a notepad.
Here's what I mean: with a notepad feature, you would dictate into the notepad, correct mis-recognitions, then you would have to cut and paste the text from the notepad to wherever you wanted the text to appear.
The way it works now, you dictate wherever you want the text to appear, then you call up the Correction window. Within that window, you can randomly select words or phrases - exactly as you would in a notepad - but instead of having to type, you can simply press a button to choose the correct word or phrase, which usually appears in the alternates list. (If it doesn't, you still have the option of typing.) When you are finished, the corrected text goes directly into the document you want it in - no cut and paste is necessary.
With that having been said, however, we are not deaf to the desires of our customers. At some point we will probably add a notepad simply because so many people have requested it. I maintain it won't make things easier, but as Tom Wolfe says "perception is reality."
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I believe everyone here understands you don't like iListen, and that's fine. My point remains that iListen works and works well. We could not stay in business if it did not, and prestigious publications such as Macworld UK certainly wouldn't be giving it awards if it did not work.
Virtually all the Mac trade magazines have given iListen very positive reviews. If someone gets less than 90% accuracy, there is something wrong with their setup (in almost every case). Most people get 98% or better accuracy with a fully trained profile.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Chuck, its not that I don't like iListen, I still have hope MacSpeech will succeed in improving iListen to the point I can use and recommend it. But as of January 1, 2008 it is not a product I can use or recommend.
I would gladly sacrifice the ability to dictate into every application in favour of copying and pasting from a pad if it was easy for me to correct errors manually or by voice. Any recognition software will have errors (you will never get 100%) so easy correction is paramount. What you overlook is that calling up the correction window requires a speech command (which itself is not always recognised and as often as not simply inserts the words"correct that" into the text), and then work within a window which does not itself work intuitively or accurately, instead of simply being able to delete the odd mistake, "a" or "and" which sometimes creeps into the text simply from breathing.
I dictated the above paragraph of this post with iListen and then called up the correction window. It took me longer to make the dozen or so corrections than it had dictate it. When I switched back to the text, I discovered that a word inserted via spellng mode (in this case, "correct that" which was otherwise recognised as a command) doesn't seem to find its way into the the correction window at all. The result of that is that if you have used spelling mode (which is useful,and much better than the Dragon equivalent), the positioning of corrections will always be wrong and the corrections will increase the number of errors in your text. What that meant in this case was that when I switched back to the text, it was littered with changes which were in the wrong places, in the middle of words or just parts of words, which then needed a manual tidy up. In the end, I certainly could have typed it quicker even though the recognition is probably around the 90% mark. Not very usable!
That is a very good reason why a pad which you can use to randomly select places for correction and which will not confuse iListen would be so much better. Give me that and accuracy over all applications any time.
I wll persist with iListen and see if it gets any better, but really, a place for simple quick corrections could make iListen a winner.
There is no doubt the existing Correction feature is one of the most polarizing features of the program. People either love it or hate it (and yes, we have tons of people who actually prefer the method we have now).
I don't recall if I mentioned this before, but since the thread is so long, I'm not going to search through it to see if I did. It might help if you understand the history behind the Correction window.
From the get-go, priority number 1 was to come out with a program that allowed you to dictate wherever you type. Anything else was secondary. Any other consideration for a feature was 86'd if it in any way conflicted with the ability to dictate into any application. Accomplishing any sort of Correction at all was so difficult that the first version of iListen (1.0) did not even have a Correction mechanism (we provided a free update to 1.1, which did).
Another consideration was our competition, which at the time was Via Voice from IBM. The first version of their product ONLY allowed dictation into a notepad. No dictation into other applications was possible. Consumers tended to split themselves into one camp or the other: either dictate into any application and put up with our Correction interface, or use ViaVoice and only dictate into their "SpeakPad" and then copy and paste.
From all the data we had available to us, it looked like the market was split pretty much at 50-50 between us and ViaVoice for at least 2 or 3 years. Then, of course, IBM bowed out of the speech recognition market in 2003 and we continued to improve iListen. All the while, we have listened carefully to all the feedback we have received regarding every facet of our program, including the Correction window.
Our biggest issue has been the constant Mac OS X updates Apple has thrown at us over the past 4 years or so. It has made it very difficult to make substantial progress in some of the areas we know could use some work, so we had to pick our battles. Knowing the "Correction" issues would not be easily solved by simple, easy to make changes, we tended to focus on accuracy issues and command functionality.
Now that we have released a Leopard-compatible version of iListen, we can turn our attention back to all the many things we have had on to put on the back burner for the past few years.
With that having been said, I would encourage all of you to pay close attention to the media and our web site later this month. I am not promising anything, nor do I want to set any inappropriate expectations, and no, I won't elaborate further, except to say there is a major Mac trade show this month that will be host to many new announcements from many companies, and in past years MacSpeech has always tended to make some of its bigger announcements of the year at that show.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I suspect that Apple or another company with more resources will eventually deliver a voice conversion and control product that will be better suited for those unable to access the keyboard or mouse.
Will check the Mac site daily in the hope MacSpeech delivers a workable solution for me sooner than I expect from potential competitors.
At first, I was impressed with its accuracy but when I shut my Mac down and again opened iListen it was as if I was using a different voice or had not trained it at all. This experiment was performed in the same day under the same environmental conditions.
The manual correction is simply a no go given there are alternative applications that allow for very easy corrections.
I keep trying Chuck, but still no luck.
I also peeked at the reviews at the Apple Store. The latest batch make my comments seem almost as if compliments. I looked at other products reviews posted there and found few with such strong negative feelings.
I am as frustrated as many of those reviewers . I really do want a speech to text option I can use with my Mac that does not require I install another OS on top of OS X.
I hope MacSpeech can make iListen work well ... as of today they are the only native application in "town".
It is interesting to hear what you have to say and the history etc. I think that the real issue is that, as at present designed, Correction simply does not work - many of the posts in this thread including mine testify to this. Until that is sorted out, either a "Pad" is needed or the program will be sort of limping along.
Sure, it might work for you and for some other people, particularly those who know the package well but for many (including me and I have many years of speech recognition experience) it is counter-intuitive and fiddly, introducing mistakes into the text etc etc. No need to go into that any further.
I really like the idea of a native Mac package and hope that iListen does well. That is why I will persist, at least until I reach some sort of understanding with it, it is improved or I simply feel that I can take it no longer. Having once been a programmer, I know how easy it is to fall in love with, and then stoically defend, something which you wrote (or use all the time) and understand. That's not the point, but ease of use is.
I look forward to the announcements and I also hope to be using iListen for a long time but it is up to Macspeech whether or not I will be.
1) Kudos to Chuck for role of Evangelist -- as stated by another poster, if the product is as good as Chuck, it should be a tremendous success.
2) Kudos to Sam for dedication to individuals with disabilities and some useful comparisons.
3) Kudos to ATPM for stepping in to stop increasingly zealous (and perhaps self-righteous) comments by Sam which appeared as if he had dedicated his being to saving people with disabilities from the scourge of iListen...
4) Kudos to Sam for subsequent self-restraint and re-examination of issues although it still appears like he is on a mission ...
5) Kudos to the rest of the community for (for the most part) not getting sucked in to the fray.
6) For me, one of the most interesting points was raised by Sam when he attempted to quantify the opportunity costs incurred by a hypothetical professional billing at $200/hr ... concluding that the person would be better off purchasing the software that enables Windows to run on the Mac and using Dragon (or even Vista alone). In my case, I am interested in out of pocket costs more than opportunity costs, although my time is not infinite nor without value either. $170 for software vs $3-400 or more is signifcant. What's the old saying, don't let perfection get in the way of good enough?
7) Of course, if Vista's speech to text is as revolutionary as described, it may indeed serve as a tipping point for this niche of the industry and cause issues for the Mac universe as a whole .. which will naturally threaten not only MacSpeech's market position but financial viability. However, this forum is an e-zine about the personal "computing" (not investing) experience. And, in the short term, I need to make a purchase decision in the next few days.
7) There has been very little discussion of the Command Mode, which is one of the key features (in addition to word processing) that I am after. How do the packages compare with regards to working in spreadsheets? in switching between applications? in processing your email inbox? making calendar entries? using a browser? Those aspects are very important to me also, as my greatest discomfort comes from using the mouse, not from using the keyboard ...
Thanks very much.
Well done Mark.
iListen's command mode is the one thing that really does blow the doors off anything else out there. The closest would be Dragon's $900 professional product, but even that does not have all the built-in capabilities of iListen.
1). Say "Open" followed by the name of an application. iListen knows over 150 applications, including all the ones most people use every day, so chances are you can open anything you would use easily. If it is already open, iListen switches you to that application.
2). If there is an application iListen doesn't know, creating a command to open it is as easy as dragging the application's icon into a window and giving the command a name. (The same technique works for all other files and folders as well!)
3). iListen comes with built-in "Global" commands to do things like cut, copy, paste, print, save, quit, etc.
4). We sell "ScriptPaks" on the MacSpeech web site that add anywhere from 50 to hundreds of commands to iListen for specific control of certain applications. For instance, the ScriptPak for Microsoft Word has about 400 commands and allows you to do almost anything with your voice to control Word that you would normally do with your keyboard or mouse. We have actually had people tell us they learned more about how Word works from our ScriptPak than anything else.
5). Any AppleScript can be made speakable. (AppleScript is Apple's built-in English-like Scripting language).
6). You can create your own commands without any knowledge of AppleScript. Creating commands to select menu items or execute keyboard shortcuts are extremely easy.
I hope that answers your questions.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Chuck I suspect you are right about that. I have not even tried the command options, I have been so focused on dictation. Will experiment with that when I get a chance.
Mark, the observation that I am "perhaps self-righteous" pains me a bit but I believe that is how I come across. Not the best way to "Win Friends and Influence People". Painful to see in print but true I suspect. I am working on it ... lets hope it is not too late for an old dog.
Mark, please let us know how you find iListen if you give it a try.
Sam
chuck maybe you might comment regarding the "tone" of the comments at macstore...
sam, you seem cantankerous and well meaning, ...and well informed... i feel if you are approaching this thing as "the loyal opposition" ...and a mac believer; you're good babe, -anything else? well you should probably move on, I think you have made your point(s)... if "negative balance" is needed, the board at mac store will suffice... [email protected]
There are a lot of people out there who either a). think speech recognition should be like Star Trek, or b). aren't willing to put in the additional effort our product requires to fully train a profile. Since they don't get instant gratification, they give up and give it a bad review.
Many reviews really have nothing to do with us at all. For instance, Apple introduced Mac OS X version 10.4.10 with a bug that made certain models of microphones go "deaf." The bug impacted Logitech and Plantronics microphones, among many others. Unfortunately, we had thousands of copies of iListen already in the channel that had Plantronics microphones in the box. We know of several "bad reviews" which were people who complained about the microphone not working, instead of contacting our support staff for assistance.
As I have said so many times before: if you buy iListen and you are getting less than 90% accuracy after reading just the first training story, something is not setup correctly and you need to contact our support department. Please keep in mind that speech recognition software is the one piece of software more susceptible to outside influences than any other application. If the microphone settings are wrong on the computer, or something is wrong with the mic, or it is in the wrong position, or if the user mumbles, iListen is going to get "garbage" and will not be very accurate. All of those things are out of the software's control, although MacSpeech, as a company, can address many of them.
At any rate, in almost every case, a bad review is from someone who did not contact MacSpeech before posting their review. Had they done so, there would be hardly even one bad review out there.
I would encourage anyone who doubts this to look at the many user groups and magazines who have responsibly reviewed iListen. In every case, the review was very positive. These entities have a responsibility to their constituents, not MacSpeech - if iListen didn't work, and work extremely well, they would not report to their other user group members or readers that it does.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I hope those on the fence follow up after making a decision.
As you have noted there is a preponderance of negative iListen reviews posted at the Apple store. I suggest you look for reviews of competing software even if it requires a different OS before you make your final decision.
The cost you incur is not in the software and a Microphone but the time invested in learning and,in the case of speech to text software, time training.
Chuck, you turn comments fast...i hope that's reflected in your customer service department...
I'm a slightly dyslexic linguistics major with a tendency to mumble -what's the best microphone available?
i have just made 5 mistakes typing this... get your boys ready chuck, i sign tomorrow...will post comments on the training process
can you talk a bit about the microsoft word pack?
I am looking forward to being an evangelist.
semper fi -sgt'77
marlon
I'll warn you in advance that no speech recognition software likes mumbling - so the more distinctly you can speak, the better. That doesn't mean you need to sound like a radio announcer - just try not to mumble when dictating.
All of the microphones we sell will give great results. The best one is the VXI TalkPro Xpress, in my opinion.
Not sure what you'd like me to say about the Word ScriptPak. It adds around 400 commands to iListen that allow you to say things instead of choose menu items or press keyboard shortcuts. You can find a list of commands on our web site at http://tinyurl.com/22xqo7.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Any chance you will give us a hint on the new product that is coming out in a few weeks?
All I can tell you is that MacSpeech has always been completely dedicated to producing the best speech recognition software for Macintosh, and there will be announcements at Macworld Expo in keeping with that commitment.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I am looking at buying iListen very shortly. I assume the "new product" is a new version.
Are you able to say what the new product is and when it will be available in Australia?
Joe
Details about the new product will be announced tomorrow at Macworld. That is all I can say at this time.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Helen
The press will be awash with the announcement, and we will also have details up on our web site (www.macspeech.com), so you will have no problems finding out about it.
That having been said, I will be happy to answer any specific questions once the announcement has been made. Just please understand that I am at Macworld all week, so I may not be able to respond immediately.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Helen
We will be announcing shortly after Steve Jobs' keynote, which is scheduled to go until 10:30 AM PST. More importantly, however, anyone with Internet access will all have the same access to the information. You can check out our web site, or any of the Mac news sites.
www.macsurfer.com is an excellent aggregator of all the Macintosh news sites.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I do hope they will offer up a product I can use and support in my postings.
Thanks Chuck,
Chris(Barry Bennetts)
Parallels and VMFusion are great ways to run Windows on Macs, but they really do not present much of a threat to us, as iListen allows you to dictate directly not only into any Macintosh application, but also any Windows application that is running inside of a Parallels or VMFusion virtual environment. Using Dragon inside of one of those environments you a). need an extremely fast machine with a lot of RAM, and b). you cannot dictate into any Mac application (your dictated text must be copied and then transfered to the Mac's clipboard, and then pasted where you want the text.) If you are getting accuracy comparable to Dragon NaturallySpeaking with iListen (which I have all along maintained is entirely possible), then it would sort of defeat the purpose to use a virtual environment for your speech recognition needs.
I am surprised, however, that you aren't seeing a tremendous advantage in your ability to control your Mac with iListen, because outside of any concerns anyone may have over accuracy or correction, that is the one area where we absolutely blow the doors off any other speech recognition product, including Apple's built-in English Speech Recognition.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thanks
Sam
Can I dictate with iListen into Microsoft Word in Windows in Parallels?
Thank you for your time.
YES! You can literally dictate anywhere you would normally type - including any Windows application that is running in a Parallels or VMFusion virtual environment, and that includes Microsoft Word for Windows (although Word 2008 for Mac will absolutely blow that away!).
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Could I copy a Mac Word document I dictated with iListen and then use an online program in Windows in Parallels to paste it to a client's server?
Thank you for your time and timely answers.
There is nothing special about a document dictated with iListen, no matter what program you use. The document itself is no different than if it had been typed. So if you could copy a document you typed with the keyboard to an online server using a program in Windows, then the answer is yes.
If you are really asking me if a Word document created on the Mac could be transferred to a client's server using Windows from within Parallels, the answer is also "yes." But if all you are doing is using FTP to transfer to the server, then just use an FTP program on the Mac and avoid Windows all together.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
not happy with your help desk people they will not come back to me. your support team are i waste of time. May i remind you? below is what you said
"Come check us out. If you purchase the upgrade and our support team a). doesn't respond to you within 24 hours; and b). can't get you above 90% accuracy within 1 week and above 95% within 1 month or less, you can have your money back."
i have had a on going complaint for 4 weeks now.and your team sill do not reply to my tickets,this is not good P.R
i hope you can be of help? if not then ilisten will go in the bin
here's hoping Denis in U.K
This is apparently a Nuance product with Nuance code and Nuance Engineers. If it lives up to the hype ... Great news for those wanting a native speech to text engine.
A quote from a posting on Macsimum :
"Upon distribution of MacSpeech Dictate to retail outlets, the iListen product will be immediately discontinued."
I don't believe many will shed tears as iListen is cremated.
"Upon distribution of MacSpeech Dictate to retail outlets, the iListen product will be immediately discontinued."
I suspect that if the new MacSpeech Dragon is equal to the existing Windows product I will simply switch and trash my iListen software. The alliance with Nuance will hopefully give the Apple community a very high quality product that in Apple Speak "Simply Works."
i wonder what chuck has to say when he surfaces from macworld.
Tuesday January 15, 2:47 pm ET-from yahoo financial.. i bet there is other-
"MacSpeech Dictate Brings State-of-the-Art Speech Recognition, Powered by Nuance, to the Mac
- New Speech Recognition Product, Based on Nuance's Dragon NaturallySpeaking Technology is All-Mac, Highly Accurate and Easy to Use -
SAN FRANCISCO--(BUSINESS WIRE)--MacSpeech, Inc., the leading provider of speech recognition solutions for the Mac, and Nuance Communications (Nasdaq:NUAN - News), developers of Dragon® NaturallySpeaking®, today announced an agreement to bring the world's best speech recognition solution to Mac users worldwide.
LINK:
http://biz.yahoo.com/bw/080115/20080115005894.html?.v=1
I just returned from the show floor. What a phenomenal day. Response to our new product has been nothing short of amazing! For the first time ever, we actually were able to allow people to try our software without first having to do the training required by iListen.
Allow me to address a few of the comments that have been placed here during the day:
- Yes, we are using the Dragon engine from Nuance. We are not, however, making "Dragon for Mac." MacSpeech Dictate is an entirely new product, written from the ground up for Mac users and Mac OS X.
- No, we are not using anyone from Nuance's engineering team, except for development support. The reference to a "new engineering team" is correct in one sense, but not another. Several members of our development team were also on the original iListen development team. They bring invaluable experience to our development process. We also have several new programmers on our team, but they are Mac programmers through and through. So yes, we have a new development team, but several team members have been with us since the beginning.
- For those of you who do not have Intel Macs, yes - you will need to get a new Mac to run MacSpeech Dictate. Given the development tools available to us, making something that would work on both PowerPC and Intel-based Macs would be cost prohibitive and add several months to our development process. Given that Apple is no longer supporting these Macs with their newly introduced applications and that Leopard will be the last release to work on PPC Macs, we could not justify either the development cost or time.
- MacSpeech Dictate will be a more expensive program than iListen. Keep in mind, we have not increased the price of iListen since its introduction over 7 years ago. Yet we have continually updated its feature set. MacSpeech Dictate will have all that is best about iListen, but require much less training to achieve acceptable accuracy. Since it will be a more expensive program, everyone who owns iListen will have to pay something to upgrade to the new program. Those who purchased in 2008 will pay less than those who purchased in 2007. Those who purchased in 2007 will pay less than those who purchased iListen in 2006, etc.
If anyone has any specific questions, feel free to email me at [email protected]. But please keep in mind that my ability to respond is severely constrained by my obligations to Macworld and related events this week. So don't be too upset if I don't respond right away.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I asked our Support Manager about your issue and the lack of response. He said they had a problem with the support database that the developer of the help desk software had to address, which could explain the lack of response. At any rate, if you send me your email address, I will see you get a response within 24 hours of when I receive it. For security reasons, don't send me your email address as a response in the forum. Please send it to me directly at [email protected].
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I am impressed with this move and hope MacSpeech Dragon performs as well as the Windows version. Please let us know if the correction options are similar to those in Naturally Speaking.
Anxious to upgrade and review.
Enjoy the glow!
Sam
I want to be very, very clear about something. This is no more "MacSpeech Dragon" than iListen was "MacSpeech Philips." Licensing a speech engine and making software to interact with it are two very different things. MacSpeech Dictate has very little in common with Dragon NaturallySpeaking other than the speech engine itself. And I can tell you first hand, that while having access to the Dragon Speech Engine is a big deal, it isn't the whole enchilada. This is 100% a Mac product, just like iListen.
So I would ask the favor that you refrain from calling it anything other than by its proper name: MacSpeech Dictate.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
If you are suggesting MacSpeech has retained the iListen model for correcting text, I believe the advantages of the Nuance engine well be negated.
Sorry for the slip ... stand corrected tis Dictate not Dragon and certainly not iListen.
I beleive there is another issue that you need to address. Up until very recently you were defending iListen as the best choice for Apple centric Speech to Text software and now it appears you may drop sales of the product altogether. The quotes from your president focus on the fact that it shares no code with iListen as if trying to distance the new product from iListen.
Given your consistent sales pitch that iListen, was an excellent product every bit as good as the Nuance product, given enough attention by the user, I would hope you give significant discounts if not free upgrades to your faithful customers. Granted we can expect and indeed demand upgrades but yours is a very unique position as the MacSpeeck Chief, Evangelist. You have convinced many to buy iListen possibly knowing the plug would be pulled.
I do hope MacSpeech Dictate incorporates the Nuance corrections options, if not, it will not fare much better than iListen as a useful tool.
Sam
It was sales of iListen that made it possible for us to fund development of a new product. Without that funding, there would be no Macintosh speech recognition product whatsoever.
That having been said, we have made every effort to be as fair as possible to those who own iListen. The new product is more expensive, so everyone needs to pay something to upgrade, and the amount is pro-rated depending on how long you have owned iListen. Someone who has not upgraded to the most recent version of iListen will pay more than someone who has, for instance.
We have had many people stop by the booth telling us how they were getting excellent accuracy from iListen. As I have so often said, it was entirely possible (and still is entirely possible) for one to get every bit as good accuracy from iListen as from Dragon NaturallySpeaking. The primary difference, as I have repeatedly said, was the time it took to train iListen to achieve that level of accuracy.
By switching to the the Dragon engine, we have eliminated that one barrier to achieving superb accuracy in the shortest possible time.
Regarding Correction, I have not inferred anything, nor I do I intend to do so at this time. Right now, our focus is on demonstrating the amazing accuracy one can achieve with MacSpeech Dictate with as little as 5 minutes of training. For many people, a Correction interface of any kind will be useless - that is how accurate this new product is.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
1. Will there be a UK English version? (and Australian, etc) If so, will it *really* deal with UK spellings (eg, realise instead of realize - drives me nuts in iListen!)
2. How much will it cost people in the UK to upgrade?
3. I'm wondering why you're not giving direct answers about the correction facility. Will users still have to switch to a correction window?
4. Will it be possible to edit with speech after corrections have been made, or will everything disappear from memory once we give the equivalent of 'commit corrections'? With DNS it is possible to do this.
5. When will MacSpeech Dictate be available?
Best wishes,
Helen
I look forward to buying and using MacSpeech Dictate. Sounds like a terrific product and will make my job a lot easier.
Well done on developing better software for us to use. I converted to mac 2 years and thanks to software developers like you, I don't see any downside in using macs over PCs - only benefits.
Joe G.
I am well aware of the spectacular accuracy and the significantly reduced need for training when using the Nuance speech technologies. It has been the focus of many of the discussions in this forum. Even though I get near 100 percent accuracy using the two speech to text products detailed in my contributions here and in our Windows 2 Apples podcast, the software would be of limited use to me and to disabled users if the correction scheme was that currently implemented in iListen. It tis a bit strange you have avoided responding to my my question re. correction. I understood the release date to be Feb. 15 and surely your can offer up details on the Dictate correction scheme.
Your comment makes it clear that the conversion of your product line from the Phillips to Nuance engine were made on the backs of customers who purchased an inferior technology. (If not inferior why discontinue iListen?)There are many such as myself that not only plunked down the cash but wasted many hours in trying to get iListen to work.
I do hope MacSpeech rewards those that, as you agree, made it possible for MacSpeech to offer a product competitive with existing technologies.
The dog in iListen's Feedback window belonged to our CEO and was more or less the mascot of MacSpeech. Her name was Lykke (pronounced "Luka"). Lykke died in late 2006 after a long and happy life. At present, we do not have plans to include her in the new product.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Yes, there will be a UK version. In fact, the English version, and it will respect all UK spellings and punctuation rules.
At this point, I do not have any information on what the UK pricing will be.
Regarding correction, I literally cannot comment much on it because that part of the program is not finished yet. I can tell you it will be dramatically different from the mechanism in place now. As I have so often said, we really do listen to our customers.
One of the issues with Correction has to do with the Mac OS itself. There are literally three different ways text gets into a document. How we employ Correction is going to depend on the method the developer used to put text in a window. It is a very daunting task. One thing I can tell you is that MacSpeech Dictate has its own Notepad window. In that window Correction will be a very seamless process.
BTW, one of the things I am demonstrating on the show floor at Macworld is saying "Select" followed by a word or phrase you want to change, and then saying the new word or phrase. It works flawlessly.
Version 1.0 of MacSpeech Dictate will be available in mid February.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Of course iListen is an "inferior" product. And version 1.0 of MacSpeech Dictate will be "inferior" to version 1.0. And version 2.0 of MacSpeech Dictate will be "inferior" to version 3.0, etc.
Surely you would not have preferred that MacSpeech simply stopped selling anything for the past year while we developed this amazing new software. And your comment that we developed the new software "on the backs" of our customers is no different than the way future versions of MacSpeech Dictate will be developed. As I stated, with a retail price of $199, we can't simply give the program away to existing users. Also, we have to pay Nuance for every copy we distribute, so each and every person who acquires a copy will have to pay something.
As I stated, existing iListen users do not have to pay full price. They pay a discounted price based on the version of iListen they have and when they purchased it. So we are rewarding existing users for purchasing iListen by offering them a discount new buyers do not get.
Regarding Correction, I am not avoiding the question - please see my response to Helen Green.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
The reality is that you knew a far superior product would be available and yet continued to evangelically promote iListen to customers. I understand the way things work when developing products ... have been there and done that for over 30 years.
I applaud MacSpeech in its alliance with Nuance. As you know, I suggested to you in emails and those reading this forum such a move was one of the obvious ways MacSpeech could offer a competitive product. I do feel that this smacks a bit of the iPhone debacle in that you have promoted iListen to generate cash flow to only replace the recently purchased product with one that may actually work well and require a further investment of time and money from your loyal customers.
I purchased mine long enough ago to be willing to pay the upgrade price but feel the most honorable course for MacSpeech is to offer free upgrades if purchased during the time MacSpeech had internally documented the plan to pull iListen from distribution.
All of that said, I will wait to see what MacSpeech ships as Dictate and will offer my feedback once I have had time to test it in the trenches.
My position on iListen has not changed and will not change: it is possible to get the same accuracy in iListen as with our new product. And we have had several people stop by the booth and tell us that.
iListen will continue to be sold at retail outlets until the supply is exhausted. Anyone who purchases iListen in 2008 will be able to upgrade to MacSpeech Dictate for only $29. That means they will be saving about $30 over purchasing MacSpeech Dictate when it comes out. We simply can't offer it for free because of our licensing agreement with Nuance - they simply will not allow it.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
When will MacSpeech Dictate be shown in road shows in the Los Angeles Area? Does Anton have a copy for demonstrations yet?
I have received messages from folks expressing their concerns re. the decision to not support non-Intel Macs. I understand why MacSpeech is not offering a universal binary but wonder how many people particularly disabled have older machines and do not intend to or do not have the resources to purchase a new Intel Mac.
If anyone has stats on the potential number of Apple users affected please post. We are considering the development of a speech to text and control product that would be hardware and OS independent. It would work with any system including Macs.
Thanks
Sam
All this time, Sam, you've presented yourself here as an individual user with no "personal or professional interest" in the various products being discussed.
Yet, you've been hyper critical of the MacSpeech product throughout and have even recently stooped to weak attempts to unfairly lambast and attack the company for having innovated and introduced a superior product to iListen, which has worked well for me, and many, many others, for years.
NOW you suddenly reveal that you are part of some mysterious "WE" that is a potential business rival to MacSpeech, and have been working on a potentially competing speech recognition product?
ANY credibility you may have had has just flown out the window as far as I'm concerned - and you owe a lot of people an apology as far as I'm concerned.
What a world!
When will first reports/reviews be available? What is the microphone status?
Sorry you feel that way. The interest in making machines available to disabled as well as more recently an OS and hardware independent speech to text and control product evolved from my experiences with iListen and this forum. I did not enter this arena with any commercial or philanthropic intentions and my comments were based on my experiences with iListen and other similar technologies.
Some time ago, I shared our interest in this kind of project with Chuck in an email and we would only consider this option if there were a need and sufficient funding to move on such a project.
I can assure you and others that my interest in making speech to text and control technologies available to disabled was born of the feedback in this forum and would be made available under a government regulated not for profit foundation.
The we I speak of is the group of very talented engineers and scientists I have worked with for many years in developing physiological monitoring devices.
You, of course, can discount my sincerity if you need but the reality is as I have stated ... our interest in a product evolved from this forum. You have the reverse order of chicken and egg. This is one of the great benefits of this kind of exchange. If my intentions were as you suspicion, I would never show my hand at this stage nor in a public forum.
The we by the way has always been identified in this forum. I have delineated my back ground and interests several times.
Tis of interest you did not make similar claims when I let the readers know I was interested in getting this kind of technology in the hands of disabled via a not for profit foundation.
It is the open exchange of this forum that you need credit with my and our broadened interests in Speech Recognition technologies. Without this forum and the feed back from folks like yourself I may never have looked back at my history with disabled and what I could possible do to make a difference.
From Chucks earlier comments, it does sound as if the new Dictate will have enhanced corrections features. This is a very encouraging bit of news.
I eagerly await shipment!
Oh, btw, mine cost me $79 and my original purchase date was in 2006.
Congratulations Chuck!
And, won't it be nice if this product is so good that Sam is silenced? Nah, what're the chances of that!
But let me say to everyone that the possibility of closing down this thread *is* on the bubble. I'm not going to type out a lengthy comment that few will probably read to explain what is on-topic for this page. Your own common sense should adequately tell you what is on-topic.
But to summarize it, don't forget that this page was a review about iListen. This page was not ever intended as a forum for speech technology in general. We're allowing discussion about MacSpeech's next generation product, but ask that it remain on the topic of how use and workflows will progress from iListen to Dictate.
In short, this page's comments is on probation, and I strongly suggest everyone who is participating seriously consider whether *their* comment is the one that will convince us to close it down. Sam, no ill will intended, but we're especially looking at you.
That's good!
Okay here is my attempt to get this off on the right foot:
Is there a way that you can get your people at MacSpeech to begin to contribute to the forum re their experiences with Dictate, to scratch the itch we all have for info about the new product?
Anything will be appreciated; transferring profiles, macro's, etc?
Thanks,
MRH
Congratulations on you new product. I think you did the right thing transferring to a new engine that would allow your product to advance with speech technology while using your in-house expertise to stay in synch with Apple.
Many of us have developed word commands for the insertion of macros into text. Could you please address the transfer of these to the new program.
Thanks
Dave
Chuck and Sam, I really enjoyed your "sparring", and i feel like it helped me to understand the issues that brought macspeech to its current direction.
Chuck, your zeal is commendable, and/but, I feel a bit miffed over the latest switch out, and am glad that I waited before ordering; whether there was an ethical lapse on your part, (that has been addressed here by others) ...well, i dunno.I guess you gotta dance with who brung ya...
Sam thank you for your insight and perseverance, and/but I think you "had a dog in that fight" as we say down here in TX. Keep me posted about your "WE" ...and as for ATPM "looking at you" ...well as DeNiro said: "you lookin at me?"
FOR ATPM: PLEASE KEEP THIS FORUM OPEN!
strictly speaking, it is not about ilisten anymore -except for folks that have that "old" software, and anything that is not about ilisten would be offtopic. Maybe we need to move to a different forum/group for the "Dictators"...I know I will need a forum after I acquire the new software. Input from experienced critics, like Sam, will be invaluable to me, and i imagine that critical input would be valuable for everyone, including macspeech...and Chuck. Anyone can read a press release and advertising copy.
Thanks, marlon
Can we look at Nuance products to see if there will be similar options for Dictate?
thanks, marlon
Greetings once again from Macworld. Allow me to remind everyone that I have limited ability to respond while at the show. My next window of opportunity will probably be about this time tomorrow night.
Before I respond to the various comments that were posted today, let me give you some fantastic news: MacSpeech Dictate won Best of Show!!!
For those wondering about iListen's future, while we won't be developing it further, we will continue to support it for the foreseeable future - certainly through the end of this year, and longer if necessary.
Keep in mind that we went through something similar when we switched from Mac OS 9 to Mac OS X. There was a point where we had to stop developing versions for Mac OS 9, and then a bit later on we stopped support as well. The situation with Intel Macs is very similar. Apple no longer supports any kind of development for older Macs, and making a commitment to support a hardware platform for which Apple no longer ships new hardware would have resulted in a delay of several months. I am not making that up.
Our licensing agreement with Philips states that we must cease shipments of any product using their technology the moment we announce an agreement with a new partner. So the last shipments of iListen went out on Monday. It will continue to be sold at retail outlets such as the Apple stores until supplies are exhausted.
Regarding any suggested "ethical lapse" - I have to say that comment sort of rubs me the wrong way. I have been nothing but honest and open with everyone. iListen remains a great option for anyone with a pre-Intel Mac, and as our many existing users will attest (including some who have posted here), excellent accuracy is very achievable.
The new product costs us more to license, and is more expensive to sell as well. We are treating existing iListen users very fairly in terms of our upgrade policy. Think of it this way: if we were to continue down a path of developing iListen and introduced version 2.0, there would be an upgrade cost involved. Also, keep in mind that the last upgrade to 1.8, which not only is Leopard compatible but also introduced some great new features, was a free upgrade. We made this a free upgrade BECAUSE we knew something else was coming.
We certainly understand the concern some of you who do not have Intel machines have, but to be honest, the Dragon engine requires some pretty hefty processing power. Even if it were economically feasible to bring that technology to the PowerPC-based Macs, the resulting accuracy would most likely not be any better than what you have now with iListen. Yes, the new engine requires the additional processing power brought to bear by the Intel chips.
Those of you who have written your own commands or have zillions of text macros (like Doctor Black) will be pleased to know that you can export those from iListen and then import them into the new product and all of them should work. We know this to be the case with our ScriptPaks, and with text macros. For obvious reasons, I can't promise a script someone else has written will work in the new product, but we think it will. You will also have many more options for the types of commands you can create, and creating them will be much easier.
In terms of acceptance here at the show, the "Best of Show" award speaks for itself. I can also add that David Pogue has been talking it up tremendously, and our sales - even though they are pre-orders only - are outpacing previous Macworld shows (at which we have always sold out of iListen).
If I have not specifically addressed anyone's comments, I am sorry - please restate, or send an email to me personally and I will do my best to respond on a timely basis (but again - don't be surprised if it isn't until tomorrow evening).
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
MacSpeech will release medical and legal vocabularies later this year. We do not have a price or release date at this time.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Congratulations! I am anxious to see Pogue's comments and review. Please let us know when and where he is posting.
It looks as if MacSpeech is sensitive to the costs of upgrading and maintaining some continuity with iListen add-ons. If you follow through with a better correction option and customer support it would seem a strong win for MacSpeech, Nuance and consumer.
Please let us know if reviews are in the works.
Cheers
Sam
Thanks
Chris(Barry Bennetts Ltd)
I read in the original review that iListen was not for people under 14. Can I expect Macspeech Dictate to work for him?
Unfortunately, there is no speech recognition software - for either Mac or PC - that will work well for pre-adolescents.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
You raise an interesting and very important question and I'm surprised Chuck feels that speech translations software in general is not well suited for adolescents. He has much experience in this area and I am not suggesting that he is wrong but want to research this more because of the potential benefits for disabled teens and children.
Chuck does Nuance advise against using their speech engine with children and teens? Would you expect better luck with adolescents using the new Dictate product?
Please share references to the issues in using iListen and Dictate with this population.
Thanks
Sam
As you can imagine, I am exhausted from a very busy and exciting week at Macworld Expo - so please excuse my brevity.
The issue with younger voices is partly the fact that their voices have not matured. More importantly, however, is their lack of acquired knowledge of grammar, which results in an inability to dictate. To put it more simply, if someone has trouble writing well (regardless of age) they will have trouble dictating.
We have found that pre-adolescents tend to not do well with dictation because handling the added skill of dictating on top of speaking correct grammar is just too much for them. In other words, speech engine expects certain things that most kids just can't do.
The age 14 limit we suggested for iListen was based on our experience. We found that by age 14, many kids could indeed master the skill of dictation and had the proper grammar skills to support it. It was sort of a tipping point. we do know of a couple of instances where 9 or 10 year old kids did rather well, but they tended to be more of the "Doogie Howser" type of kid, if you catch my drift.
We also know of 17 or 18 year olds who have tried iListen and could not master it, so the 14 year old age limit is by no means fixed. It is simply where we thought the cutoff was for our ability to adequately support our customer base. Below that age and we found the average child simply did not have the skills necessary to have success with the program.
That having been said, the new engine will certainly provide more accuracy - but we do not believe it to be enough to result in a product that has sufficient accuracy to be usable for the average person ages 13 or under. As far as Nuance is concerned, they do not recommend their product be used by anyone that is a pre-adolescent. They did try supporting kids as young as 9 with Dragon NaturallySpeaking 4 (they are now on version 9). But they have backed off of that since then due to poor results, although they do still officially support teens.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
The enhanced Dictate and Mac should prove a good combination for disabled. Please let us know if,once released, pre-teens and teens are able to use it.
chris
Thanks very much for the information on what types of problems kids run into with the software. I appreciate having enough information to make an informed decision.
I have read about kids with motor or dyslexia problems being sucessful with voice recognition software. As noted by others, kids can be very focused when they are given an avenue for success.
Does Dictate work with voice recorders?
What time frame can we expect Dictate to be available for purchase?
Thanks,
Jeanne
The accuracy ratings are based on reading a story. You won't get that much accuracy while dictating thoughts into a word processor.
This technology has a long way to go.
Pecan
Thanks for your patience as I recover from a very hectic Macworld.
Jeanne - the first version of MacSpeech Dictate will not work with voice recorders directly. That feature is planned as an add-on for later this year. You can pre-order MacSpeech Dictate now. We are aiming at mid-February for release.
Pecan Logic - If you just say "car" you aren't giving the recognizer enough information to make a valid choice. Try saying "I drove my car to the store PERIOD" and see how it does. BTW, you could also try saying "I drove our car to the store PERIOD"
All any speech recognition software can do if you are speaking single words is to make the best guess as to what you said. With no context around it to determine meaning, you will very often get the wrong word.
Chuck Rogers, Chief Evangelist
MacSpeech., Inc.
While I am getting up to speed with MacDictate for medical transcription when I get my MacPro I will be working with Dragon in Windows in Parallels. Can I use the same USB headset for both?
Thank you for your time.
We won't be testing mics until after 1.0 ships. But that having been said, there is no reason for you not to try what you have and see what kind of results you get.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
The Macworld show special was only available on the show floor, as we are not shipping yet. We will have a different introductory special once MacSpeech Dictate is shipped.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
For those upgrading from iListen, there is no microphone with the crossgrade. The microphone you received with iListen will work fine with MacSpeech Dictate. Limitations on microphones are due to the Mac OS X, not the program you are using.
If you purchased iListen without a microphone, you can purchase a MacSpeech-certified microphone from our web site.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I apologize for the intrusion
it is of my opinion through the problems I am experiencing with iListen and the help team, that when the new software is launched we will probably be faced with the same problems with the help team that is not dedicated or caring to come back to existing customers using iListen 1.7 or 1.8
You say that you have many questions about the new software. Am I (expendable) as an old customer, not worthy of help?
i am thinking that your interest should be in helping your old customers to keep them on board so they will buy the new software, but at this rate. (You think I'll buy your software) regarding the way that your help people are treating me.
How much time does it take for somebody to pick up the phone and contact me or( email me with help not excuses) to help me through my difficulties with your software , is this to much to ask in this day and age.
Can you tell me when your UK Web site will be operating I am trying to download the English version of your upgrade, but all I get for the last week is a set of swinging curtains telling me you are at Mac World.
Not Happy
Denis UK
Absolutely no customer is expendable. You need to understand, however, that the week before Macworld, the week of Macworld, and the week after Macworld will ALWAYS be when you get the absolute worst support from us. We are a small company and simply can't keep up for those three weeks.
Also, our web site has been inundated with traffic and has had problems as a result. If anyone at MacSpeech could tell you when it will be operational, we would. All I can tell you is that our web team is working on it as fast as they can.
In regards to how much time it takes to contact you, our support team tells me they have sent several responses to your inquiries, so perhaps you should check your junk mail box to see if they somehow got in there. Our support manager also responded to me today that he is going to look for your phone number so he can call you. (If it isn't in one of your emails, then that would explain why you haven't received a call.)
In short, we are all working long hours to keep up with demand, and we probably won't be caught up until later this week or the beginning of next. I wish I could tell you we were as big as Apple or Adobe or Microsoft and could just snap our fingers to get something done, but it just isn't that simple.
Please bear with us. Resolving your issues are very important to us, but you are requesting a turnaround time that is simply impossible for us to accommodate as the time surrounding Macworld Expo is absolutely always the very worst time of the year for us, and this year, doubly so due to the introduction of the new product.
BTW, in the interest of full disclosure, I would like to point out to everyone that I forwarded an email I received from Denis earlier today to both our support manager and CEO. Both responded almost immediately. Here is what our CEO sent Denis at 12:55 CST today:
"Hello Denis,
"Our technical support department is very very busy right now.
"It is not that we have no interest in helping you, but indeed there are many many questions about the new software despite the fact that we are not shipping anything until the middle of February. They will be getting back to you, but it will take longer than it ought to.
"Thank you for your understanding.
"Andrew Taylor"
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Now that MacSpeech is embracing the far superior Nuance technologies, I have to wonder what it is about the Mac that makes voice input so tricky. Granted you need a USB analog to USB digital converted such as the VXI interface shipped with iListen but what would prevent a Nuance certified headset or microphone from being used with the Nuance port to Apple ala the MacSpeech licensing of of the Nuance speech engine?
Is there something about the Mac that makes audio input via a USB converter particularly challenging?
I am not sure I fully understand the question, but I will take a stab at it. First, it may be worth a quick mention of why we need a USB adapter at all.
All Macs since the very first Bondi Blue iMac and G3 PowerBook have had a "line in" sound port, but no "mic in" port. The difference is important. On Windows machines, most sound cards have inputs for both line in and mic in. Line in assumes an amplified signal, where mic in assumes an unamplified signal. In other words, if you plug an analog microphone into a line in port, there is not enough signal strength there to get any kind of decent results. Before going any further, I also need to mention that Macs want an audio signal (called "gain") about 10db higher than most Windows machines. This has been independently verified by Martin Markoe of eMicrophones.com. (This becomes important a bit later.)
OK, so you need a USB adapter of some kind to bring mic level sound into a Mac. There aren't that many chipsets out there for USB audio, but each manufacturer who uses a particular chipset can produce their own firmware. Those of you who have many microphones may have noticed they show up as "AK5370," "VXI 7.0x," "C-Media Headset" (even when the mic is not a headset), "Plantronics," and even "Unknown." When something shows up as "unknown" it means the manufacturer didn't label the firmware they are using, btw.
One more thing you need to understand: all microphones are analog - even "USB" microphones. What that means is that the sound is listened to by an analog mic, then converted to digital before being sent through the USB port. When the Mac gets the digital stream, it is identified as audio and then decoded back to analog by the Mac OS. From there, we can do something with it.
So here is where sound in on the Mac becomes "so tricky." If you take a microphone that was designed without the Mac in mind, it usually has a signal strength too low to get good results. This is perhaps a definitive example of "Garbage In, Garbage Out." If you search the web for the right things, you will find references from Skype for Windows users who complain that Mac users are harder to hear. This is because they are using the Logitech microphones, none of which were designed with the Mac's higher gain requirement in mind. (Further corroboration of the 10db difference mentioned above.)
With iListen, this was especially important. With Dictate it may be less important simply because of the higher accuracy, but also, our engineers are now able to concentrate exclusively on Core Audio, so additional tweaking may improve results and broaden the gamut of microphones we are able to certify.
As a FYI, we are not certifying any additional microphones until after we ship MacSpeech Dictate 1.0. This is because our audio engineers are still tweaking the code, and we wouldn't want to provide false information about any given microphone based on code that has not been fully optimized.
So the bottom line is that just because you have a microphone that works well on a PC does not mean it will work on a Mac.
In the interest of completeness, I will also add the following: we occasionally hear from someone who has professional microphones costing $400 or more who are astonished to find out their expensive mic does not work as well as our models that sell for less than $100. For the most part, this is because their expensive microphones lack the noise canceling properties that really have quite a bit to do with the accuracy one is able to achieve.
I hope this answers your question.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Thanks for the info.
Sam
He was able to get a very impressive 99.1 % accuracy with minimal training but laments the lack of Nuance "features like audio playback of what you said, a simple "add word" command, legal and medical versions, and non-English language kits" and more importantly that Dictate "lacks voice correction" requiring the user "fix transcription errors by hand". This is very disappointing but he hints MacSpeech may introduce a usable correction option in version 1.1.
He closes by commending MacSpeech for adopting the far superior Nuance speech engine and vastly improving ease of set up and recognition. I hope that revision 1.1 follows shortly after the release of 1.0. Accuracy is very important but without a good correction scheme this kind of product becomes unduly limited when used by physically disabled.
Re. the recent complaints about MacSpeech support, this has been one of my most favorable experiences with MacSpeech when trying to make iListen work. Support was prompt and courteous but unfortunately iListen was simply not a usable product for me regardless of how much support was offered.
Still looking forward to the 1.0 release! Is there further word on prerelease discounts?
Thanks to all who post. Are there any other speech recognition forums of note? Is there any dialog between the disabled community (especially war vets)and the speech recognition development folks? Where?
In the kinds of Dictate 1.0 version, there is a "u.s. teen" option...is that to be responsive to teen jargon: "dude/awesome" etc., or is there an identifiable way to think of voice quality/timber of teen users that is different from older users...
Since I am occasionally afflicted with "californicated dialect patterns" -and was actually in the redevelopment team of the word "dude" in 1981...and "duh/no duh" in 1969 -I wonder if it would be like totally awesome if I could get this -like- really cool dude speak recognition for my radical speech recognition package? Can I get a "Jive" or "Valley-girl speak" upgrade in the 1.1?
Like whoa! -thanks,
marlon
austin texas dude
The teen english language model is not so much vocabulary oriented as it is to the unique properties of the adolescent voice. MacSpeech Dictate has a tremendous vocabulary, so doing "teen speak" should be no problem regardless of the language model you are using.
Regarding other voice forums, I only know of one, which is the MacVoice list. You can find out more at the following link:
http://www.themacintoshguy.com/lists/MacVoiceFAQ.html
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
My only realm of conservatism is with technical equipment. If it works, don't change it; and if it isn't broken, don't fix it. However Chuck, I'm trusting you on this one. Even though I was pleased with Ilisten on my G4 and am now eminently pleased with Ilisten on my new intel Imac, I put in for Dictate today.
Question. After I install Dicatate can I still use Ilisten on the same computer so I can bring the new program up to speed? Hedging my bet a little.
Say hi to your wife for me. My daughter went to Cabana Cove for lunch today. Lucky girl.
Thanks
Dave Black
We believe the new program will provide significantly higher accuracy sooner for most people. It's not that similar accuracy cannot be obtained with iListen, just that for most people, iListen takes longer to train.
You can have iListen and MacSpeech Dictate on the same computer, as long as you don't run them at the same time. You can also import all your text macros from iListen into MacSpeech Dictate.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I suspect everyone is holding their breath and comments until the release of Dictate. Will you meet your target of Feb. 15? Will the 1.0 be shipped with a nuance like correction mode?
Thanks
We are on target for MacSpeech Dictate 1.0 shipping on February 15th. The first shipments will be for pre-orders and to distribution. It will take us a bit of time to catch up, so anyone ordering on the 15th may not get their copy for a week or two.
As we have already announced, version 1.0 will not have Correction. We knew this when we announced at Macworld Expo. We decided then it would be better to get this product into the hands of end users sooner, rather than later. Correction will be added in a free update as soon as it is ready.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I may want to wait until you have a release with correction ... will this effect the cost of switching for me or others who decide to wait in moving to Dictate?
Thanks
The pre-order price for those who already have iListen is $79. This will go up to $99 after we start shipping, so you save $20 by ordering now.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
When will it be available in the UK, please, in UK English? And how much will it cost?
thanks
Helen
You can order your crossgrade from iListen from our UK web site (www.macspeech.co.uk). The cost, including cost for shipping, will be available to you there.
It should start showing up in UK stores sometime in March.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Regarding correction, if Dictate doesn't have a correction facility, how will we correct? Will we still have to open a special correction window, as in iListen?
thanks
Helen
I can't comment on what we are doing in Correction except to say we are taking the time to get it right.
In terms of how you will correct the very infrequent mis-recognition in version 1.0, you either select the word or phrase by voice and re-speak it, or you simply double-click the word and re-type it.
What our beta testers are reporting is that MacSpeech Dictate 1.0 is so accurate that they do not miss Correction in the least.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I understand from your last few postings that my cost will now be $100 to shift to Dictate. Since time is more valuable to me than the cost of shifting to the Nuance engine, I would rather wait until a Nuance like correction scheme is shipped. Will this wait further increase the cost of shifting?
Second:
I keep seeing bloggers commenting on how "Apple is working to dramatically improve how well Leopard will work for people with disabilities". It seems many see Apple moving in the same direction Microsoft has by building in more advanced Voice Recognition and Control features. I would appreciate you comments on the possible competition from Apple as OS X evolves.
I have no information on what our sales team plans to do after version 1.0 is out the door, but I suspect the crossgrade price will remain at $99. I am perplexed by your unwillingness to save $20, however, as the update that includes Correction will be free to everyone who either crossgrades or buys version 1.0.
Regarding Apple's future plans, we have no special inside information, but I do have some insight on the matter, and I would like to stress these are my personal opinions and not those of MacSpeech.
I believe you will see Apple move forward with Command and Control by voice, but not Dictation. Steve Jobs has gone on record stating he does not believe dictation is something most people either want or would use, and further, it seems he isn't interested until computers can do 100% recognition without having to do Correction.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
To begin with I had not been made aware the deadline for the $80 switch to Nuance was expiring until the end of the dead line was posted Feb. 14. I did not read your reply until the 15th and assumed I had missed the dead line. I had not received an email as to the dead line. If you had previously posted the dead line I missed it.
However, if I did get 1.0 I could not resist trying to use it and I know from working with other products it would be of little use to me without a working correction option. I am in contact with several who have been promised a review copy and am anxious to get their feedback.
I continue to feel the correction option is essential for most users and certainly disabled. I trust this is a priority at MacSpeech.
Thanks for your feedback.
The Crossgrade offer is still available from our web site.
Regarding Correction, we have several beta testers who used to use Dragon NaturallySpeaking on Windows or on their Mac from within Parallels or VMFusion. They don't miss the Correction feature. In fact, one of them is recovering from a broken arm and hasn't used Dragon NaturallySpeaking for over a month!
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
That said, I suspect that the significantly increased accuracy and ease of training offered by the Nuance engine will make Dictate 1.0 useful to many.
I will try to test Dictate before going to press with our site focusing on computer based products for disabled.
Thanks
Sam
Thanks very much.
Louis
Contact our sales department at
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Please don't use this forum for questions regarding MacSpeech Dictate, as it is intended as a response to the iListen review the folks here at ATPM did a while back.
If you have questions about MacSpeech Dictate, please email [email protected] or email directly at [email protected].
Thanks for your cooperation on this matter.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
What exactly is the shipping plan, now?
I do not know who qualifies as a 'Founding iListen Customer'; so I am wondering when to expect delivery. I purchased 1 1/2 -2 years ago - where does that place me in the line?
Thanks,
Michael
"Founding customers" are about 1000 people who took a chance on a small company named MacSpeech over 10 years ago and bought a T-shirt for $99 without a promise that we would ever ship a product. They are the original investors in MacSpeech, other than friends and family of the first employees. In return for their support, they were promised first dibs on new products as they are released.
After our Founders, orders placed at Macworld will be fulfilled, followed by all those who crossgraded, in the order their crossgrade request was received. After that, pre-orders get shipped to new, non-iListen customers, followed by shipping product to distribution and resellers.
The entire process will take several weeks.
Again - I am happy to answer these questions, but since they are not about iListen they would be better sent to me personally. chuck.rogers [at] macspeech.com.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
So this is just for IListen now? Is there a forum for the new product? Chuck -it would be nice to have the feedback from other than "evangelicals"? Are there other forums for voice recognition programs and ideas? -Sam?
[email protected]
Thanks, looking forward to the new product when things shake out, and thanks to all in this forum for helping me get up to speed in the subject area.
Marlon, Texas
As MacSpeech Dictate has not been fully rolled out to the public yet, there are no forums for it.
There is a general discussion list for Macintosh Voice software at www.themacintoshguy.com/lists/MacVoiceFAQ.html.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
People will increasingly interact with computers using speech or touch screens rather than keyboards, Microsoft Corp. Chairman Bill Gates said.
"It's one of the big bets we're making," he said during the final stop of a farewell tour before he withdraws from the company's daily operations in July."
I can't help but feel that Jobs is also focused on enhancing the speech engine in OS X. Chuck, how much in Dictate is dependent on enhancements in Leopard and should I expect to see better performance in Leopard than Tiger? What kind of performance improvement if any has MacSpeech seen in iListen users who switched to Leopard?
If Gates is right and Apple keeps its paws off of Speech to Text enhancements, MacSpeech seems poised to offer the kind of technologies that will make MacSpeech a key player in moving Apple forward in sales.
Would be interested in your feedback on the state of cross grade shipments as well. We are holding off on opening the Assistive Devices site until I can evaluate Dictate.
Do you have any feedback from non-beta tester iListen users who have made the cross grade?
Thanks
Sam
As MacSpeech Dictate just started shipping to our Founding customers within the last week, it is a little early to be hearing from anyone yet.
Regarding Tiger vs. Leopard, MacSpeech Dictate is programmed to take advantage of enhancements in Leopard, but that doesn't mean Tiger users will notice any difference, especially in terms of accuracy, as the engine itself doesn't do anything different between Tiger or Leopard.
To summarize, the only way those using Tiger are limited are in ways that are due enhancements to Leopard that aren't available to Tiger users anyway.
Shipments are happening slowly as we ramp up production. We expect the entire process to take several weeks.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Sam
I hope that you can help with this, regarding cross grade/upgrade and registration
I haven't used iListen for quite a while, but when I started it up today I was told that I must first register. I am pretty sure that I registered with macSpeech when I purchased iListen (at the Apple Store in regent Street, London) latish last year, but being a trusting kind of person set off along the route which MacSpeech was charting for me, into the registration application.
There I was told that to cross grade MacSpeech Dictate (which is something that I am really keen to do), I must first upgrade to 1.8. I was happy to oblige, so selected the upgrade option (UK English) only to be told -
"Error NSURLErrorDomain 100: http://www.macspeech.com/updates/iListen.UK.1.8.Updater.dmg"
with no further explanation. Not very useful.
Struck by the similarity to the microphone compatibility problems I had with the installation process, where I was told that you knew that there was a problem but you left me to find my own solution, I struck out in a new direction.
On the registration page, there is a link to something helpfully called "[email protected]",
which is recommended for people who have "questions or problems". Ever optimistic, I tried that, to be told -
"Cannot show URL: mailto:[email protected]"
Next, I saw and decided to try "For further assistance with registration please visit the Registration Help page".
There I found 2 useful suggestions. Link to the "Update Center" or download a "standalone registration application". As I was trying to get registered (even though I was sure that I already had) I went for that one first. I guess that you will agree that I was entitled to be disappointed to be told -
"Error NSURLErrorDomain 100: http://www.macspeech.com/updates/MSRegistration.app.zip"
None daunted, I found my way back and tried the other option - follow the link to the "Update center". I found myself at a page which looked familiar - yes, indeed, it was where I had first started!
What is going on? Do I have to upgrade first and, if so, how?
Len
Problems like this need to be directed to MacSpeech Support. I don't control the web site behavior, and the quickest route to getting a resolution would be to contact our support staff directly. You can do that by going to www.macspeech.com/support and click the "Submit Ticket" link.
I'd also like to remind everyone that this thread is for comments in response to iListen. The good folks at ATPM are being very lenient right now, but this isn't a catch-all place for anything related to MacSpeech or speech recognition in general.
Thanks for your understanding and cooperation in this.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Looking forward to it.
Anyone know of a forum specific to Dictate?
Michael
I realise that this is not really the place for my problem. I will get a support ticket. I posted it here because when I did report something to support previously they never replied and secondly because I fully expect them to tell me they won't deal with it because I seem not to be registered and, of course , the problem on which I need help is that I can't register!
What I really wanted to know from you is whether it is an essential requirement in order to cross grade to Dictate, that one first upgrades to iListen 1.8?
Regards
Len
First, it is not a requirement that you register to receive support. Our Support site has a registration option, but that is only for the convenience of our customers. (It allows you to track your tickets more easily, as well as see your ticket history.)
Second, it is not necessary to have 1.8 in order to cross grade. The product you get when you cross grade is exactly the same as if you purchase brand new, only the price is different. In other words, the cross grade version of MacSpeech Dictate does not check to see that iListen is installed before it will install.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
The Dictate Quick-start flyer repeatedly refers to the "amazing accuracy" you should expect when using the product. Those who have struggled with trying to make iListen work will be blown away. However, those using other products such as Dragon will be less impressed but breathe a sigh of relief that Nuance finally ported their speech engine to the Macintosh.
There is essentially no usable correction scheme which in my opinion severely limits its usefulness and certainly makes it impractical for many who are physically disabled. It seems to have some of the same quirks I noted in iListen. For example when I use the command to select text and then delete it, not only is the text deleted but the word delete also appears. Several times when correcting text manually text previously deleted appears when I began dictating. Once after manually deleting text the application stopped responding to my dictation within Pages. The dictate application showed that my speech was being monitored but no text was displayed. The only way I was able to resume dictation was by shutting down Dictate and restarting.
David Pogue, in his last review, suggested several bugs would need to be squashed before shipping,however, it appears that more work needs to be done. Although accuracy it is greatly improved a workable correction scheme needs to be implemented before it can truly compete with the Windows version.
I, as I am sure many others do as well, often find that not only do I want to correct errors in speech to text translation but also to go back and modify phrasing and syntax. This is best done manually when using the current version of Dictate and, as I've noted above, at least once manual correction seemed to have prevented the application from responding to further dictation.
All of this said, Dictate is light years ahead of iListen and with improved correction options and the extermination of a few bugs, Dictate will offer Mac users leading edge speech to text technology.
Again as in iListen, the program seems to retain deleted or corrected text that may be randomly inserted when creating a new paragraph or simply adding punctuation.
I suspect, I and most others will be able to use Dictate after some experimentation with the correction options shipped with Ver. 1.0.
Unlike iListen, Dictate crashed when I launched VMWare and Vista. VMWare was capturing the USB Audio interface when Dictate crashed. Dictate seems less capable than iListen when using VMWare.
It is a pleasure to be able to dictate on my Mac using a native application. Fingers crossed the bugs are quickly addressed and better correction added soon.
Sam
victormedina.typepad.com
The blogger is an attorney elated with Dictate. He noted the same problems I have but goes as far as saying Dictate "most likely saved my career."
Well done review.
Do you know when Dictate will be shipped in the UK? I ordered and paid for it almost a month ago!
thanks
Helen
I don't have any information on when specific orders are being shipped, but you can always check on the status by logging into your account on the MacSpeech web site.
As an FYI, credit cards are not charged until orders are shipped. Also, there is a difference between an authorization and a charge. When a charge is authorized (which happens at the time you place the order), the credit card company puts a hold on the amount of your purchase but does not charge it to your account. This way, the money is reserved for when the merchandise ships - (so you don't go over your credit limit), but no interest accrues on your purchase until it is actually charged.
If your card HAS been charged, then that means it has been shipped. When it ships you will also receive an email with a tracking number. Again, if you log onto your account in our online store you can see if it has shipped or not, and if it has, get your tracking number (in case the email notification of shipment got caught by your spam filter).
If your order status is listed as "pending" then it has not shipped yet. All orders are shipped in the order in which they were received.
If you still have questions, please send an email to sales[at]macspeech.com (replacing the "[at]" with the "at" sign) and our sales team will check on things for you.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I assume MacSpeech is aware of the problem with Dictate losing track of the dictated text. The Do Select, Do Delete, Insert before etc. commands become useless if the text is manually edited. Several times complete paragraphs were deleted by Dictate when I was seeking to modify one word.
This flaw seems to be a carry over from iListen and is a major issue when trying to correct or modify a document. I am going to try to use Dictate to complete a few reviews of the product and see how well I can work within the limits prescribed by this problem.
I did figure out how to add canned phrases by poking around the global options under new commands. I only knew to look because Pogue stated Dictated supported this option. Will MacSpeech release a manual and or video tutorials?
Thanks
Sam
The problem you are experiencing is not an artifact from iListen as there is not even one line of code in MacSpeech Dictate that was taken from iListen. It is an issue with how Mac OS X processes text. Our development team is working on it, but for now, the same rule applies: you can use your voice to correct, but once you use the keyboard, MacSpeech Dictate has no way of knowing what was typed and can therefore become confused. Again, this is an issue with how Mac OS X processes text and not a problem with MacSpeech Dictate itself.
Think of it this way: you have two keyboards connected to your computer and one mouse. You have control of the keyboard and mouse, but the other keyboard is controlled by a blind typist. She's very good, never misspells a word, but is a little deaf in one ear and sometimes types the wrong word.
OK - the only way this blind typist can accurately do her job is by counting the number of characters she types. So when you dictate to her, she types away, keeping track of exactly how many characters she types. If she types the wrong word, you tell her what word it was and she counts back the correct number of characters and makes the correction you tell her to make.
But suppose you decide it would be just quicker if you just do the correction yourself. So you use your mouse to reposition the insertion point, select the text you want to change, and type it in on your keyboard. What you have just done is throw the character count off for your blind typist. From that point, she isn't going to be able to accurately enter the text you say in the correct place, nor will she be able to find mistakes accurately when you ask her to correct them.
The above illustrates the limitation imposed on us by the way Mac OS X handles text. Similar issues are imposed by Windows with Dragon NaturallySpeaking, btw, its just that they have had years to work with certain programs to work out the issues. As we have just got the basics working right now, we haven't had the opportunity to make adjustments for individual applications (including our own Note Pad) yet. I can tell you our development team is working on it, however.
Yes, there will be a manual coming along shortly.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I have almost finish a second review of Dictate and find its accuracy reduces the need to correct conversion errors.
Even though there is no shared code between iListen and Dictate the lost cursor problem is the same. I hope your engineers can over come the Mac OX limitations and improve the programs ability to edit text. Even with this limitation, I am enjoying the advantage Dictate gives me in working with my Mac.The longer I work with it the more proficient I become at working around this issue.
This is truly a significant improvement over iListen and I wish your team well in resolving the issues posed by the Mac OS.
It works great, and recognition is excellent. Oh, and I am using my Sennheiser ME3 Noise canceling mic on my iMac with great success.
Sure wish we had a Dictate forum...
Michael
It would provide a forum for users of all assistive technologies such as iListen and Dictate.
Just finished a review of Dictate and podcast of same using Dictate. Once I began to learn how to deal with the temperamental cursor and spontaneous appearance of deleted text it added significantly to my productivity.
Approaching 500 (!!) comments on this page, everyone should note that we will probably deactivate new comments on this page soon—perhaps by the end of this month. The author has already proposed that we do so, and Chuck Rogers is not opposed to it.
Hopefully, the author of this original review might be interested in publishing a new review of Dictate in a future issue of ATPM.
I hope the information remains logged on the net.
Thanks again for chance to comment and for the other APTM articles and forums.
Sam
I hope you are planning to post a review of the new MacSpeech product, Dictate, so that comments about that product may be posted. This forum has been very valuable.
Donna Pointer, iListen user
If I run back to the Apple Store and buy one of the two remaining iListens, will the discount for upgrading to Dictate still be valid a few years from now when I upgrade to an intel Mac? How long will tech support for iListen continue?
Another problem. I have allergic post-nasal drip. Can iListen ignore throat clearing and coughing? Can this problem be solved by simply moving the mic away?
Potential Future Customer
Regarding MacSpeech Dictate, the cross grade offer is only available for a limited time. It will not be available after existing stock of iListen at our resellers is exhausted. The cross grade offer is only $29 for anyone who purchases iListen in 2008, so you should go ahead and place your order for MacSpeech Dictate, so you will have it at the cheapest available price.
Regarding your allergies, there are actually two different issues here. The first is when you sneeze, cough, or clear your throat. Speech recognition software converts sound to text. It can't distinguish between the sound of a cough or sneeze, or that of a door slamming or sound on the TV for that matter. If you feel a sneeze coming on you should turn the mic off before you sneeze and then turn it back on afterwards. The same is true of clearing your throat, coughing, etc.
The other issue is that your voice will sound different to the program when you are suffering from allergies or a cold. This can result in reduced accuracy. What many people do is to create another voice profile for these occasions. When they have an allergy attack (or a cold) they simply switch to the profile they created for that situation. The resulting accuracy still won't be as good as your regular profile, but it has the benefit of not applying changes to your regular profile that was a result of your voice sounding different due to the cold or allergy attack.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
I have finally received my Dictate crossgrade disks. When I tried to create a profile, however, i discovered that the English data disk I was sent contains not UK English but US English - it asks me to read 'period' instead of 'full stop' and spells 'realise' 'realize'. I've emailed MacSpeech Uk support but was wondering if you have any comments on this. Has the UK version not been created properly?
thanks
Helen
All versions of MacSpeech Dictate come with 6 versions of English: US, UK, Australian, Indian, Southeast Asian, and Teen. What happened is when you created your profile, you did not choose the UK language model from the language pop-down menu.
If you create a new profile and select the UK Language model from the pop-down, everything should be fine. If you have further problems with this, you would need to contact our support department directly.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Helen
In AUDIO MIDI SETUP (in UTILITIES in APPLICATIONS), when the audio input format is set to 1 ch - 8 bit, there are further small but discernible breaking noises. I have checked that iListen uses this 1 ch - 8 bit setting and will pick up this noises. When the input format is changed to 1 ch - 16 bit, the breaking noises disappear but this format unfortunately is not used by iListen. This breaking noise is the second problem. When the two problems are compounded together, as is the case when the power adapter is plugged into the iBook G4 and iListen is operating, iListen will pick up both the humming and breaking noises. I think this has made the dictating accuracy of iListen quite poor so far.
I wrote to MacSpeech support three days ago and have yet to receive their reply. Any advice on these problems with the Plantronics Audiio 85 headset microphone?
There was a bad batch of Plantronics microphones that had a hum in them. Not sure if that is the exact problem with yours or not, but support should be able to determine that and get you a new microphone if you bought it with iListen.
Contact them again if they still haven't gotten back to you.
Chuck Rogers, Chief Evangelist
MacSpeech, Inc.
Just a quick note to let you know that MacSpeech and I have parted company due to differences of opinion on several matters. Although I will continue to be active with this forum, I will no longer be representing MacSpeech when posting.
Chuck Rogers, former Chief Evangelist, MacSpeech
With Chuck's announcement, and because this review specifically covered the iListen product which has been EOL'd, and also because ATPM requests all comments remain on-topic to the product reviewed, ATPM has decided the time has come that further comment discussion on this page would be of no use. Therefore, effective immediately, this thread will be closed.
ATPM has already requested a review copy of Dictate and an author to pair with it. Watch for our review of Dictate in the coming months.
Thank you all for reading About This Particular Macintosh.