Vista Speech Recognition


Microsoft Speech Recognition has been available as a part of Windows Vista since its beginning. Recently, interest seems to have increased in this feature, probably because of the release of Windows 7 Beta and the hope that there will be some needed improvements made to this nice but somewhat buggy application.

The application was new to me so I was more than a little concerned with how hard it would be to get it up and running. No reason for concern here. Just a matter of buying a microphone, plugging it in, and Vista recognizes it and makes it operational. Next, I went to Start/ Control Panel / Speech Recognition Options/ Start Microsoft Speech Recognition, and followed a 5 minute tutorial that only permitted the expected commands. No mistakes were allowed. This made it seem all too quick, easy and I was actually a little excited about getting started.


The tutorial finished, I opened the actual program and BINGO, excitement over. Once the real program was opened things became a little more difficult.

Even though I was using a relatively expensive microphone the program didn’t seem to recognize the correct commands. Not knowing the correct command would be expected in this situation, but there were too many instances of using the correct command and the program just not responding properly.

After about 6 hours of use over several days I decided that, if I were ever injured or disabled, it would be a nice free program to “try”, but I was glad I had use of both arms, my mouse, and my keyboard. The whole thing was too much like a comedy skit for me. The dogs were getting nervous with my repeated monotone commands to close___
___close that___OK___ address bar___ open address bar___ location___ ok___ OK____ number. The first day had the wife laughing but by the second day she and the dogs were showing more than a little irritation with the constant monotone blasts. In fact using the program was about like training an out of control dog. Sometimes it listened and sometimes it was off chasing squirrels. Some of the videos included below get that point across.

Because I had Windows 7 up and running I gave it a try hoping there would be some improvement with the new OS. Alas, this was not the case. The application seemed identical to me. There may well be some improvements that I didn’t notice but overall the results were exactly the same.

If you really need this type of software I’m sure it could get the job done, but it wouldn’t be fun on a daily basis. It’s not something that I would use for creating documents in Word instead of manually typing. Some of the articles I’ve read indicated about a 90% accuracy rate and others indicated a 60% accuracy rate. I didn’t see anything near that.

Is it a good useful program? Yes, I’m sure it’s been a life saver for anyone unable to use the keyboard and/or mouse for information input. The problem for me is that there is a big difference between being able to open a couple of documents or programs and actually getting large amounts of work done. Using the program and becoming accustomed to its quirks would require a lot of time, certainly more than a few days. It’s a program that I would use and be glad to have if it was my only option, but it’s not a program to use in order to save time creating a word document or because it’s so easy to navigate using voice commands.

Maybe in a couple of years, if Dragon Naturally Speaking keeps the pressure on Microsoft, that situation will change. I’ve viewed and included a couple of telling UTube videos for your enjoyment. Keep in mind that your results may vary. Mine certainly did.


MS Document


Vid I

Vid II


Vid IV

Paul Nielsen SSMS Vid

 1,098 total views,  1 views today

(Visited 1 times, 1 visits today)

47 thoughts on “Vista Speech Recognition”

  1. I must admit:
    After reading my post a day later, i think i should ask “Alfred” to start dictating my word docs cause man the grammer i use sometimes 🙂
    My old school teachers would be so disappointed 🙂

  2. I must ammend my previous statements somewhat:
    1. The “macro fuction” is not built-in with Vista but download from MSoft.
    2.For those wanting to try the above set-up in there home but have no Vista….try “E-Speaking”. Not as intricate as Vista’s wrt creating macros for multiple actions for one command, but you are able to create your own commands plus it has a 100 built in commands for the basic stuff.
    3. I thus agree with those who stated that the people who spent five minutes on a tutorial are never going to the full effect of what this technology can offer. If 5 mins was all it took for you to write off this tech, then it’s better you rather leave it to those who put in the effort for research and training of their SR software.
    By the way, i’m not in IT, i’m a 26 year old Project Manager, which if i could do this with nothing more than a bit of research, then so you YOU!!
    Ciao, Jon

  3. I’ve spent the last few days reading up on reviews wrt “speech recog” for Vista and XP and i must say that after reading all the reviews (including this entire article), 90% of all comments (good or bad) are purely about “dictation”.
    I’ve been using “speech recog” for over a year now and i’m still in love with it. Granted, I don’t use the dictation function much, but i have “mics” setup in every room of the house, with a small speaker for responses, and when i wake up in the morning, i don’t need to see my pc screen or even be anywhere near my “office” and i can instruct my “home network”, i called him Alfred, to play a particular song or playlist and he’ll auto start up Winamp, select the playlist and/or song and start playing ,no probs.
    I’ve gone further and created macros to intruct Alfred to check if back-ups have been done and he’ll respond accordingly. Thats a simple “If/else” statement built into the XML documentation. Sometimes it actually feels as if he’s AI.

    I’m getting carried away here again, but people are so amazed that i’m able start up movies on my TV by simply giving a single instruction and BAM, a series of actions take place onscreen, and the movie starts.

    IF this is something you would like to have in your home, but not pay a fortune for Home Automation, then this is the answer. And i did all this with Vista built-in.

    I’m just sad that so many others (as per this article) have written of this technology when i’m having so much fun!!

  4. Software Diplomat

    There is something I always wondered about voice recognition software:

    It uses your voice to execute commands, right? So if you were to say “beep” for “Open Outlook” and “boop” for “Close Notepad”, couldn’t you program it to understand R2D2? 😉

  5. For those who are questioning whether I finished the training routines the answer is yes. Again it’s not that the program doesn’t work or that it’s not useful. Like anything else there is an amount of time necessary to become proficient with any new program of this sort. So it works, it’s useful, it’s just not something I’ll use on a regular basis, probably because I don’t have a need for it.

  6. I have used the voice recognition that is part of Vista and find it to be a huge improvement over XP. I used a head set with a microphone and spent about an hour training it. I have gotten well above 90% accuracy. While there is a lot to knock about Vista the 3 things they got right were VRS, the snipping tool and movie maker.

  7. I disagree with this article completely. I’ve used two separate mics, one on a headset and one stand alone, and both types have worked nearly flawlessly.

    There’s a few training routines you can complete which will further the quality of WSR. I’m not sure if you completed that entire routine?

    Anyway, I’ve been incredibly impressed by the results. It’s been incredible on some grammatical corrections as well. For example, it will accurately type (for me) the following sentence:

    There were two knights, with their swords gleaming, who fought bravely on a night which was too cold.

    All three homophones, without any change by me, were properly typed by WSR. While this program isn’t perfect, it’s certainly on the right track.

  8. General Electric spent several hundred thousand
    dollars in early 70’s on VRS (voice recognition software)without success. Forty years later it still is not perfected; try calling companies using “VRS” driven menu’s, and the common response is “I am sorry I did not understand.”

  9. I’m an engineering manager and I share an office with two colleagues. My work is about 50% analysis and 50% communication. I cannot imaging sharing an office with others who are talking to their computers all day long. Phones and cell phones are bad enough! The mouse and keyboard are alive and well in my work environment.

  10. Only 6 hours with any Speech Recognition program and it’s going to make mistakes. Much like the Dragon series, you have to keep training it to your voice…the more training you do with it the better it understands you.
    Of course, the first time you go through speaking so that MSR learns your voice only goes towards navigation (open this, close that, select “app”).
    So yeah, only spending 5 minutes training Windows Speech Recognition is only going to create the “base line”, hence all the errors you were getting.
    If the reviewer took the time to train MSR more (like spent an hour training it) and then did the review, that would make more sense.

  11. voice recognition

    I have tried Voice recognition. I wish it worked well (or will one day). If you have a cold, need to clear your throught, cough, sneaze, blink… (ok, maybe you can exclude blink) you get many errors. It seems once the errors start they just continue to multiply. I found it VERY frustrating and time consuming. I can type faster with 2 fingers! Like I said earlier, I hope it works well one day soon.

  12. We have tryied to use voice software, and it worked only some of the time,as for the person calling some us clowns well maybe he should be in our mocasines for a while it took me 20 min. to type this as my hands and arms don’t work very well. and we speak with a soutwestern acsent

  13. Michael D. Kast

    …am a techno-peasant, but don’t mind typing or using a mouse, or pad. In the physical world there is a concept re: physical human ergonomics. That said, I can’t say much for the instructions that Vista has, i.e., poor mental ergonomics as per my experience with it. Makes me afraid to venture further with voice recognition programs just for that reason alone. From what I’ve read from you good folks, puts me on the fence, but what would win and get me over the fence would be a voice recognition interface via cloud computing connected to a program, perhaps in a Super Computer! Isn’t that how it is done on the Star-Ship Enterprise?

  14. Is there a software program that recognizes vocalized music, singing, that can convert to printed out musical notation?

    It would be nice if I could “hum” the notes of songs I write, then immediately see the sound as a musical score.

  15. i would hope keyboards and mice never become obsolete.
    not only are there the deaf and speach impaired but there are also those like me who do not care to have to talk all day long. some of us actually do not have to hear ourselves and are quite happy with peace and quiet.

    i dont use my headset and i dont use voice at all and i get so sick of people wanting me to use voice and i do always refuse. i have no needs or desires to talk to my computer or to talk to anyone via voice.

    i only came to read this because of the newsletter putting it as ‘obsolete’ as if to say we lose choice to keep mouse and keyboard. i would hope that is not the case. let all of the lonely people have their way but dont let that take away from those of us who are at peace with self.

  16. Stuart Bennett

    several years ago i utilized “Dragon Dictate” just to see how easy it was, yes there’s a basic tutorial but then it goes on to a more comprehensive one that last for about 3 hours, your reading text from things like alice in wonderland etc.

    i found it fairly good, although in typing it did make the occasional mistakes, however while it recognized commands very well i am not sure i could really get used to giving verbal commands all the time.

    bare in mind this was years ago and the software is probably better now, i haven’t ever tried speech recognition in vista so i can’t compare the 2 but from what your saying i probably wouldn’t enjoy it much.

  17. Years ago I used Dragon Naturally Speaking, IIRC version 5. When working in the recommended mode of continuous speech it was absolutely hopeless, despite me going through the whole training section numerous times. When it went ‘crazy’, after one or two sentences, it was impossible to make any sense of the text. Worse, it seemed to stay in ‘crazy’ mode, even after a pause and starting a new sentence.
    My English is pretty good but although my mother tongue is German and I have an accent. I thought that the training session was designed to take care of that. It didn’t.

  18. I find Vista’s VR is better than Dragon, because it is simpler. For instance, correcting is easy and quick. It gets better with use, and becomes used to one’s voice. The main problem is it doesn’t come with a manual.
    For instance, How does one capitalise the first letter of the word after a double quote? Put the quotes in later seems to be the best way, but it’s awkward. “He isn’t here,” is the correct form. “he isn’t here,” is not. The programmers should have picked up on this.
    Sometimes it has difficulty ‘learning’ a new word. This can be done in Dragon, but not easily in Vista.
    The most important thing is: No Background Noise. So if someone else is in the room, or the TV is going, forget it and use the keyboard.

  19. I got interested in speech recognition with Win3.1. It was pretty impressive way back then, but it does take motivation and lots of practice. As Bruce Cyr said, though, losing the use of your hands and/or arms would probably give most of us plenty of motivation to work with the program. I for one appreciate Microsoft’s efforts in this arena and hope they keep working to improve the application. I’m sure many of our Wounded Warriors will find it helpful.

  20. American accents typically sound muddled, so you might have to pretend to be British to get your program to work properly. Also, I don’t know if any voice-recognition software works with dialects.

    The only voice-recognition software I have ever used heavily was Opera’s built in voice-browsing. Though I enjoyed it, [it entertained and confused the cat since she thought I was trying to teach her a new trick] I have an easier time just using the keyboard and mouse despite how bad my eyesight is.

  21. I have tried windows speech recognition on my laptop, along with voice recognition on my gps mapping program in a big rig, and yes, you need a good headset mic to make it work, and need to set for background noise, but i did find that speaking “monotone” was less effective than actually just speaking normally, but of course being Canadian, no heavy ‘merican accent might help too. worked great…handsfree, nothing better.

  22. I have been a user of Dragon NaturallySpeaking since it first came out. I have the latest version installed on my new computer. I admit, this does give me an advantage over those were just starting to use Dragon NaturallySpeaking. I do suggest that a person expect to create at least two or probably three Dragon voice files. I have one file labeled as “natural voice file”, one labeled “I have a cold”, and one “who am I?”. The last is a special file that I created when I realized there were times that my voice for some reason was very hard to interpret. I knew that’s because during those times my wife would keep asking me “what did you say?” Either way, this reply is dictated using the version 9 of Dragon and I have yet to touch the keyboard! I will agree that to obtain high accuracy requires considerable training not just in the training programs offered by and taking the additional time to train words or phrases that Dragon has misinterpreted. When an individual says “what did you say?” We politely repeat, however, when Dragon, does strange and wonderful things we get mad as Dragon rather than politely repeating what was misunderstood so that Dragon becomes trained to that particular variation of words.

    I admit I have seldom used Dragon to control the computer so I am not proficient at that however I know it can be done and if the arthritis in my hands becomes much worse indeed become proficient at that particular use.

    By the way I’m using a wireless microphone to dictate this reply.

  23. I use Dragon seven for several years now, recently upgraded to version 10. And I’m using it to dictate this note.

    With the improvements in version 10 I use my keyboard about 4% of the time. I use my mouse perhaps 10% of the time. The point I’m trying to make is that after I get it fully trained I will be retiring both mouse and keyboard.

    I also use at work and I find I am between 150 and 175% more productive.

    I have not tried Microsoft and I do not intend to try Microsoft. Microsoft should focus on streamlining their operating system and Microsoft office and stay out of other people sandbox until they get their own act together. On my other machine I am using Ubunto which I find is several times faster than Microsoft. As soon as a good speech recognition package is available for Ubunto I will switch over both machines and bye-bye Microsoft.

  24. I strongly agree with Marco’s post above. My system apparently has most of the features that he mentioned. However, in my case my professional installer put in an extra big sound card. We had to sacrifice the modem that was inside the tower in order to make room for it. However, I don’t need that modem because my computer is on cable, and I send faxes without the computer. Dragon naturally speaking works fine for me. My home is quiet without background noise. I try to speak distinctly. And I spent a little time training the Dragon.

  25. As I was saying, this Microsoft product works well with other Microsoft programs. Don’t try using this with Firefox, for example.

    And BTW be careful when you say, because it might be understood as

  26. As a person with disabilities, I’ve used DNS and Dragon Dictate before that. I found DNS to be such a memory and CPU hog that I bought a 6GB system with 64-bit Vista. Then I discovered that DNS had no plans to upgrade to 64-bit (they now write that they have “begun a Beta”). So my introduction to VSR was born of necessity.

    Once I got over the differences in the command language, I’ve gotten pretty good at using VSR. It is a Microsoft productDon’t try using this a

  27. I am still waiting for speech recognition to get to a decent level. I am not disabled or anything like that, but I have a number of applications that I want to use my computer for that simply cry out for this stuff.

    The poster that said he worked on speech recognition said that apart from the disabled (reason enough to get this stuff developed), there was no real reason for speech recognition. Sadly I disagree, as I can think of several reasons. The first is home automation. I would like to automate my next home. Speech recognition would be great in that aspect, so you don’t have to use a remote or station to do things. The second is car PCs. Speech recognition is a huge advance here. Third is entertainment centers. Would be nice to no longer absolutely need a remote.

    Unfortunately I don’t think we will get far with speech recognition until computers get MUCH better, and we learn to use multi core processors to their full capability. I have a feeling that we need to use several different comparison engines to reduce false positives.

  28. I just bought a new laptop with and discovered the word recognition program by accident. I had just loaded DNS 9.5 onto the computer but it would not let me get my the initial recording steps sayoing that the sound system and mike were bad,{ I am paraphrasing). I spend the time necessary on the VR that it required and found the program to be as good as DNS 9.5. I couldn’t remember some of the commands so I ran the tutorial again paying more attention to what was written above the commands that they wanted you to say and learned even more the second time. I think the program is phenomenal and would highly recommend it to anyone who wants to spend a little time learning how to use it. The command,”show numbers” is fantastic when it comes to operating any program where you are not sure how to specify a particular command. I found that microphone location is a big part of whether the program understands what you are saying. I use the headset that came with DNS which I also think is a great program.

  29. Due to a pinched nerve in my neck rendering my right arm virtually useless, I used Dragon Naturally Speaking 9 for about 6 months in 2008. It worked OK. However, now that I have had neck surgery and have the use of my arm again, I find that I just don’t even think to use Dragon any more. Easier to point and click and type. My point and click is way more accurate. My typing is another story.

  30. I’ve only used Dragon NS for word processing ~(letters, documents and email) and have found it to be very accurate. The key to high level accuracy is a well specfied PC, i.e. a good processor, plenty of RAM, fast HD drive and a good graphics card with its own RAM; and a good microphone headset. VR is an application program and for it to give the best performance it needs the best operating environment. As for navigating round a PC I think a Keyboard & Mouse is easier; if you think about it it’s more mouse than Keyboard.

  31. Interested to read what Don Piller had to say. I have been interested in speech recognition since I first came across it w-a-a-y back when we started pioneering in office automation. Lack of keyboard skills was seen as a stumbling block to getting office staff in general and managers in particular directly involved in IT. Started with an Apple II (if I remember correctly). Ruddy thing wouldn’t acknowledge my simple “Good morning” if I had a bit of a cold or had had a heavy night out.

    It didn’t work then and it doesn’t work now.

    My most recent experience was several years ago, when I had retired after 30 years working in IT. A friend, who had ambitions to write a book, sought my help to set himself up with voice recognition software because he couldn’t – or wouldn’t – take the time to acquire typing skills. I told him it would not work, but he took me to a very slick IBM demo of ViaVoice. He was very persuasive – and I loved a challenge anyway – so we invested much time and money in the necessary hardware and software and self-training. I pulled out after a while, but he doggedly carried on. Eventually, on the verge of insanity, he gave up and admitted defeat.

    The good news is that he had gained enough keyboard expertise through having to correct the voice recognition software’s errors that he was able to carry on writing! I would like to be able to finish by saying that, having got rid of the frustrations and time-wasting caused by speech recognition, he went on to finish his book, which became a best-seller but… the keyboard and mouse are still a problem for a lot of people.

  32. This is without a doubt one of the best trips I have made to you site after reading about speech recognition on VISTA. Had never paid any attention until now. This is the coolest ever. I’ve trained my laptop and now it’s time for the PC to learn. Worked great on here with just the microphone built in. I have been receiving you emails and coming to your site for years now. Thanks again for making me look at what is actually on my computer.

  33. I have used Windows speech recognition for several years , I do not have Vista but use Windows XP Operating System with Office 2003 and find speech recognition invaluable. Your review gives a completely false picture of the application. The problem may be your voice! Perhaps it does favour those who have been verbal communicators in their careers .It may not suit everyone , nor is it foolproof , but then neither is typing .All e-mails and articles which I write are largely done with speech recognition. I wouldn’t be without it . Another advantage is spontaneity -you write as you talk , and this seems to give a naturalness to the printed word .The initial training sessions are important, and completing these correctly makes a huge difference.As a retired teacher who still coaches pupils, they are amazed when they see it demonstrated.

  34. In spite of WSR’s defficiencies, I am sure this experience is rooted in some wrong behavior, or a bad microphone. WSR, esp. with the Macro feature add-on (downloadable for free from Microsoft)definitely is more advanced in terms of accuracy and efficiency than Dragon NS 10. The accuracy rate can reach 99% rather than 90% as indicated. However, certain voices, or certain articulations, probably get different results from different applications. Some native speakers of English get worse results than non-natives like me, if their enunciation is of the blurred type.

  35. Vista Speech Recognition and Dragon both work well. Some of the clowns on these videos need training, the first part of which should be to learn to be smarter than the program, followed closely by learning the program itself.

  36. Speech recognition is not a game or casual application you can pick up and run out of the box. But if you have lost the full use of your hands and/or arms, or if you don’t know how to type but producing text is part of your job (think lawyer, doctor, journalist, programmer), then you will have the motivation to learn the art of speech recognition dictation. Like typing or old-fashioned dictation, it requires motivation, practice and skill. The nay-saying posters don’t seem to meet these requirements — for their sake, I hope they never reach a state where they have to meet them.

    Dictated via DNS Pro 10

  37. I’ve used Dragon Naturally Speaking 6 & 9 as well as trying to use Win Vista’s speech recognition software and pray that I’m never in a position to be totally dependent on any of the three as I can think of little else that would assure giving up computers by me. It’s too bad you can’t sue for False Advertising as it is neither “natural’ nor “speech recognized”. Might as well burned the two hundred dollars that between DNS 6 & 9 I spent.

  38. I’m a freelance translator, so speech recognition software is a good idea for me. I don’t know about Vista speech recognition- but I use Dragon Naturally Speaking 9.0 and it saves me a lot of time. Of course, you have to keep an eye on it make sure it doesn’t go crazy from time to time, but I guess I get about 95% (or better) accuracy after a bit of practice- better than my tiepin- sorry, typing.

  39. I worked on Speech Reco back in the 80s. We had two main problems at that time. They both still seem to exist. 1) We spend several hundred thousand dollars and 20 years educating the smartest thing known to mankind, that being human children, to allow them to recoginize simple commands. 2) After a few applications like disabled or dark room work there are no prevailing apps that just cry for speech reco.

  40. Microsoft made a big improvement in speech recognition with Vista. I’m a long-time user of Dragon Naturally Speaking, but I prefer Vista’s speech recognition because of its more attractive interface and how well it integrates with the operating system. It may not be quite as accurate, but on my system the difference is negligible. However, it does seem to be more sensitive to background noise than Dragon is. Perhaps the dog and the wife were contributing to its less-than-stellar accuracy. Microsoft speech recognition is one of Vista’s undiscovered gems, and was hoping your review would be more encouraging to users.

  41. Tiggerr Schleicher

    That’s odd. My 14 year old son loves it. He didn’t do anything differently than the regular set up. He does however use a headphone with a microphone attached. Maybe that is the key. He goes to a virtual school and does 3 days at home, and it’s kind of fun hearing him say Start listening. Although I will admit when I first heard Stop Listening, I thought he was hearing things. 🙂

Leave a Comment

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.