Vista Speech Recognition and the latest System Access Beta

For most of this week, I have run a new beta of System Access almost exclusively on my Vista computer.  I’ve stuck to System Access because “Monster” Matt Campbell, over a few days hacking last week, has created the best out-of-the-box solution for the Vista Speech Recognition system.  I have not tried J-Vist yet and I expect it will perform well as everything else that I’ve tried that Brian has written works great. 

 

Window-Eyes and JAWS without J-Vist work poorly in the various Speech Recognition dialogues and not at all when dictating into Word 2007.  With Both WE and JAWS, I could complete the speech recognition tutorial and perform some of the training tasks.  For both JAWS and Window-Eyes a user must use the JAWS Cursor or Window-Eyes Mouse Cursor extensively to get the speech recognition features to talk at all.  In System Access, Monster Matt has, excepting fairly minor bugs (remember I did say it is beta software) really made speech recognition a pleasure to use.

 

When dictating into Word 2007 with either JAWS or Window-Eyes, absolutely nothing gets read back to the user.  Setting screen echo to “all” (a technique that works for me in Word 2003 with JAWS and Dragon) didn’t cause JAWS to read back any of the information I had dictated and in Notepad and WordPad caused JAWS and Window-Eyes to speak far too much and I assume that Vista must be repainting the edit window far more often than one would expect.

 

With this System Access beta, the text I dictate is read back after the recognition engine processes the information and I can tell when the recognition system has made a mistake and correct it.  The correction dialogue works very nicely with SA as does the spelling dialogue in which a user can speak the correct spelling of a word to the system and, from which, the recognition software’s accuracy will improve.

 

Using JAWS, I issued the verbal “correct that” command and, with the JAWS cursor, I could find the correction dialogue laying atop my MS Word file with its text intertwingled with the text in my document.  One can sort of use JAWS if they don’t mind issuing a lot of SayLine keystrokes to hear what they have dictated and poking around a lot with the JAWS cursor to find the correction dialogues.  It is nearly impossible to use JAWS or Window-Eyes as a hands free solution with Vista Speech Recognition; System Access can be used very nicely and, excepting some odd situations, an SA user can go almost entirely hands free after the recognition system is trained well enough to understand your voice.

 

So, System Access, the lowest priced screen access program, without the user needing to purchase any additional software, outperforms the two leading screen readers by a substantial margin in the Vista speech recognition system.  For me, having a hands free system that I can use when my RSI problems flare up badly, is a major improvement to an otherwise difficult situation.

 

Once the Serotek guys post this as an update, I recommend people give it a try and, if you are patient (a virtue required of all speech recognition users), I will bet you find it impressive.

 

I would also like to tip my hat to the guys at Microsoft for doing a terrific job with the voice recognition features in Vista.  Having faithfully used it for a few days now, I would say that it performs similarly to Dragon Naturally Speaking and SA users will soon have access to it by default.

 

— End

 

 

Subscribe to the Blind Confidential RSS Feed at: http://feeds.feedburner.com/ Blindconfidential

Published by

chris.admin

I'm an accessibility advocate working on issues involving technology and people with print impairment. I'm a stoner, crackpot, hacker and all around decent fellow. I blog at this site and occasionally contribute to Skepchick. I'm a skeptic, atheist, humanist and all around left wing sort. You can follow this blog in your favorite RSS reader, and you can also view my Twitter profile (@gonz_blinko) and follow me there.

7 thoughts on “Vista Speech Recognition and the latest System Access Beta”

  1. Ever thought of trying a brain computer interface. It’s not invasive and BCI systems are starting to become more viable. I did think of combining BCI and reinforcement theory for your favourite CEO once but that’s a story best told in private.

  2. Hi Chris

    This is becoming a regular thing, me contributing to your blog, I hope that is OK.

    I read your comments concerning the System Access support for Vista speech recognition. In fact I discovered this update purely by accident the day previous to your posting. I think some of the things Serotek have done are quite interesting and are delivered within the spirit of the System access philosophy – what you hear is what you get. The output from System Access is quite automatic which can have advantages and disadvantages. I think people have to remember, as you Chris have pointed out, that JAWS gives you an enormous number of tools to customise the output you hear, and I don’t mean through scripting. I mean with tools like the keyboard manager, virtual viewer, etc. To that end, J-Vist follows through with that philosophy by having a series of dictation modes for various situations or no text echo at all for those of us who use Braille, which is very important for many of us who do read it. Anyway, we wish them well with the update.

    To broaden this out a little though, I am interested in why people feel that Freedom Scientific should have supported Windows Vista speech at all? Not on this blog I have to say, but many people have criticised the fact that Freedom Scientific have not supported Windows Vista speech “out of the box”.

    I see Vista speech similar to the Sonar product. Like the good guys from Dancing Dots, with J-Vist we will be providing very good context sensitive help which is virtual viewer based, and an in depth manual in a variety of formats including DAISY. Local support is also essential from people who have experience in this field, especially in terms of quality microphone provision and so on.

    Are we really saying that Freedom Scientific should be providing this kind of input? Isn’t it better for the customer if products within specific areas of technology are properly supported both after purchase and within the documentation accompanying them? I think on the scale of priorities Freedom Scientific I am sure have other applications they need to support to serve a greater user base and that Speech Rec would rank fairly low at the moment.

    Again, thanks for the blog.

  3. Hi Brian,
    i think people are criticizing FS for not supporting dictation in Vista because it is a feature built into the operating system, and people shouldn’t have to spend extra money to give them support to built-in OS functions. While it probably wasn’t high on the priority list for the reasons you mentioned, they should at least provide a free update or something. Not that I’m trying to criticize you or the scripts, but i think screen readers should provide full access to the OS, and people shouldn’t have to buy third-party scrripts to access functions that other screen readers give for free.

  4. Brian Hartgen makes some great points. We applaud Brian’s work across the board. Brian Hartgen has been a long time, totally unbiased, supporter of Jaws For Windows, WindowEyes, and System Access. We find his reports to be informative and well produced. T And T is also a reseller of Serotek products. While I will leave my opinions about development priorities for other AT developers out of this post, I do agree with Brian that voice recognition is not something that any AT company would put at the top of it’s list. It certainly was not at the top of ours at Serotek, but, our customers wanted some kind of voice recognition experience before our final FUI application is released next year, so we gave it to them.

    Brian has done a great job on this product and I agree with his post 100%. The truth is that, I, as a blind person: am excited to see so many choices for blind AT users. I don’t want any company to go out of business or lose it’s market share. why? Overall we have so many blind folks to serve, that we have enough blinks to go around world wide for all the AT companies and then some.

    I think it is important for a blind person to take the time to look at all the AT products on the market and make an informed choice. This should be done by new users and veteran users as well. The mainstream computer market is changing every day so yesterday’s AT solution may or may not work for you now. We don’t need to take any business from any AT company because, besides the thousands of people that have attempted to learn AT and a computer and given up, we have millions of elderly users coming in to this space within the next 5 to 10 years.

    While many state agencies have a cookie cutter AT package that they give to the uninformed, most states will give the client the solution they ask for. This is important to know if you prefer one technology over another. If they don’t give you the package you want, after you research it, then contact the client assistant program for that state agency and I assure you that they will give you the product you are asking for. The key here is that the blind consumer push and push and push some more, until they get what they need instead of what the state thinks they do.

    We, don’t believe that if you are a secretary working with Word, Outlook, IE, and other more common tools of the trade, you really need an AT solution that is going to take hours and hours of training on features that you may or may not ever use. However, if you are working in custom applications that require set files or scripting TODAY, then you may want another solution. Make informed choices folks. Demand explanations as to why one AT package is being purchased exclusively if that is the case in your state. When you hear the reasons compare what you hear to what you have learned and then make your choice. Stop allowing agencies for the blind to make your AT choices. While I agree to some point with BC’s view about needing several AT solutions if you want the most access, let’s remember that BC is a hacker at heart and wants to do the most he can do. For many people having more than one AT solution is information overload. Hence, Brian Hartgen’s view of our product is correct. We call it automatic as apposed to stick. Hey, nobody will argue that stick gives you more control over a car, but, nothing beats automatic for ease of use. Both methods will get you where you want to go and interestingly enough once you learn stick or auto you usually get to where you are going in the same amount of time no matter what you choose. Well, I am assuming the people that understand that have driven. grin.

    So in conclusion, hats off to Brian Hartgen and T and T. They have pulled another rabbit out of the hat and I am sure that between both products, users at each end of the scale will enjoy a rich experience without feeling overwhelmed.

  5. Howdy Comrades!
    This is quite intriguing, but what do cars have to do with the price of tea in China? Onward through the fog!
    Regards,’
    Chairman Mal
    Power to the Peeps!

  6. I totally agree with Mike Calvo on the point about consumer choice! I held a receptionist position at a nonprofit organization not too long ago, and one of my duties was to type memos and other items for the staff, from dictation. I was paid only for part of the time, but that’s another story. I’ll just say it had nothing to do with me personally or my job performance. Anyway, when I first got this job, I needed a screen reader to perform my duties. The nonprofit was kind enough to provide a laptop which I could use. I am and have always been a JAWS user by trade and thus have felt most comfortable with that screen reader. I told this to my VR counselor at the time, but apparently he didn’t take that to heart because a demo version of Window-Eyes was purchased for me. Not that WE is a bad screen reader, but I had no knowledge of how to use it. Fortunately, an adaptive tech instructor who was working at the Chicago Lighthouse at the time was kind enough to give me a Braille cheat sheet. Not only was I at a disadvantage because I was only working with a demo copy of WE, but the laptop ended up dying and needed to be replaced. I forget how this happened, but I eventually got a full copy of JAWS like I had requested. Somebody had messed up the settings but everything ended up working fine. Moral of my story: Don’t always rely on the rehab agency for your employment needs. For some VR may in fact work quite well, but for me and for many others it doesn’t work at all. Meantime, keep up the great work Mike, Matt, and everybody else at Serotek!

Leave a Reply to Jake Cancel reply

Your email address will not be published. Required fields are marked *