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.