VisorCentral.com (http://discussion.visorcentral.com/vcforum/index.php)
- Springboard Modules (http://discussion.visorcentral.com/vcforum/forumdisplay.php?forumid=10)
-- voice recognition module (http://discussion.visorcentral.com/vcforum/showthread.php?threadid=5033)

I think this one is inevitable if possible to make. The module could work with the microphone and create text on the screen as quickly as you speak. Eliminating the need for a stylus and competing effectively against Clarion's new voice controlled AutoPC

Would that be asking too much out of code that needs to fit into a 32MB limit. Going by the size of the code of the voice recognition systems on PC that are available in the market today, I would guess that a good reliable voice recognition system will need more than the available code size on the Visor and the Springboard. My guess is that VR is still possible but not with a lot of accuracy. I would stand to be corrected.
The only VR system I tried on a PC is the "Naturally Speaking" package from Dragon. My first impression was good but my friend was not very happy with it.
I would love to lazily talk to my PalmOS device instead of writing.

PalmOS-er.

Dragon NaturallySpeaking is a great product. Starts out pretty darn accurate, and the more you use it, the more accurate it gets. And our social worker here w/ cerebral palsy is incredibly more productive after we installed Dragon on his machine.

Dragon also puts out a product called NaturallySpeaking Mobile. It comes complete w/ a little digital dictating device. You speak into the device (Doctor's notes, etc.) and bring the device back to your PC, plug it in via a serial cable, and the PC downloads the digitized audio and runs voice recognition on it.

So, what we need is some way to do the voice recording on our Visor and download the voice file to the PC for processing. Perhaps a converter from "whatever audio format the Visor w/ springboard creates" to "whatever audio format NatSpeak needs".

It's going to be tough to do more than limitted-vocabulary speech recognition directly on the Visor. (NatSpeak requires a high-end Pentium and scads of ram.)

But the hybrid system has real possibilities.


------------------

john

[This message has been edited by john (edited 10-07-1999).]