Interesting ideas. I'm fond of having a voice recognition front end that several different systems/engines could be plugged into. I consulted my father about this in order to figure out where NS and VV's GUI lets him down. Unfortunatly, as I've said - I make time out of nowhere - and it was pretty unlikely that anything would come of it, so all that's left is a load of sketches and ideas for various functionality within it.. Due to my dad not being completely unable to use a keyboard, but having severe dyslexia - he's not a normal case - for one thing, he actually prefers (or prefers them for certain things) some of the older versions of voice recognition systems like VV and NS I think. So - as I was saying - he's not a normal case, so some of the ideas might not be generically useful - but I think that some of them were quite good.
I really do like the idea of a frontend - to be honest (without knowing anything) if people could apt-get Sphinx and it'd just run like NS or VV on their computer - in terms of a GUI being available, and with a training wizard - I'm sure loads of people would try it and it would generate the kind of support needed to make it decent. We mustn't forget that although we probably don't see voice recognition as a cool subject - most computer users would find it quite a cool thing to play with. When I was sketching things out though - the frontend began to seem more like a framework (pardon that word), or at least more than a frontend - with the eventual functionality it'd hope to have. I'd certainly be up for anything to do with a project like this. Also - without wanting to sound too pesimistic towards ideas that aren't my own - there's been a fair amount of criticsm aimed at VV and Sphinx - personally I find it hard to believe that there's a system that could be reverted into the copyright of one person that could hold any ground against the best that Sphinx has to offer. I'll stop there though - I am very interested to see information about this though. If it has a GUI that'd be useful, if not, it underlines the fact that we need to make one.... Chris Hayes P.S - this response has been somewhat rushed. On 23/02/07, Eric S. Johansson <[EMAIL PROTECTED]> wrote:
Eric S. Johansson wrote: > as I was constructing my response, and was almost finished when it hit > me about what's wrong with the model proposed. it is the equivalent of > raw natural text. Full function natural text sucks a little bit. The > broken, unable to correct consistently, natural text is horrible and > ruins voice models. What you're proposing has even less functionality > than a broken natural text. I'm sorry, that was too harsh. I was interrupted by one too many things while I was writing that bit and I forgot to go back and edit it. Again, I apologize for being careless. today might be a day to stay away from the keyboard unless I'm writing code. :-) --- eric -- Speech-recognition in use. It makes mistakes, I correct some. -- Ubuntu-accessibility mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-accessibility
-- Ubuntu-accessibility mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-accessibility
