Has anyone thought about implementing voice/speech recofnition in the Z? A 'speech to text' application would be very useful I'd think, especially when the Z (atleast my 5600) has such excellent voice recording capabilities.
Unfortunately my own programming knowledge = 0
Therefore I am not capable of developing anything for the Z at all, but am willing to be a beta tester.
Anyone interested?

Hi!

I have been using such a prog for two monthes. It was "Dragon Naturally Speaking" under win xp. It is not that usefull because of so many recognition errors... I finally learned how to use a keaboard...

I have two objections :
1) can a 400mhz/64mo ram be enough? my "old" 850 mhz with 128 mo wasn't (too long sentences made it "forget" some words.)
2) is it really usefull? I mean if your Z can record, why don't you play your recorded voice and have it recognizedby a voice recording recognition on your pc?

Cheers.

flite generates speech, Sphinx recognises speech.

There are binaires available for both.

Have a play.


Si

lardman, i've been googling but i couldn't figure out whether there is a sphinx binary for the z. sorry if i missed an obvious place, thx

The instructions for downloading and compiling Sphinx can be found at http://www.speech.cs.cmu.edu/cgi-bin/cmusp...ldAndRunSphinx4
Is anyone interested in compiling this for the Z?
I'd have tried but my programming expertise can comfortably fit in a nutshell with the nut and room to spare.

yes, there is work being done on something like this. I can't say who, or what, or when, if ever, it will get released.

That's good to know. If it is ever released, I hope there'll be a version for the 5600 (since that's the one I have).

I've already compiled it (about a year ago in fact).

http://sgp.zaurii.net/binaries/sphinx2-0.4_arm.ipk
http://sgp.zaurii.net/binaries/sphinx2-0.4_arm.tar.gz
http://sgp.zaurii.net/binaries/sphinx3-0.1_arm.ipk
http://sgp.zaurii.net/binaries/sphinx3-0.1_arm.tar.gz

Not sure about the packaging.

Enjoy.


Si

Wow, this sounds cool.

lardman, how well does it work on a C860? I guess you have to wangle some way of getting a mic set up for the Z for this to work though.

Hey Lard man ,
I'm playing around with this....thanks for the info

I have had no joy so far....can you give some basic advice to get us started? It would be a great help.

Thanks ads

Ps c750 Qtopia cacko rom
ads

woooooooooow!

I MUST give this a try...

Does it work with pdaxrom?

As far as I know it has more chance of working with pdaX than anything else but I'm using Qtopia and so far no joys...

I really wanna get it going so I can use my zaurus to launch and control programs using only speech....

ie mp3 player, flite for reading books and some database retieval ( address retrieval) without having to pull out my zaurus

Its a pretty distant dream but if I can get the building blocks then I'll be happy.

ads

@ads: To tell the truth I don't know.

If I remember correctly, sphinx is mostly just the libs which allows you to create your own apps which can do speech recognition. That said, there are a couple of example apps in the tarball (I think). I just compiled it as someone had asked me to, I'm not sure that I ever got round to using it myself, though I think others did. Assuming the devnet is back up (and has its old data) you could try a search there.

I'll have a look at it during the week and see whether I can remember (but you're probably better off just reading the release notes/README/INSTALL files in the tarball as then you'll know pretty much as much as I do).

@alan:

It's command line so it ought to.


Si

Lardman, thanks for that....I have had a look at the release notes, and it gets kinda involved....I was looking for an easy solution.

I'll have a real go at it later when I get the time but it looks like a weekender to get it all going.
Thanks for all your help.

Ads