voice recognition to text
 

Hi, I don't have a N800 yet, but I have been watching since the 770 and am very fascinated by its potential.

I wondered if there is an application for converting voice, from the mic to text on the screen.

My elderly father is really hard of hearing, he does have hearing aids, but I never know if he really hears me. He can still read though<G>.

Is there an application for this device to translate my voice into text, through the microphone, for him to see. I figure I would face the N800 toward him.

thanks for any help

Re: voice recognition to text
 

While there is a text-to-speech engine (flite) available, what you want is the other way around ... a speech recognition app. I have not heard of one for the N800. Unfortunately, I would suspect that there won't be since speech recognition is pretty CPU intensive and the N800 isn't a huge number cruncher. You never know though. Someone might find an efficient enough algorithm that matches the capability of the N800 so I would never say never.

Re: voice recognition to text
 

Doing some digging around, it looks like there is something that would be appropriate is someone could do some integration work for the N800 (or maybe even the 770.

http://www.speech.cs.cmu.edu/pocketsphinx/

Re: voice recognition to text
 

It's important not to forget that the 770 and N800 are *Internet* tablets. In the first instance if I were developing something like this I'd go for creating a service I could connect to - think media server style. Much easier to do the development that way. It might even create a better product - you could speak into one 770 and your father could read on another.

It might be worth suggesting such a product/service to these folks - http://www.spinvox.com/

Re: voice recognition to text
 

IBM did it on my Pentium 75 with 64 MB of RAM, ten years ago, with a software-only solution. Granted, that was OS/2, so the rest of the world will probably have to wait another decade.

So, no: voice recognition is not that CPU-intensive (these days). It is, however, quite algorithm-intensive, which seems to be what is lacking in the postmodern world.

Re: voice recognition to text
 

Well I have Sphinx compiled.. now its just a matter of figuring this blasted thing out and pointing the 64Mb speech base lib to the mmc... more news soon :)

Re: voice recognition to text
 

Ok this sucker is more time-intensive than I am willing to spend. If anyone wants to help out post to this thread. We need a Sphinx-expert to configure the speech templates and a /dev/dsp pocketsphinx_continuous script.

Re: voice recognition to text
 

The section of that website that caught my attention was:

Since the 770 and N800 seem to capture audio at 8000 Hz sampling rate (based on the maemorecorder abilities), this voice model might be the way to go. Plus it is only 8 MB or so rather than the 25 MB that you speak of.

They are available here:
http://www.speech.cs.cmu.edu/pockets...linterp.tar.gz

Re: voice recognition to text
 

This is what caught my eye on the Carnegie-Mellon site (http://cmusphinx.sourceforge.net/html/cmusphinx.php):
"Note however that Sphinx is not a final product. Those with a certain level of expertise can achieve great results with the versions of Sphinx available here, but a naive user will certainly need further help. In other words, the software available here is not meant for users with no experience in speech, but for expert users."

Aren't we in over our heads here?

(BTW, the phrase "users with no experience in speech" is kinda funny)

Re: voice recognition to text
 

http://www.kiecza.net/daniel/linux/c...ndex.html#toc5 another option but there are some compilation issues related to the mixer.