Link to home
Start Free TrialLog in
Avatar of dosyl
dosyl

asked on

Voice recognition

Hi, I'd like to make a pgm. that recognize the voice(in French but i don't believe the language is important), to be able to write in WordPad or something else.

I don't know where to begin. Someone can help me? If the answer is very good i'll increase the points.
Avatar of Matti
Matti
Flag of Finland image

Hi!

Tha language is important, Ms hasmighty good English tools in Speech SDK.

The recon works so that you need to say certain number of words a listed and then it will form the command's You can control program operations Like say clearly:

Start Wordpad
Do it

and so on.



Matti




 
Hi Matti!

I've been asking an even easier question than this, and I got one answer the third time that helped me.

I only asked how to read the inputlevel or outputlevel from the soundcard, and someone gave me the link to audiolvl.exe which is a sample application that can read the in and out level of any Windows SoundSystem soundcard - as far as i know. You gan eventually store the history of what comes in, and then compare it. But this is something that's VERY hard

The file is found at:
http://www.microsoft.com/downloads/release.asp?releaseid=12845

and suggest you take a look.

This was the ONLY valuable answer I got to my questions...
Avatar of dosyl
dosyl

ASKER

Matti, you told the language is important, then do i need a software or a special ActiveX for the French language?
Sorry! My comment was for Dosyl, of course! Not Matty.... Sorry  :)
ASKER CERTIFIED SOLUTION
Avatar of Matti
Matti
Flag of Finland image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of dosyl

ASKER

PPC, i looked the pgm. Audiolvl, but how can i do to recognize a WORD, that pgm. can capture the voice but how to recognize a WORD???
I only wanted to say that what you're asking of is very much. But maybe you can use this to recognize words too.

I've never tried though...

What I think this sample does is that it sums up all volume at all frequencies, and returns it as the present volume. You could maybe store the last 5 seconds of "volumes" at all time, and compare them to some stored in certain files... If difference is less than accuracy then... BANG...

I don't know - still. But this is maybe a beginning, I thought
Avatar of dosyl

ASKER

PPC, your idea is good, but very hard; i'll look the possibilities.
Have you tried using the Microsoft Agent control? - I think this includes voice recognition, and is nice and easy to use from VB.

Doing it yourself wouldn't be too hard though. Providing you have a phD in computational linguistic processing and a spare three years ;)

Good luck!
Avatar of dosyl

ASKER

andyclap i never try. Do you think it can recgnize French?
Avatar of dosyl

ASKER

Where can i find Microsoft Agent control?
Avatar of dosyl

ASKER

I found MsAgent but if  i see right: MsAgent do Text to Speech, not Speech to Text??????
Sorry if I misled you, I though it did, but the last time I used it was a few years ago. If you don't get anywere with Agent, Look into Matti's suggestion of the speeck SDK.
Avatar of dosyl

ASKER

I found the SDK and read about that, first we need W98 or more, and need VC++. I want to have a suggestion at 38 years old must i begin to learn VC++, i only learned C at school in 1992, is it too difficult??
Funny comment!!!
Hi!

We are about same age.
And the C situation is quite similar I had VC 1.0 in 93. and now I had learned VC ++ 5.0 and 6.0 about two years not do much whit the MFC side but MIDL and Win32 goes and if you are API related VB developer you have propably missed the skill of Win32 developer and this is one most addvanted feature in VC . I don't like to use it but if no other bossibility then I have to.


About the SDK
Is it version 4.0a or 5.0 beta.


Matti
Avatar of dosyl

ASKER

Thank's Matti for your comment, i'll try to find a VC++ from another guy to see if i can remember my C.

I have SDK 5.0 beta.

It seems the only way to talk to the computer: Microsoft SAPI or Dragon SAPI.
Avatar of dosyl

ASKER

I accept this question because it seems the better; i didn't try it.