Author Topic: Speach Recognition and Voice Synthisis?  (Read 4440 times)

Mega

  • Sr. Member
  • ****
  • Posts: 300
  • Who GNU?
Re: Speach Recognition and Voice Synthisis?
« Reply #15 on: November 28, 2010, 08:14:00 am »
see Clippy wouldn't be the Clippy we know today if not for his daughters

Would the Clippettes be called Subs of Clippy.bas as the main module, compiled miniature clippies with similar code, or clippy include files? I'm sure that all the PEEKs and POKEs over the years can make any program run out of stack space with some people, but like every good declared variable, Clippy still has absolute value. :)

James
Just an old-school programmer of covox speech things and such...

Mega

  • Sr. Member
  • ****
  • Posts: 300
  • Who GNU?
Re: Speach Recognition and Voice Synthisis?
« Reply #16 on: November 28, 2010, 10:44:04 am »
By the way, a few years back I saw an o. source speech SDK for linux on sourceforge. I never did try it out, but they said it was comparable to Microsoft's SpeechSDK5.1 which is pretty good. Although for the Win32 side you could probably be able to make use of API calls through a raw/ring0 asm loader to pass things to QB64 with the Microsoft Speech SDK, that would totally leave out the Linux port and any future ports that could (and probably should) use a universal SDK implementation. I know it would be a ways off, but it may be possible to use that standard GPL'd lib speech recognition package as a foundation for calls to audio input hardware of the machine you're using if there's enough interest in it...

Just an old-school programmer of covox speech things and such...

iamdenteddisk

  • Hero Member
  • *****
  • Posts: 2735
Re: Speach Recognition and Voice Synthisis?
« Reply #17 on: February 21, 2012, 01:45:56 am »
ummm, perhaps he could colaborate with me on my naomi project?
it has all the AI and text in speech out, also most of the SR stuff Is done but I just need a bit of help with the actual wav comparison "confirming words" the rest is there, allready working..


iamdenteddisk

  • Hero Member
  • *****
  • Posts: 2735
Re: Speach Recognition and Voice Synthisis?
« Reply #18 on: February 21, 2012, 01:53:10 am »
here is a link to that project and the update to current version will be up ASAP..
http://dl.dropbox.com/u/24429764/NAOMI_Project.zip

Robert Claypool

  • Sr. Member
  • ****
  • Posts: 463
Re: Speach Recognition and Voice Synthisis?
« Reply #19 on: March 20, 2012, 09:15:33 pm »
iamdenteddisk please take a look at http://www.qb64.net/forum/index.php?topic=5640.0 where I have a set of speech routines using similar if not the same wav files, but uses the QB64 sound play system which allows for music to be played under the voice or even voices.

iamdenteddisk

  • Hero Member
  • *****
  • Posts: 2735
Re: Speach Recognition and Voice Synthisis?
« Reply #20 on: March 31, 2012, 11:10:44 am »
realy nice, I will check that out Robert. I used standard wav files and the intent was to use sync for that..

I have found out how to use the mic and stuff aswell as found a sample of file comparison. I am now working on how to do FFT with the compare sub, so I can then test first for silence and then recognise words are anything not silence. then comparison of those input words to set of pre recorded words to identify the meaning of input..

if you have any ideas or advice in this area any input will be greatly appreciated.. 

MajorHeadrush

  • Newbie
  • *
  • Posts: 41
Re: Speach Recognition and Voice Synthisis?
« Reply #21 on: January 31, 2016, 08:00:29 am »
Actually im just trying to find a way to integrate an existing speech recog package into the program

Dragon was a pience of crap. I don't know if it got any better, I see it is still around. I still remember my wife's little brother trying to teach Dragon to print swear words back in 1990-something. It didn't do it worth a damp!

Pete

MajorHeadrush

  • Newbie
  • *
  • Posts: 41
Re: Speach Recognition and Voice Synthisis?
« Reply #22 on: January 31, 2016, 08:10:41 am »
I am just trying to integrate the microsoft speech recog package that i got online and is freeware into my program... im not much of a programmer out side of quick basic... the software seems a lot like c... has calls.. passes variables etc... i thought someone with more expertise could have a look at it.

Heres where i got started.

https://msdn.microsoft.com/en-us/library/office/hh378466(v=office.14).aspx


would be thoroughly impressive if anyone got a speech recognition engine off the ground all i managed to pull off myself was a stupid clap detector which was set off by any short duration sound... does qb64 support microphone input yet ??? anyway for this I would suggest looking into the methods others hve used in other languages... this site might help you you will need a large speech library for any language and these guys are working on an open source library for several (spoken) languages http://www.voxforge.org/ hope that helps you.
btw would love to see your original source
welcome to QB64 feel free to look around there re a large (uncounted) number of programs on this site that are just floating around and I hope that the size of some of the larger topics does not intimidate you as long as the comments are:
1 constructive/ helpful
2 asking for help/understanding
3 funny
they are valued

keybonicplague

  • Full Member
  • ***
  • Posts: 212
  • I'll think of something to put here eventually.
Re: Speach Recognition and Voice Synthisis?
« Reply #23 on: January 31, 2016, 03:23:30 pm »
Quote
I wrote a program a while back in 1984
lol!!!  I was 1!!

lol. I was -1. His program predates me.

Robert Claypool

  • Sr. Member
  • ****
  • Posts: 463
Re: Speach Recognition and Voice Synthisis?
« Reply #24 on: February 02, 2016, 05:02:49 pm »
I was born in 1976, QB45 has a CALLS statement and C has a pascal statement for handling code from different languages, QB64 uses DECLARE statements to achieve most of the same effect, but I think they miss too many edge cases.

MajorHeadrush

  • Newbie
  • *
  • Posts: 41
Re: Speach Recognition and Voice Synthisis?
« Reply #25 on: May 03, 2016, 02:15:52 am »
keep in mind this isnt about writing speech recognition software... this is about finding some that a qb64 program can call and use. As far as performance you have to be careful. I thought this new software windows 7 has sucked but then i stopped trying to enunciate... when i would just speak normally not trying to speak in a way i thought would help the program all the sudden its recog jumped to like 99% accuracy... i even tried speaking a little fast and it did really well and this was without training the software which my program has no time for.

the Julius software looks like a good second candidate... both it and the ms one are open source as of this post.

the key is making the jump from info and packages available for free and hopefully open source in case we need to tinker with the code to actually implimenting it in a qb64 program... hopefully mine. So again here is where im at in my progress.... what i didnt like was the way microsoft installed countless programs on my computer to do this. I really want to get it down to just the program that does the recog... something that will stand alone in the programs directory and translate the speech into text.

so heres where im studying the ms stuff

https://msdn.microsoft.com/en-us/library/office/hh378380(v=office.14).aspx
https://msdn.microsoft.com/en-us/library/office/hh378466(v=office.14).aspx
https://msdn.microsoft.com/en-us/windows/uwp/input-and-devices/speech-recognition

next i will look into Julius which does seem like it might actually work better for me than this unnecesarily complicated ms install

http://julius.osdn.jp/en_index.php

MajorHeadrush

  • Newbie
  • *
  • Posts: 41
Re: Speach Recognition and Voice Synthisis?
« Reply #26 on: December 07, 2016, 05:46:44 pm »
ive made several planned improments to convo... it gets more like talking to a person you know with every advancement.. i love that... but i still have not come up with a way for it to take speech and turn it into text for convo to use... still having to type everything in...                                 

MajorHeadrush

  • Newbie
  • *
  • Posts: 41
Re: Speach Recognition and Voice Synthisis?
« Reply #27 on: January 10, 2017, 02:17:15 am »
I was born in 1976, QB45 has a CALLS statement and C has a pascal statement for handling code from different languages, QB64 uses DECLARE statements to achieve most of the same effect, but I think they miss too many edge cases.

I basicly hacked my way around the problem... instead of trying to find a way to call on microsofts voice recognition in the program (which i still hope to do but when i finally find an expert to help me) i simply use it to create a file, save it and then tell convo that it is there... convo then accesses it... it works just fine... hopefully it will someday just listen and reply but in the mean time this is kind of working for me

Pete

  • Hero Member
  • *****
  • Posts: 7205
  • Cuz I sez so varmint!
Re: Speach Recognition and Voice Synthisis?
« Reply #28 on: January 10, 2017, 10:20:50 am »
Quote
I wrote a program a while back in 1984
lol!!!  I was 1!!

lol. I was -1. His program predates me.

I was 1 when this thread started!  ;D

Well kuDOS for your Windows 10-acity.

Pete
Give a man a program, and he does one thing for a day. Teach a man to program, and he does one thing for a lifetime.