Page 1 of 1
Chatterbox - STT / TTS / TTA project. Part 1
Posted: Fri 11 Oct 2013, 21:29
by greengeek
Chatterbox STT / TTS / TTA project
(Speech-To-Text / Text-To-Speech / Text-To-Action)
(Part 2 etc of this project is here)
My “chatterbox
Posted: Fri 11 Oct 2013, 21:30
by greengeek
reserved
Posted: Fri 11 Oct 2013, 21:30
by greengeek
reserved
Posted: Fri 11 Oct 2013, 21:30
by greengeek
reserved
Posted: Fri 11 Oct 2013, 21:31
by greengeek
reserved
Posted: Fri 11 Oct 2013, 22:45
by greengeek
The first method I have tried is to record a .wav file (using mhwaveedit) with the appropriate question ("Welcome to Puppy, would you like me to play your music collection randomly"). I called this Qplay.wav and placed it into /root.
Then I write a script to use aplay to play the .wav, and place that script in /root/startup.
Code: Select all
#!/bin/bash
sleep 3
aplay /root/Qplay.wav &
Works fine but I probably want to do the same thing using an ogg or mp3 instead of a wav, (so maybe there is an mp3 alternative to aplay?). I also want to be able to produce that sound bite without direct recording - ie: I want to be able to produce the wav or mp3 using a program that turns text into speech. (I tried flite but for some reason it will only produce mp3s for me - I can't get the wav side of it working yet)
The other option is to work direct from a text file and have something like espeak driven from the command line and just have that script in /root/startup. (I will give that a go once I have read the documentation...)
Posted: Fri 11 Oct 2013, 23:20
by smokey01
arecord will make a .wav file
lame will convert the .wav file to .mp3
mpg123 will play the .mp3 file
Puppy normally has the three files above included.
Some useful but old info below:
http://www.faqs.org/docs/Linux-HOWTO/MP3-HOWTO.html
I have used a command line mp3 recorder in the past but I can't remember it's name.
Posted: Fri 11 Oct 2013, 23:38
by H4LF82
ok the TTS part is already done for us.
I have had lots of success with espeak on Lucid
http://www.murga-linux.com/puppy/viewtopic.php?t=25046
Check it out!
Put a script in the /root/startup directory named Qplay and fill it with the following code...
Code: Select all
#!/bin/sh
espeak -f /path/to/Qplay.txt
exit
Save it.
Posted: Sat 12 Oct 2013, 17:31
by greengeek
Yes, that espeak method seems to work quite well. Did you try any other voices?
Posted: Sat 12 Oct 2013, 17:41
by H4LF82
i have not tried other voices yet.
Posted: Sat 12 Oct 2013, 18:43
by greengeek
smokey01 wrote:arecord will make a .wav file
lame will convert the .wav file to .mp3
mpg123 will play the .mp3 file
Puppy normally has the three files above included.
Thanks Smokey. I seem to lack the mpg123 in Pemasu's Upup3992 so will try some other pups aswell.
Posted: Sat 12 Oct 2013, 21:36
by smokey01
greengeek,
madplay will also play mp3 files and is generally part of most Puppy distros.
Home page for mpg123 if you want it.
http://mpg123.org/
Download binary or source:
http://mpg123.org/download.shtml
Posted: Sun 13 Oct 2013, 06:32
by greengeek
Thanks - yep, madplay does the trick on Upup. Definitely a space saver to be able to use mp3 rather than wav.
Video of talking computer helping disabled
Posted: Thu 17 Oct 2013, 13:52
by Ted Dog
http://www.ksat.com/news/ut-medicine-ta ... index.html
After some phone calls and replies they have posted the interview of the main doctor I've spoken with / short demo-ed the puppylinux offerings.
They show the commercial 'pricy' tools to support ALS but low cost and free would help the other 95% of people without ALS.
WHY I'M I HAVING PROBLEMS WITH BLANK MSG, URLS, CAN SOME MOD TAKE ME OFF THE WATCH LIST!!!!
Posted: Tue 29 Oct 2013, 18:41
by greengeek
Just a reminder to myself to do some more work on flite and to investigate svox-pico too