Hi, i found this:
http://perlbox.org/
Look @ the features, it's quite interesting. It uses festival, sphinx2
and has (for now) a kde plugin only. I didn't manage to force sphinx
use /dev/dsp2 (my own tux mic entry) instead of /dev/dsp. It could be
an interesting base for tuxdroid, even if it's a tk interface (mm...
no antialiasing :p Anybody using amsn will know what i'm talking
about) and it is perl..
I tried testing xvoice, which seems quite interesting too (but uses
ibm's viavoice, sphinx support not ready afaik): it allows diction.
It's features are great, but i don't know how usable it is.
I quickly evaluated espeak, which gives good results (english though),
festival (i didn't try the good models...) which is undoubltly less
working than acapela, and others i could'nt make work
I tested the mic ( aplay /dev/dsp2 ), and there's this constant noise,
but apparently sphinx2 is robust against parasites. This one is
constant, which is good, but when the motors light up.... We can't
hear a thing apart of this. Plus, mouth has to be wide open to get
better sound.
Interesting links:
Lots of testing TODO :)
http://www.linux.com/howtos/Speech-Recognition-HOWTO/software.shtml
http://www.cstr.ed.ac.uk/projects/festival/mbrola.html (the "good"
models for festival; non-free though)
French resources:
http://kubuntu.free.fr/blog/index.php/2006/09/24/121-synthese-vocale-en-francais-sous-linux
KDE: commandes vocales http://kubuntu.free.fr/blog/index.php?p=43
See you next week.
BTW: why is'nt the wiki used anymore?
|
