VoxForge
Hi nsh,
>Massepur is doing this robot with sphinx:
Freaking amazing! Now that is how a robot should perform... :)
The competition here in Canada has a lot of catching up to do! All I saw was robot mini-Sumo, line following and 'firefighting' (i.e. searching a 10*10 foot maze and putting out a candle in a 'room').
Ken
Hi nsh,
amazing!!!, great work...seems that we have lot to do here at monobotics, but yes we are also doing our stuff all in linux and hopefully soon add some speech recognition capabilites...
congratulations ..
Mauricio
please dime si vez mi archivo de sesion, la pagina me indica que termino pero no lo veo en la lista, tal vez es un tema de actualizacion...lo subi como un attachment en:
http://www.voxforge.org/home/audacity
saludos
Mauricio
Hi Ken,
Mauricio tells that in some of the recordings there is a "chick" noise at the beginning of his phrases (may be noise generated by his keyboard). Could this noise afect to the training process?
Could it be removed in some easy way (may be some program that let you remove the first two seconds of a wav).
Thanks in advance
Hi Ivan,
>Could this noise afect to the training process?
Yes.
If you don't mark it as noise in your transcription (i.e. prompt line), your acoustic models might expect certain words to contain a click sound in them... and thus decrease your recognition accuracy.
>"chick" noise at the beginning of his phrases [...]
>Could it be removed in some easy way
Two approaches to address this:
trim start [length]
Trim can trim off unwanted audio from the beginning and end of the audio. Audio is not sent to the output stream until the start location is reached.
Ken
Hi Ubanov,
I created a new Trac repository (http://www.dev.voxforge.org/projects/es) and Subversion repository (http://www.dev.voxforge.org/svn/es/) where I committed your code.
I also sent you email on how to login to these so you can make changes.
If anyone else needs access to either of these, please let me know.
thanks,
Ken