I'd like to know if and/or how your system can perform the following:
* Voiceprint Authentication - eg. Voicevault and Nuance VocalPassword
* Out of Grammar utterances - eg., a non word spoken sound (grunt, cough), a non human noise (bang, clank), or out of grammar spoken word. This is kinda like VAD? What I would like to do is to detect the user speaking or making some noise but it is out of grammar. Furthermore, be able to set the trigger level (volume) of this noise to filter out potential more distant and lower volume sounds.