I understand that background noise affects speech recognition negatively. If I want to automatically detect this type of noise, what should I look for in the audio samples? The ultimate goal is to inform the user to go somewhere quieter!
The following is of interest. During recognition...
1) Will a few high peaks have an effect?
2) Will a constant low/medium background sound have an effect?
3) Are there particular frequencies to look for?
4) Other properties?
Thanks
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I understand that background noise affects speech recognition negatively. If I want to automatically detect this type of noise, what should I look for in the audio samples? The ultimate goal is to inform the user to go somewhere quieter!
The following is of interest. During recognition...
1) Will a few high peaks have an effect?
2) Will a constant low/medium background sound have an effect?
3) Are there particular frequencies to look for?
4) Other properties?
Thanks
Yes.
Yes.
Model frequences are pointed in feat.params, usually between 100Hz and 6800Hz.
Echo is also harmful.
Thanks! I suppose detection of #1 and #2 should be sufficient in most cases.