my experience :
you should bandlimit the input to speech bandwidth for more consistant
results when using dirty speech (speech with some occassional background
noises, voices with lots of bass or sibilance)
I use 200-3000Hz 1dB, -40dB at < 140, > 3500 Hz)
Don't forget that microphone frequency responses are very bumpy and
lumpy !
A 3 band parametric Q=0.7 EQ can assist to optimise a specific voice.
-glen Vk1XX
_______________________________________________
Freetel-codec2 mailing list
Freetel-codec2@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
