
Technical discussions related to Speech Coding (all itu and other vocoders, ACELP, CELP, AMR, etc)
|
Dear all: I found that when a female speech was coded by G.723.1, the speech sounds more "low" (high frequency components degradation). However, this situation is less noticeable in the case of male speakers. In fact, I had tried GSM and G.729 before, and female speakers always get more low quality outputs than male. I am not sure but I think that this "curse" may be caused by the over- simplification of the residual signal (or fixed-codebook in the case of CELP) because residual take care of the more unpreditable part of speech (high frequency components). Am I right or there is another explanation? Thanks in advance! Koro Chen |