In cepstral speech analysis, it is relatively easy to understand why subtracting the mean removes any "channel" effects (including that of recording equipment and the vocal tract). There is a good explanation of that here. But why does dividing by the variance (after the mean has been subtracted) also give a similar (additional) improvement?

Subtraction in the quefrency-domain corresponds to de-multiplication in the frequency domain, which corresponds to de-convolution in the time domain. Can a similar analogy or explanation be given for division in the quefrency-domain?

  • $\begingroup$ Are you trying to say that CVN on it's own is giving similar results to the CMN? Or do you mean CMVN? $\endgroup$ – jojek Oct 20 '15 at 15:51
  • $\begingroup$ No. I edited the question to be a bit clearer. I admit the question is two-fold, but the main thing I am after is the last part: To understand the time-domain and/or frequency-domain meaning of dividing the cepstrum by its variance. $\endgroup$ – rudolfbyker Oct 21 '15 at 9:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.