As I know, we can get the cepstrum by taking the Inverse Discrete Fourier Transform (IDFT) of the log of the power spectrum.

enter image description here

enter image description here

And we can get the Mel-frequency cepstral coefficients (MFCCs) by taking the Discrete Cosine Transform (DCT) of the log filterbank energies.

But I didn't understand the following two points:

  • Why do we use the term "cepstral" in MFCCs when we use DCT instead of IDFT?

  • How does MFCCs relate to the cepstral domain according to the reference here?

  • $\begingroup$ Can you please have a look at this as cited in this and then perhaps amend the question with further points that may have remained unclear? $\endgroup$
    – A_A
    Jan 28, 2020 at 10:56
  • 1
    $\begingroup$ Take a look here. DCT can be treated as an IDFT in this case. $\endgroup$
    – jojeck
    Jan 28, 2020 at 11:16


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.