[hts-users:00060] Re: How the MFCC features converted to real speech wave?

["Heiga Zen (Byung-Ha Chun)" <zen@xxxxxxxxxxxxxxxx>]

Hi ShaoHuang,

ShaoHuang Pin wrote:

> I am wondering how to convert the MFCC features to speech sound, thanks.

Many people misunderstand that HTS uses MFCC for spectral parameter.
It uses "mel-cepstrum", not MFCC.
By using MLSA filter, mel-cepstrum features can be converted to speech 
waveforms.
Following papers will provide you the detail information about 
mel-cepstrum and MLSA filter.

Fukada, T., Tokuda, K., T., K., Imai, S.,
"An adaptive algorithm for mel-cepstral analysis of speech,"
Proc. of ICASSP'92, Vol. 1. pp. 137–140.

Imai, S.,
"Cepstral analysis synthesis on the mel frequency scale,"
Proc. of ICASSP'83), Vol. 1. pp. 93-96.

Best regards,

Heiga Zen (Byung-Ha Chun)

--
 ------------------------------------------------
  Heiga Zen     (in Japanese pronunciation)
  Byung-Ha Chun (in Korean pronunciation)

  Department of Computer Science and Engineering
  Graduate School of Engineering
  Nagoya Institute of Technology
  Japan

  e-mail: zen@xxxxxxxxxxxxxxxx
     web: http://kt-lab.ics.nitech.ac.jp/~zen
 ------------------------------------------------