/extract_features_using_world

using world vocoder to extract features and make data for training neural networks

Primary LanguageMakefile

extract_features_using_world

  1. first, put all your wav files in wav dir;
  2. second, go to extract_features dir and extract features;
  3. third, go to copy_synthesis dir and validate your features using copy synthesis;
  4. (option)fourth, go to make_cmp dir and composite your features to cmp files (for nn training);
  5. (option)fifth, move your binary label here and make frame-aligned prepared_label and prepared_cmp(for nn training);