/Mel-Cepstral-Distortion

Calculation of MCD (dB) between two speech waveforms

Primary LanguageJupyter Notebook

Mel-Cepstral Distortion

This repo is concerned with finding the Mel Cepstral Distortion between .wav files.

Dependencies

  • librosa
  • pyworld
  • pysptk

Installation

Tested on a Python verison 3.6.2 in linux VM environment

conda create -n mel-cep-distortion python=3.6
conda install jupyter
conda install -c conda-forge librosa
pip install pyworld=0.2.11
pip install pysptk=0.1.18

Usage

See the Jupyter notebook for example usage

TODO:

  • Update dependencies and packages used in README, more detailed installation
  • Include scripts on extracting acoustic audio features