Proyecto EDCAV Audio

Singing Voice Separation

Singer Classification

Singer classification using pitch and mfcc. Implemented on MATLAB 2018a. Using Audio System Toolbox.

List Files:

• feature extraction list file (list of files to extract features from): featureExtractionListFile.txt
• training list file (list of files to train from): trainListFile.txt
• test list file (list of files to test from): testListFile.txt
• classification output file (classification result): outputListFile.txt
• list of (used) dataset singers: singers.txt

Functions:

• extract_feat: extract pitch and mffc from audio file
• train_model: trains classification model from features
• classif: classify from features using trained model
• main: script to call functions