AudioExp

Audio stream parsing experiments using Mel-Frequency Cepstral Coefficients (MFCCs) and machine learning.

Training

See 'usage' section in TrainKerasTensorFlowForTaggedSounds.py.

Visualization

Compare waveform in amplitude domain to MFCCs graphically. See 'usage' section in PlotWav.py.

WAV to MFCC CSV

Convert a wav file to a .csv file containing MFCCs using WavToMfccFeatures.py.

Finding samples in a .wav file

When you have a trained model, use AnalyzeSoundStream.py.