1. pip install -r requirements.txt
2. Have cd01 and phase1 directories

python utils.py

1. Numpy arrays of short speaker segments are in the features directory.
2. Catalog identifies a speaker identity to the path to the mfcc features for a speech segment by the speaker.