Table 6.

Average classification performance per feature for training and testing the convolutional neural network from scratch over 100 training epochs.

Feature	Training scratch CNN		Testing scratch CNN
	Average accuracy	Standard deviation	Average accuracy	Standard deviation
SPEC	0.77	± 0.09	0.65	± 0.06
Chroma	0.63	± 0.06	0.55	± 0.03
MFCC	0.68	± 0.07	0.64	± 0.06
MelSpectrum	0.80	± 0.11	0.67	± 0.07
PowerSPEC	0.84	± 0.13	0.67	± 0.04
RAW	0.77	± 0.15	0.57	± 0.03
Tonal	0.69	± 0.13	0.53	± 0.02
All features	0.67	± 0.07	0.60	± 0.02