EURASIP Journal on Audio, Speech, and Music Processing

Table 1 Data sets used in the experiments

From: High level feature extraction for the self-taught learning algorithm

GTZAN database			ISMIR 2004 database
Set	Pieces	Hours	Set	Pieces	Hours
GT-50	500	0.69	IS-20	120	0.17
GT-100	1000	1.39	IS-50	300	0.42
GT-250	2500	3.47	IS-100	600	0.83
GT-500	5000	6.95	IS-250	1500	2.08

Back to article page