Fig. 3From: Comparative study of singing voice detection based on deep neural networks and ensemble learningSCNN structure. The input is a spectrogramBack to article page