Facing the problem of vocal discrimination in pop music,the authors propose applying MFCC parameters as features,and Support Vector Machine(SVM) as classifier.

• 针对流行音乐中人声的发现问题,使用SVM分类器针对MFCC特征进行训练和分类。