SVSGAN: Singing Voice Separation Via Generative Adversarial Network | IEEE Conference Publication | IEEE Xplore