Credit: Music Audio Processing Workshop by HKUST Note: semantic info