A unified model for zero-shot music source separation, transcription and synthesis.
Liwei LinGus XiaQiuqiang KongJunyan JiangPublished in: ISMIR (2021)
Keyphrases
- unified model
- source separation
- audio features
- music emotion recognition
- independent component analysis
- blind source separation
- audio visual
- low level
- visual features
- music information retrieval
- temporal structure
- denoising
- single channel
- music retrieval
- spatio temporal
- audio signal
- simple linear
- signal processing
- feature space