Demucs: Deep Extractor for Music Sources with extra unlabeled data remixed
arXiv:1909.01174
Abstract
We study the problem of source separation for music using deep learning with four known sources: drums, bass, vocals and other accompaniments. State-of-the-art approaches predict soft masks over mixture spectrograms while methods working on the waveform are lagging behind as measured on the standard MusDB benchmark. Our contribution is two fold. (i) We introduce a simple convolutional and recurrent model that outperforms the state-of-the-art model on waveforms, that is, Wave-U-Net, by 1.6 points of SDR (signal to distortion ratio). (ii) We propose a new scheme to leverage unlabeled music. We train a first model to extract parts with at least one source silent in unlabeled tracks, for instance without bass. We remix this extract with a bass line taken from the supervised dataset to form a new weakly supervised training example. Combining our architecture and scheme, we show that waveform methods can play in the same ballpark as spectrogram ones.
References in corpus (7)
- WaveNet: A Generative Model for Raw Audio
- Progressive Growing of GANs for Improved Quality, Stability, and Variation
- Deep Scattering Spectrum
- Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders
- SampleRNN: An Unconditional End-to-End Neural Audio Generation Model
- Wave-U-Net: A Multi-Scale Neural Network for End-to-End Audio Source Separation
- Towards Unsupervised Single-Channel Blind Source Separation using Adversarial Pair Unmix-and-Remix
Cited by in corpus (7)
- The Cone of Silence: Speech Separation by Localization
- Complex Spectral Mapping With Attention Based Convolution Recurrent Neural Network for Speech Enhancement
- CatNet: music source separation system with mix-audio augmentation
- Transfer Learning with Jukebox for Music Source Separation
- Semi-Supervised Singing Voice Separation with Noisy Self-Training
- Using Python for Model Inference in Deep Learning
- GSEP: A robust vocal and accompaniment separation system using gated CBHG module and loudness normalization