D3Net: Densely connected multidilated DenseNet for music source separation

5 Oct 2020  ·  Naoya Takahashi, Yuki Mitsufuji ·

Music source separation involves a large input field to model a long-term dependence of an audio signal. Previous convolutional neural network (CNN)-based approaches address the large input field modeling using sequentially down- and up-sampling feature maps or dilated convolution. In this paper, we claim the importance of a rapid growth of a receptive field and a simultaneous modeling of multi-resolution data in a single convolution layer, and propose a novel CNN architecture called densely connected dilated DenseNet (D3Net). D3Net involves a novel multi-dilated convolution that has different dilation factors in a single layer to model different resolutions simultaneously. By combining the multi-dilated convolution with DenseNet architecture, D3Net avoids the aliasing problem that exists when we naively incorporate the dilated convolution in DenseNet. Experimental results on MUSDB18 dataset show that D3Net achieves state-of-the-art performance with an average signal to distortion ratio (SDR) of 6.01 dB.

PDF Abstract


Results from the Paper

Ranked #6 on Music Source Separation on MUSDB18 (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Music Source Separation MUSDB18 D3Net SDR (vocals) 7.80 # 4
SDR (drums) 7.36 # 3
SDR (other) 5.37 # 4
SDR (bass) 6.20 # 7
SDR (avg) 6.68 # 6
SDR (vocals) 7.24 # 8
SDR (drums) 7.01 # 8
SDR (other) 4.53 # 11
SDR (bass) 5.25 # 16
SDR (avg) 6.01 # 11