Filterbank Design for End-to-end Speech Separation
By Manuel Pariente, Samuele Cornell, Antoine Deleforge, Emmanuel Vincent. 2019
75 citations recorded by the source corpus.
Single-channel speech separation has recently made great progress thanks to learned filterbanks as used in ConvTasNet. In parallel, parameterized filterbanks have been proposed for speaker recognition where only center frequencies and bandwidths are learned. In this work, we extend real-valued learned and parameterized filterbanks into complex-valued analytic filterbanks and define a set of corresponding representations and masking strategies. We evaluate these filterbanks on a newly released noisy speech separation dataset (WHAM). The results show that the proposed analytic learned filterbank consistently outperforms the real-valued filterbank of ConvTasNet. Also, we validate the use of parameterized filterbanks and show that complex-valued representations and masks are beneficial in all conditions. Finally, we show that the STFT achieves its best performance for 2 ms windows.
Subfields: Artificial Intelligence, Signal Processing
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Yossi Adi
- Lior Wolf
- Eliya Nachmani
- Manuel Pariente
- Samuele Cornell
- Joris Cosentino
- Emmanuel Vincent
- Antoine Deleforge
In Signal Processing
- Yossi Adi
- Lior Wolf
- Eliya Nachmani
- Samuele Cornell
- Manuel Pariente
- Antoine Deleforge
- Joris Cosentino
- Emmanuel Vincent
Citations in the corpus
- Voice Separation with an Unknown Number of Multiple Speakers (2020, methodology): “, 2018b), Learnable filter bank (Pariente et al., 2020), and DPRNN.”
- LibriMix: An Open-Source Dataset for Generalizable Speech Separation (2020, background): “We see that for two-speaker mixtures, Conv-TasNet outperforms ideal masks in clean conditions and is on par with them in noisy conditions, as in [4,32].”
- Wavesplit: End-to-End Speech Separation by Speaker Clustering (2020, unknown): “We similarly remixed WHAMR!, Model ∆ SI-∆ SDR SDR Chimera++ (Wichern et al., 2019) 9.9 – Conv-TasNet (Pariente et al., 2019) 12.7 – Learnable fbank (Pariente et al., 2019) For instance, the results on WHAMR! are comparable to the reconstruction accuracy from clean inputs (WSJ0-2mix) with models prio”