Multi-Channel Deep Clustering: Discriminative Spectral and Spatial Embeddings for Speaker-Independent Speech Separation

By Zhong-Qiu Wang, Jonathan Le Roux, John R. Hershey. 2018

253 citations recorded by the source corpus.

Subfields: Computational Mechanics, 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 Computational Mechanics

1. Jinyu Li
2. Zhuo Chen
3. Sanyuan Chen
4. Shujie Liu
5. Yu Wu
6. Takuya Yoshioka
7. Gordon Wichern
8. Tomohiro Nakatani

In Signal Processing

1. Jinyu Li
2. Zhuo Chen
3. Shinji Watanabe
4. Takuya Yoshioka
5. Dong Yu
6. Nima Mesgarani
7. Yi Luo
8. Marc Delcroix

Citations in the corpus

• Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation (2018, background): “Previous studies have shown the benefit of extending speech separation to multichannel inputs [67], [68], [69], particularly in adverse acoustic conditions and when the number of interfering speakers is large (e.”
• Continuous Speech Separation with Conformer (2020, background): “com And in [20, 21], the multi-channel extension separation has been shown to be effective.”
• Continuous Speech Separation: Dataset and Analysis (2020, background): “On top of, or aside from, these algorithmic improvements, researchers also sought more effective input features [17, 18] and model architectures [19, 20, 21].”
• Bootstrapping Single-channel Source Separation via Unsupervised Spatial Clustering on Stereo Mixtures (2018, methodology): “Our training, validation and test data are from the publicly available(1) spatialized version of the Wall Street Journal mix dataset with two speakers (wsj0-2mix) [6, 28].”
• Phase-aware Speech Enhancement with Deep Complex U-Net (2019, background): “Another important direction is to extend the proposed model to deal with multichannel audio since accurate estimation of phase is even more critical in multichannel environments (Wang et al., 2018).”
• SpeakerBeam: Speaker Aware Neural Network for Target Speaker Extraction in Speech Mixtures (2019, methodology): “For different experiments, we report results for four different versions of the dataset, namely WSJ0-2mix [10] for single-channel 2-speaker experiments, WSJ0-3mix [10] for single-channel 3-speaker experiments, WSJ0-2mix-MC [48] for multi-channel experiments and our own modification of WSJ0-2mix, WSJ”
• Separate Anything You Describe (2023, unknown): “Besidesthemapping-based approach,deepclustering-basedmethods[32],[33]havealso shownpromisingresultsonspeechseparationbasedonthe learningofarepresentationspacewithcontrastiveobjectives.”