A Consolidated Perspective on Multimicrophone Speech Enhancement and Source Separation

By Sharon Gannot, Emmanuel Vincent, Shmulik Markovich Golan, Alexey Ozerov. 2017

609 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

In Signal Processing

Citations in the corpus

Single Channel Voice Separation for Unknown Number of Speakers Under Reverberant and Noisy Settings (2020, background): “Over the years, many attempts have been made to tackle this separation problem considering single microphone [1, 2], multiple microphones [3, 4], supervised and unsupervised learning [5, 6].”
Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation (2018, background): “porate multiple input audio channels may prove advantageous when more than one microphone is available. Previous studies have shown the beneﬁt of extending speech separation to multichannel inputs 10 [67], [68], [69], particularly in adverse acoustic conditions and when the number of interfering spe”
Neural Target Speech Extraction: An overview (2023, methodology): “Fixed beamforming [2], [16]* 1985 — — ✓ — ✓ — ✓”
Global and Local Simplex Representations for Multichannel Source Separation (2020, background): “M ULTICHANNEL BLIND audio source separation (BASS) aims at extracting the individual source signals from multi-microphone recordings of a mixture with several concurrently active speakers [1].”
WHAMR!: Noisy and Reverberant Single-Channel Speech Separation (2019, methodology): “Re-verberation times were chosen to approximate domestic and classroom environments [9] (as we expect these to be similar to the restau-rants and coffee shops where the WHAM! noise was collected), and further classiﬁed as high, medium, and low reverberation based on a qualitative assessment of the…”
Multi-Channel Deep Clustering: Discriminative Spectral and Spatial Embeddings for Speaker-Independent Speech Separation (2018, methodology): “Our method is shown to give much better separation performance in the stereo case in terms of signal-to-distortion ratio (SDR) than the oracle multi-channel Wiener ﬁlter (MCWF) [9], model-based EM source separation and localization (MESSL) [18], and GCC-NMF [19] algorithms.”
ReVISE: Self-Supervised Speech Resynthesis with Visual Input for Universal and Generalized Speech Enhancement (2022, neutral): “As a standard practice in multichannel speech processing, we consider beamformed audio that is derived from the four non-binaural channels as input [5, 13, 18].”
FaSNet: Low-Latency Adaptive Beamforming for Multi-Microphone Audio Processing (2019, background): “Beamforming , also known as spatial ﬁltering, is a powerful micro-phone array processing technique that extracts the signal-of-interest in a particular direction and reduces the effect of noise and rever-beration from a multi-channel signal [1].”
Speech Processing for Digital Home Assistants: Combining signal processing with deep-learning techniques (2019, methodology): “From these covariance matrices, the beamformer coefficients of most common beamformers can be readily computed [21].”