Deep Learning for Talker-Dependent Reverberant Speaker Separation: An Empirical Study

Delfarah, Masood; Wang, DeLiang

doi:10.1109/taslp.2019.2934319

Cited by 22 publications

(12 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dereverberation algorithms such as the weighted prediction error (WPE) method [16] are crucial for good performance. Systems for combined separation and dereverberation of a target speaker have been proposed in [17,18]. A recurrent neural network for joint separation and dereverberation is trained to learn the clean target signal in the time-frequency domain from the reverberant mixture [17].…”

Section: Introductionmentioning

confidence: 99%

On End-to-end Multi-channel Time Domain Speech Separation in Reverberant Environments

Zhang

Zorilă

Doddipatla

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

This paper introduces a new method for multi-channel time domain speech separation in reverberant environments. A fullyconvolutional neural network structure has been used to directly separate speech from multiple microphone recordings, with no need of conventional spatial feature extraction. To reduce the influence of reverberation on spatial feature extraction, a dereverberation pre-processing method has been applied to further improve the separation performance. A spatialized version of wsj0-2mix dataset has been simulated to evaluate the proposed system. Both source separation and speech recognition performance of the separated signals have been evaluated objectively. Experiments show that the proposed fully-convolutional network improves the source separation metric and the word error rate (WER) by more than 13% and 50% relative, respectively, over a reference system with conventional features. Applying dereverberation as pre-processing to the proposed system can further reduce the WER by 29% relative using an acoustic model trained on clean and reverberated data.

show abstract

Section: Introductionmentioning

confidence: 99%

On End-to-end Multi-channel Time Domain Speech Separation in Reverberant Environments

Zhang

Zorilă

Doddipatla

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…The main motivation is that jointly separating and enhancing may be too difficult for a single network to learn, and modularization may allow the networks to focus on specific tasks. Two-stage approaches have previously been explored for denoising plus dereverberation [19,20], separation plus dereverberation [21], and denoising plus separation [11].…”

Section: Cascaded Modelsmentioning

confidence: 99%

WHAMR!: Noisy and Reverberant Single-Channel Speech Separation

Maciejewski

Wichern

McQuinn³

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

123

View full text Add to dashboard Cite

While significant advances have been made in recent years in the separation of overlapping speech signals, studies have been largely constrained to mixtures of clean, near-field speech, not representative of many real-world scenarios. Although the WHAM! dataset introduced noise to the ubiquitous wsj0-2mix dataset, it did not include the addition of reverberation, generally present in indoor recordings outside of recording studios. The spectral smearing caused by reverberation can result in significant performance degradation for standard deep learning-based speech separation systems, which rely on spectral structure and the sparsity of speech signals to tease apart sources. To address this, we introduce WHAMR!, an augmented version of WHAM! with synthetic reverberated sources, and provide a thorough baseline analysis of current techniques as well as novel cascaded architectures on the newly introduced conditions.

show abstract

“…When W is fixed, each element of S j will be fixed at s j (f, n) = w H j (f )x(f, n). Now, since the terms that depend on z j and c j in (16) are given as…”

Section: Fmvae Algorithm a Ideamentioning

confidence: 99%

“…Meanwhile, given the recent advances achieved by deep VOLUME 4, 2016 neural network (DNN)-based speaker separation methods, including deep clustering (DC) [9], [10] and permutation invariant training (PIT) [11], [12], a discriminative approach has recently proved powerful in monaural source separation tasks, including both speaker-dependent and -independent scenarios [13]- [16]. The general idea is to train a DNN that predicts TF masks or TF embeddings from a given mixture signal based on spectro-temporal features.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

FastMVAE: A Fast Optimization Algorithm for the Multichannel Variational Autoencoder Method

Kameoka

Inoue

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Deep Learning for Talker-Dependent Reverberant Speaker Separation: An Empirical Study

Cited by 22 publications

References 34 publications

On End-to-end Multi-channel Time Domain Speech Separation in Reverberant Environments

On End-to-end Multi-channel Time Domain Speech Separation in Reverberant Environments

WHAMR!: Noisy and Reverberant Single-Channel Speech Separation

FastMVAE: A Fast Optimization Algorithm for the Multichannel Variational Autoencoder Method

Contact Info

Product

Resources

About