Recent trends and challenges in speech-separation systems research &amp;#x2014; A tutorial review

Ananthakrishnan, K.S.; Doğançay, Kutluyıl

doi:10.1109/tencon.2009.5396022

Cited by 7 publications

(5 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This means that the typical pitch range for both male and female from 80 to 500 Hz or time lags from 2 to 12.5 msnow corresponds to 50 to 313 samples, i.e. pitch state space which consists of the union of 0, 1, and 2 pitch hypothesis, since this algorithm tracks up to 2 pitches simultaneously,is described by (1). HMM search to be biased towards S 2 .…”

Section: Hmm Multipitch Trackingmentioning

confidence: 99%

See 1 more Smart Citation

A Hybrid Approach for Co-Channel Speech Segregation based on CASA, HMM Multipitch Tracking, and Medium Frame Harmonic Model

Eldin¹,

Youssif²

2013

IJACSA

View full text Add to dashboard Cite

Abstract-This paper proposes a hybrid approach for cochannel speech segregation. HMM (hidden Markov model) is used to track the pitches of 2 talkers. The resulting pitch tracks are then enriched with the prominent pitch. The enriched tracks are correctly grouped using pitch continuity. Medium frame harmonics are used to extract the second pitch for frames with only one pitch deduced using the previous steps. Finally, the pitch tracks are input to CASA (computational auditory scene analysis) to segregate the mixed speech. The center frequency range of the gamma tone filter banks is maximized to reduce the overlap between the channels filtered for better segregation. Experiments were conducted using this hybrid approach on the speech separation challenge database and compared to the single (non-hybrid) approaches, i.e. signal processing and CASA. Results show that using the hybrid approach outperforms the single approaches.Keywords-CASA (computational auditory scene analysis); cochannel speech segregation; HMM (hidden Markov model) tracking; hybrid speech segregation approach; medium frame harmonic model; multipitch tracking, prominent pitch.

show abstract

Section: Hmm Multipitch Trackingmentioning

confidence: 99%

“…Competing speech is the most difficult kind of interference because of the similarity of temporal and spectral characteristics between target and interfering speeches. Work on speech segregation dates back to 70s [1].…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Approach for Co-Channel Speech Segregation based on CASA, HMM Multipitch Tracking, and Medium Frame Harmonic Model

Eldin¹,

Youssif²

2013

IJACSA

View full text Add to dashboard Cite

show abstract

“…Monaural source separation or single-channel source separation works with two learning methods, supervised learning (models can use previous experience to produce outcomes), and unsupervised learning (models do not have previous experience). Existing review articles describe supervised single-channel speaker separation algorithms in either signal processing [21]- [23] or in the time-frequency [24], [25] domains. The conventional single channel speaker separation techniques such as computational auditory sense analysis (CASA) [26], non-negative matrix factorization (NMF) [27], [28] in the signal processing domain and deep learning-based deep clustering (DC) [29], deep attractor networks (DANet) [30], permutation invariant training (PIT) [31], in T-F domain have been reviewed in [32].…”

Section: Introductionmentioning

confidence: 99%

State-of-the-Art Analysis of Deep Learning-Based Monaural Speech Source Separation Techniques

2023

View full text Add to dashboard Cite

The monaural speech source separation problem is an important application in the signal processing field. But recent interaction of deep learning algorithms with signal processing achieves remarkable performance improvement for speech source separation problems. This paper explores the numerous state-of-the-art deep learning-based monaural speech source separation algorithms in the timefrequency (T-F), time, and hybrid domains. The motivation, algorithm, and framework of different deep learning models for monaural speech source separation are analyzed. The benchmarked algorithms in the T-F domain can be categorized as deep neural networks (DNN), clustering, permutation, multi-task learning, computational auditory sense analysis (CASA), and phase reconstruction-based techniques, whereas the state-of-the-art time-domain approaches can be categorized as CNN, RNN, multi-scale fusion (MSF), and transformer-based techniques. The end-to-end post filter (E2EPF) is a hybrid algorithm combining T-F and time-domain works to achieve enhanced results. Time-domain models have shown improvement in separation performance compared to the T-F and hybrid domain models with small model sizes. Methods in T-F, time, and hybrid domains are compared using 𝑆𝐷𝑅, 𝑆𝐼 − 𝑆𝐷𝑅, 𝑆𝐼 − 𝑆𝑁𝑅, PESQ, and 𝑆𝑇𝑂𝐼 as quality assessment metrics on some benchmark datasets.

show abstract

“…Speech enhancement allows for the extraction of the desired speech signal from a mixture of speech with interfering sounds coming from different sources. These methods can be used in hearing aids [ 1 ], smartphones [ 2 ], or, as a pre-processing step, in automatic speech or speaker recognition [ 3 ].…”

Section: Introductionmentioning

confidence: 99%

Speech Enhancement by Multiple Propagation through the Same Neural Network

Grzywalski

Drgas

2022

Sensors

View full text Add to dashboard Cite

Monaural speech enhancement aims to remove background noise from an audio recording containing speech in order to improve its clarity and intelligibility. Currently, the most successful solutions for speech enhancement use deep neural networks. In a typical setting, such neural networks process the noisy input signal once and produces a single enhanced signal. However, it was recently shown that a U-Net-based network can be trained in such a way that allows it to process the same input signal multiple times in order to enhance the speech even further. Unfortunately, this was tested only for two-iteration enhancement. In the current research, we extend previous efforts and demonstrate how the multi-forward-pass speech enhancement can be successfully applied to other architectures, namely the ResBLSTM and Transformer-Net. Moreover, we test the three architectures with up to five iterations, thus identifying the method’s limit in terms of performance gain. In our experiments, we used the audio samples from the WSJ0, Noisex-92, and DCASE datasets and measured speech enhancement quality using SI-SDR, STOI, and PESQ. The results show that performing speech enhancement up to five times still brings improvements to speech intelligibility, but the gain becomes smaller with each iteration. Nevertheless, performing five iterations instead of two gives additional a 0.6 dB SI-SDR and four-percentage-point STOI gain. However, these increments are not equal between different architectures, and the U-Net and Transformer-Net benefit more from multi-forward pass compared to ResBLSTM.

show abstract

Recent trends and challenges in speech-separation systems research — A tutorial review

Cited by 7 publications

References 25 publications

A Hybrid Approach for Co-Channel Speech Segregation based on CASA, HMM Multipitch Tracking, and Medium Frame Harmonic Model

A Hybrid Approach for Co-Channel Speech Segregation based on CASA, HMM Multipitch Tracking, and Medium Frame Harmonic Model

State-of-the-Art Analysis of Deep Learning-Based Monaural Speech Source Separation Techniques

Speech Enhancement by Multiple Propagation through the Same Neural Network

Contact Info

Product

Resources

About