Nonlinear Kronecker product filtering for multichannel noise reduction

Itzhak, Gal; Benesty, Jacob; Cohen, Israel

doi:10.1016/j.specom.2019.10.001

Cited by 11 publications

(8 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It should be noted that the formulation of (9) was already suggested in [20] in the context of multichannel noise reduction in the frequency domain. However, in this work, the quadratic approach is applied to a singlechannel observation vector in an arbitrary linear filtering domain, in which the interframe correlation is considered.…”

Section: Quadratic Maximum Snr Filtermentioning

confidence: 99%

“…However, in this work, the quadratic approach is applied to a singlechannel observation vector in an arbitrary linear filtering domain, in which the interframe correlation is considered. Additionally, while the optimal filters suggested in [20] are designed to minimize the squared output energy and may be seen as the quadratic approach counterparts of the conventional MVDR and LCMV, this work provides a more general perspective to derive quadratic filters and proposes the quadratic maximum SNR filter h max as a special case.…”

Section: Quadratic Maximum Snr Filtermentioning

confidence: 99%

“…The modified version is obtained from the Kronecker product of the observations' vector and its complex conjugate. In its mathematical formulation, this approach is similar to the approach presented in [20] in the context of multichannel noise reduction. On the contrary, while in [20] the essence of the innovation is the direct utilization of higher-order statistics, the key idea in this work is a generalization of the single-channel linear filtering approach.…”

Section: Introductionmentioning

confidence: 99%

“…In its mathematical formulation, this approach is similar to the approach presented in [20] in the context of multichannel noise reduction. On the contrary, while in [20] the essence of the innovation is the direct utilization of higher-order statistics, the key idea in this work is a generalization of the single-channel linear filtering approach. We demonstrate that by focusing on the estimation of the desired signal magnitude in the transform domain, we are able to achieve further reduction of the background noise.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Quadratic approach for single-channel noise reduction

Itzhak

Benesty

Cohen

2020

J AUDIO SPEECH MUSIC PROC.

Self Cite

View full text Add to dashboard Cite

In this paper, we introduce a quadratic approach for single-channel noise reduction. The desired signal magnitude is estimated by applying a linear filter to a modified version of the observations' vector. The modified version is constructed from a Kronecker product of the observations' vector with its complex conjugate. The estimated signal magnitude is multiplied by a complex exponential whose phase is obtained using a conventional linear filtering approach. We focus on the linear and quadratic maximum signal-to-noise ratio (SNR) filters and demonstrate that the quadratic filter is superior in terms of subband SNR gains. In addition, in the context of speech enhancement, we show that the quadratic filter is ideally preferable in terms of perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI) scores. The advantages, compared to the conventional linear filtering approach, are particularly significant for low input SNRs, at the expanse of a higher computational complexity. The results are verified in practical scenarios with nonstationary noise and in comparison to well-known speech enhancement methods. We demonstrate that the quadratic maximum SNR filter may be superior, depending on the nonstationary noise type.

show abstract

Section: Quadratic Maximum Snr Filtermentioning

confidence: 99%

Section: Quadratic Maximum Snr Filtermentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Quadratic approach for single-channel noise reduction

Itzhak

Benesty

Cohen

2020

J AUDIO SPEECH MUSIC PROC.

Self Cite

View full text Add to dashboard Cite

show abstract

“…This paper considers introducing some very lowcomplexity schemes to suppress the artificial residual noise components, so that speech quality can be improved at a very low cost. It is well-known that, compared with DNN-based methods, many conventional monaural speech enhancement have much lower computational complexity [27][28][29][30], and their performance is highly dependent on the estimation accuracy of the noise PSD.…”

Section: Introductionmentioning

confidence: 99%

Low-complexity artificial noise suppression methods for deep learning-based speech enhancement algorithms

Zheng

et al. 2021

J AUDIO SPEECH MUSIC PROC.

View full text Add to dashboard Cite

Deep learning-based speech enhancement algorithms have shown their powerful ability in removing both stationary and non-stationary noise components from noisy speech observations. But they often introduce artificial residual noise, especially when the training target does not contain the phase information, e.g., ideal ratio mask, or the clean speech magnitude and its variations. It is well-known that once the power of the residual noise components exceeds the noise masking threshold of the human auditory system, the perceptual speech quality may degrade. One intuitive way is to further suppress the residual noise components by a postprocessing scheme. However, the highly non-stationary nature of this kind of residual noise makes the noise power spectral density (PSD) estimation a challenging problem. To solve this problem, the paper proposes three strategies to estimate the noise PSD frame by frame, and then the residual noise can be removed effectively by applying a gain function based on the decision-directed approach. The objective measurement results show that the proposed postfiltering strategies outperform the conventional postfilter in terms of segmental signal-to-noise ratio (SNR) as well as speech quality improvement. Moreover, the AB subjective listening test shows that the preference percentages of the proposed strategies are over 60%.

show abstract

Differential constant-beamwidth beamforming with cube arrays

Itzhak

Cohen

2023

Speech Communication

View full text Add to dashboard Cite

Nonlinear Kronecker product filtering for multichannel noise reduction

Cited by 11 publications

References 18 publications

Quadratic approach for single-channel noise reduction

Quadratic approach for single-channel noise reduction

Low-complexity artificial noise suppression methods for deep learning-based speech enhancement algorithms

Differential constant-beamwidth beamforming with cube arrays

Contact Info

Product

Resources

About