Masking-based β-order MMSE speech enhancement

You, Chang Huai; Koh, Soo Ngee; Rahardja, Susanto

doi:10.1016/j.specom.2005.05.012

Cited by 23 publications

(9 citation statements)

References 24 publications

(47 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Consequently, the estimation of the relative threshold offsets obtained using (18) and/or (20) are also affected. This is because the residual noise of the preliminary estimated speech may severely change the original tonality of the speech signal [19]. As a result of the inaccurate relative threshold offset, the estimation error of the noise masking threshold increases.…”

Section: Modified Relative Threshold Offset and Noise-masking Threshomentioning

confidence: 99%

Perceptual improvement of Wiener filtering employing a post-filter

Alam

O’Shaughnessy

2011

Digital Signal Processing

View full text Add to dashboard Cite

Section: Modified Relative Threshold Offset and Noise-masking Threshomentioning

confidence: 99%

Perceptual improvement of Wiener filtering employing a post-filter

Alam

O’Shaughnessy

2011

Digital Signal Processing

View full text Add to dashboard Cite

“…Conversely, if the difference between the two is larger than the back-fitting threshold value, then the KBF process of step 6 is iterated until convergence. Applying the auxiliary function approach [20] to a quadrature form of the spectrogram gradients, the back-fitting threshold value is calculated at each iteration.…”

Section: ) Determination Of Music/voice Enhancementmentioning

confidence: 99%

“…In single-channel speech enhancement methods based on WbE, the cost function includes both a power law β and a weighting factor α. To obtain a significant noise reduction and an improved speech estimation of weak speech components both in terms of objective and subjective measures, adaptive calculation methods of parameters α and β, of the cost function, have been proposed and published in the literature [19]- [21].…”

Section: Appropriate Calculation Of β J and α Jmentioning

confidence: 99%

Music/Voice Separation Based on Kernel Back-fitting Using Weighted ?-order MMSE Estimation

Kim¹,

Kim²

2015

ETRI J

View full text Add to dashboard Cite

Recent developments in the field of separation of mixed signals into music/voice components have attracted the attention of many researchers. Recently, iterative kernel back-fitting, also known as kernel additive modeling, was proposed to achieve good results for music/voice separation. To obtain minimum mean square error (MMSE) estimates of short-time Fourier transforms of sources, generalized spatial Wiener filtering (GW) is typically used. In this paper, we propose an advanced music/voice separation method that utilizes a generalized weighted β-order MMSE estimation (WbE) based on iterative kernel back-fitting (KBF). In the proposed method, WbE is used for the step of mixed music signal separation, while KBF permits kernel spectrogram model fitting at each iteration. Experimental results show that the proposed method achieves better separation performance than GW and existing Bayesian estimators.

show abstract

“…A typical speech enhancement system [7,10,14,15,21,22,78,79,80] that exploits masking properties is described in Fig. 2.6.…”

Section: The Speech Receiver Modelmentioning

confidence: 99%

“…10 shows an example of the IS curves and the average IS value for a noisy speech signal.The frame length was set to 240 samples with an overlap of 180 samples and a sampling rate of 8kHz. In practice, the IS distortion is obtained by averaging the smallest 85% ∼ 95% of IS values over all the frames of the speech signal under examination.…”

mentioning

confidence: 99%

Speech enhancement methods based on masking properties

You¹

Self Cite

View full text Add to dashboard Cite

I would like to express my sincere gratitude to my supervisor, Professor Koh Soo Ngee, and my co-supervisor, Associate Professor Rahardja Susanto, for their invaluable guidance and pertinent advices during my study. Moreover, I would like to thank their constant encouragement, their enormous time and effort that they have made pertaining to this research and thesis writing. I am deeply grateful to my parents, Mr. You Bao Hui and Ms. Zheng Hui Yu, and my wife, Jiang Ping, and my two sons, Li Ming and Li Hong, for their endless love, support and encouragement. I would also like to thank my final year project students from National University of Singapore, Ms. Chen Kexin, Mr. Tan Chun Wee and Ms. Sem Pak Ling, for sharing with me their projects and many interesting and helpful discussions.

show abstract

Masking-based β-order MMSE speech enhancement

Cited by 23 publications

References 24 publications

Perceptual improvement of Wiener filtering employing a post-filter

Perceptual improvement of Wiener filtering employing a post-filter

Music/Voice Separation Based on Kernel Back-fitting Using Weighted ?-order MMSE Estimation

Speech enhancement methods based on masking properties

Contact Info

Product

Resources

About