Separation of multiple speech sources by recovering sparse and non-sparse components from B-format microphone recordings

Jia, Maoshen; Sun, Jundai; Bao, Changchun; Ritz, Christian

doi:10.1016/j.specom.2017.12.010

Cited by 23 publications

(7 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, a highly accurate DOA estimation and sources counting approach for multiple sound sources localization by conducting the DOA estimation in these SSZs has been proposed in Reference [ 28 ]. A soundfield microphone consists of four co-located microphones placed at the four non-adjacent corners of a cube [ 30 ] which are referred to as Front Left Up (FLU), Front Right Down (FRD), Back Left Down (BLD), and Back Right Up (BRU) microphones, respectively. The raw signals recorded by soundfield microphone are called A-format signal i.e.,

, where n and k represent the frame number and the frequency index, respectively.…”

Section: Proposed Methodsmentioning

confidence: 99%

Multiple Sound Sources Localization with Frame-by-Frame Component Removal of Statistically Dominant Source

Jia

Bao

et al. 2018

Sensors

Self Cite

View full text Add to dashboard Cite

Multiple sound sources localization is a hot topic in audio signal processing and is widely utilized in many application areas. This paper proposed a multiple sound sources localization method based on a statistically dominant source component removal (SDSCR) algorithm by soundfield microphone. The existence of the statistically weak source (SWS) among soundfield microphone signals is validated by statistical analysis. The SDSCR algorithm with joint an intra-frame and inter-frame statistically dominant source (SDS) discriminations is designed to remove the component of SDS while reserve the SWS component. The degradation of localization accuracy caused by the existence of the SWS is resolved using the SDSCR algorithm. The objective evaluation of the proposed method is conducted in simulated and real environments. The results show that the proposed method achieves a better performance compared with the conventional SSZ-based method both in sources localization and counting.

show abstract

, where n and k represent the frame number and the frequency index, respectively.…”

Section: Proposed Methodsmentioning

confidence: 99%

Multiple Sound Sources Localization with Frame-by-Frame Component Removal of Statistically Dominant Source

Jia

Bao

et al. 2018

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…A high MUSHRA score means better listening quality. The reference methods were the separation algorithm based on independent component analysis (ICA) [10], the separation algorithm based on expectation maximization (EM) [36], the joint sparse and non-sparse components separation method (JSNCS) [14], and the separation method based on linearly constrained minimum variance beamforming (LCMV) [9].…”

Section: Experiments Conditionsmentioning

confidence: 99%

“…The test results are shown in Figure 10. Condition "Proposed" is the proposed separation method; condition "JSNCS" is the reference method proposed in [14]. Condition "LCMV" represents the separation method proposed in [9].…”

Section: Objective Evaluationmentioning

confidence: 99%

“…Subsequently, improved methods have been proposed, which focus on both TF components dominated by one sound source (i.e., sparse components) and the other remaining TF components (i.e., non-sparse components). The separation method proposed in [14] uses a clustering algorithm with a dynamic threshold for the sparse component separation. Then, the non-sparse component is recovered by using the "local-zone stationarity" assumption.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Multi-Source Separation Approach Based on DOA Cue and DNN

Jia

Pai

2022

Applied Sciences

Self Cite

View full text Add to dashboard Cite

Multiple sound source separation in a reverberant environment has become popular in recent years. To improve the quality of the separated signal in a reverberant environment, a separation method based on a DOA cue and a deep neural network (DNN) is proposed in this paper. Firstly, a pre-processing model based on non-negative matrix factorization (NMF) is utilized for recorded signal dereverberation, which makes source separation more efficient. Then, we propose a multi-source separation algorithm combining sparse and non-sparse component points recovery to obtain each sound source signal from the dereverberated signal. For sparse component points, the dominant sound source for each sparse component point is determined by a DOA cue. For non-sparse component points, a DNN is used to recover each sound source signal. Finally, the signals separated from the sparse and non-sparse component points are well matched by temporal correlation to obtain each sound source signal. Both objective and subjective evaluation results indicate that compared with the existing method, the proposed separation approach shows a better performance in the case of a high-reverberation environment.

show abstract

“…Some statisticalaproperties of the sourcesaprovide a base for separation of sound sources [14]. The common statistical assumption made in our model are that the sources are www.ijew.io statisticallyaindependent, statisticallyaaorthogonala and nonstationarya [7][5] [15].…”

Section: Overview Of Proposed Algorithmmentioning

confidence: 99%

Speech Sources Separation Based on Models of Interaural Parameters and Spatial Properties of Room

Israr¹,

Khan²,

Khan³

2020

IJEW

View full text Add to dashboard Cite

This paper presents the extended evaluation in different reverberant scenarios for different mixtures of speech having interfere at one of the six angles {15 0 , 30 0 , 45 0 , 60 0 , 75 0 , 90 0 } of the model which implementing spatial covariance with interauralaparameter for improving the performance of MESSL.Thetbinaural spatialtparameters, such as interauralaphase difference IPD and interauralalevel difference ILD and spatialacovariance areamodeled in the short-timeaFourieratransform. The parametersaof the model areaupdated with the expectation-maximization algorithm. The performance of the model is checked in term of Signal-to-distortion ratio (SDR) and the perceptual evaluation of speech quality (PESQ), and the results confirmed that the performance of this proposedamodel is improved in highly reverberant rooms.

show abstract

Separation of multiple speech sources by recovering sparse and non-sparse components from B-format microphone recordings

Cited by 23 publications

References 23 publications

Multiple Sound Sources Localization with Frame-by-Frame Component Removal of Statistically Dominant Source

Multiple Sound Sources Localization with Frame-by-Frame Component Removal of Statistically Dominant Source

A Multi-Source Separation Approach Based on DOA Cue and DNN

Speech Sources Separation Based on Models of Interaural Parameters and Spatial Properties of Room

Contact Info

Product

Resources

About