A two-sensor Gauss–Seidel fast affine projection algorithm for speech enhancement and acoustic noise reduction. (September 2015)

• Record Type: Journal Article
• Authors: Darazirar, Ilyes; Djendi, Mohamed
• Is Part Of: Applied acoustics. Volume 96(2015)
• Page Start: 39
• Page End: 52
• Publication Date: 2015-09
• Subjects: Speech enhancement -- Two-sensor -- Forward structure -- LMS -- BSS -- Gauss–Seidel; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2015.02.010
• Languages: English
• ISSNs: 0003-682X

Abstract

Graphical abstract: Highlights: We propose a new Gauss–Seidel fast affine projection algorithm. The proposed algorithm is less complex even with long adaptive filters. This algorithm is efficient in speech enhancement and acoustic noise reduction applications. The proposed algorithm has less steady state error and enhances the speech signal intelligibility. The proposed algorithm is recommended for the speech enhancement applications. Abstract: This paper deals with the problem of noise reduction and speech enhancement by adaptive filtering using two-sensor configuration. The well-known Gauss–Seidel fast affine projection (GSFAP) algorithm is used in various applications such as acoustic echo cancelation (AEC), and adaptive noise control. In this paper, we use the GSFAP algorithm in noise reduction and speech enhancement application. Moreover, we derive a new two-sensor GSFAP (TS-GSFAP) algorithm for speech enhancement and acoustic noise reduction application. The application of the proposed TS-GSFAP algorithm to the forward blind source separation (FBSS) structure leads to a significant improvement of the output speech signals quality even in noisy conditions. A fair comparison of the proposed TS-GSFAP algorithm with other standard two-sensor type algorithms is presented. This comparison is based on the evaluation of several objective and subjective criteria such as: the segmental signal-to-noise ratio (SegSNR), the Cepstral distance (CD) and the system mismatch (SM), the perceptual evaluation of speech quality (PESQ), and the mean opinion score (MOS) criteria. The obtained results show the good performances of the proposed TS-GSFAP algorithm.