Two-channel forward NLMS algorithm combined with simple variable step-sizes for speech quality enhancement

Blind 2-Microphone Acoustic Noise Reduction Algorithms Using Efficient Variable Step-Size Adapted by Minimizing the Intercorrelation Function

“…( ) n x 2 [11,18]. In 2microphone convoluting system of Figure 2 -a-, the noisy signals are attained from the next relation:…”

“…-a--b- Figure 1 -b-denoted the simplified model of convoluting mixture system that is detailed and presented previously in [12], where, the two signals are statistically independents. In the detailed study presented in [12,14], directs impulses responses h11(n) and h22(n) are considered as unit Dirac impulse 𝛿(𝑛) [2,11,12,13]. The following equations provide the formulas of noisy speech:…”

“…Recently, in digital signal processing applications, the adaptive filtering algorithms are commonly utilized to reduce the acoustic echo and noise. Several important papers have treated the reducing of noise signal and enhancing of speech signal using the combination of different adaptive algorithms with the 2-microphone source separation techniques [8][9][10][11]. In study [12][13][14], the authors have been demonstrated that there exists a mathematical equivalence between the problem of unknown impulse response identification and source separation problem.…”

“…Noting that, a large step size value suggests the faster convergence of adapting filter, while a small value of this parameter gives low misadjustment and good speech quality. Many variable algorithms and techniques have been considered as solution for cancelling the acoustic echo and identification system [15][16][17] and others are proposed in 2-microphones noise reduction application [3,11]. In [3,11], we have proposed five variable 2-microphones adaptive algorithms that are given a fast convergence rate for identify the impulses responses and low maladjustment of coefficients, but we have noted that the complexity of calculation is seriously increased compared with theirs FS versions.…”

“…Many variable algorithms and techniques have been considered as solution for cancelling the acoustic echo and identification system [15][16][17] and others are proposed in 2-microphones noise reduction application [3,11]. In [3,11], we have proposed five variable 2-microphones adaptive algorithms that are given a fast convergence rate for identify the impulses responses and low maladjustment of coefficients, but we have noted that the complexity of calculation is seriously increased compared with theirs FS versions. Noting that in these recently proposed algorithms, the used functions of VS parameters is based on minimization the two output error signals.…”