Abstract:For distortion problem existed in speech enhanced by traditional adaptive wavelet packet threshold algorithm,we propose the adaptive wavelet packet threshold speech enhancement algorithm combined with improved subspace.The new algorithm proposed first makes Karhunen Loeve transform on noisy speech to get its eigenvalues,which are then dealt with by adaptive wavelet packet threshold algorithm to remove part of the noise subspace.Then it adopts recursive least squares algorithm (RLS)to estimate the noise eigenvalues,and revises the eigenvalue estimation deviation caused by traditional subspace algorithm.Finally the algorithm uses the derived new eigenvalues processed by adaptive wavelet packet threshold algorithm to subtract the noise eigenvalues so as to eliminate all the noise subspace and eventually restores the pure speech by KL reverse transform.Simulation results show that this new algorithm has better enhancement effect compared with traditional adaptive wavelet packet threshold algorithm,and reduces the speech distortion as well.And what’s more,in the circumstance of low SNR,the new algorithm gains much higher improvement in signal-to-noise ratio (SNR)and segment SNR (SSNR)of the enhanced speech.

Wavelet Thresholding for Speech Enhancement Based on Adaptive Noise Estimation

Speech enhancement using time-frequency wavelet threshold with auditory masking

Speech Enhancement Using the Combination of Adaptive Wavelet Threshold and Spectral Subtraction Based on Wavelet Packet Decomposition

Adaptive speech enhancement based on wavelet in high noise environment

An Improved Speech Enhancement Algorithm Based on Wavelet Transform

Research on Speech Enhancement Based on Perceptual Wavelet Transform

Speech enhancement algorithm based on wavelet packet adaptive threshold revised by posteriori SNR

An Improved Wavelet-Based Speech Enhancement Method Using Adaptive Block Thresholding

Speech enhancement using adaptive threshold based on bi-orthogonal wavelet packet decomposition

An improved wavelet packet threshold function for speech enhancement method

Adaptive Multiresolution Speech Enhancement Algorithm Based on Wavelet Transform

Noise Reduction of Speech Signal Based on Wavelet Transform with Improved Threshold Function

Speech Enhancement Based on the Multi-Scales and Multi-Thresholds of the Auditory Perception Wavelet Transform

An Adaptive Speech Enhancement Method Based on Hearing Wavelet Packet Transformation

ADAPTIVE WAVELET PACKET THRESHOLD SPEECH ENHANCEMENT ALGORITHM COMBINED WITH IMPROVED SUBSPACE

New threshold function of wavelet packet speech enhancement algorithm

Wavelet De-Noising Method by Dynamic Threshold Based on Separation of Voiced Signal and Unvoiced Signal

Speech Enhancement Based on Noise Estimation and Auditory Masking Properties

Research on New Speech Denoising Based on Spectral Subtraction and Wavelet Threshold Method

A Continuous Differentiable Wavelet Threshold Function for Speech Enhancement

Adaptive Wavelet Thresholding for Time Varying SNR Signal Denoising