基于噪声被掩蔽概率的优化语音增强方法

Optimizing Speech Enhancement Based on Noise Masked Probability

利用听觉系统的掩蔽特性,提出了一种优化的语音增强方法。研究表明,噪声被语音掩蔽的概率是噪声强度和听觉掩蔽阈值的函数。考虑到噪声在带噪语音中的出现具有不确定性,各语音谱分量的最终估计由对带噪语音的谱分量和用传统的增强方法估计的谱分量的加权求得,加权因子由噪声被掩蔽概率确定。语音增强性能的评估结果表明,这种优化的语音增强方法在减少语音失真与加强噪声抑制之间取得了良好的折衷,减少了语音的听觉失真, 有效地抑制了音乐噪声,提高了增强语音的清晰度。

An optimal approach for enhancing a speech signal degraded by uncorrelated stationary additive noise, which exploits auditory perception properties, is proposed. The speech spectra estimate is performed in two cases: noisy speech spectra for noise masked and classical estimate for noise unmasked. Taking account into the uncertainty of the noise presence, the enhanced speech signal spectra are obtained by a weighted sum of these two estimates, where the weights are given by the noise masked probability. The performance of the proposed speech enhancement approach has been evaluated with speech distortion and informal listening tests. Comparing with Azirani's method and classical estimator, results show that a better compromise between reducing speech distortion and reinforcing noise suppression has been made, speech distortion has been decreased apparently, musical noise has been suppressed and speech articulation has been improved.