论文部分内容阅读
听觉模型已应用于语音信号处理的许多方面,并已取得了较好的结果。论文根据目前应用较为广泛的听觉模型,提出了模型中各部分的逆变换。首先通过对自相关图谱逆变换并经迭代运算重构相位信息恢复神经发放率,再经半波整流逆运算恢复负半部分信号,最后对描述内毛细胞、突触模型的相关方程进行逆运算和Gammatone逆滤波,构成整个听觉模型的反演方法。作为应用论文提出了一种基于听觉模型逆变换的含噪语音增强方法。实验结果表明,该方法对含噪语音有很好的降噪效果,特别是当信噪比较低时,该方法较常用的方法更为有效。论文提出的听觉模型逆变换方法可应用于语音增强等领域。
Auditory models have been applied to many aspects of speech signal processing and have yielded good results. According to the currently widely used auditory model, the paper proposes the inverse transformation of each part in the model. First, the autocorrelation map is inverse transformed and the phase information is reconstructed by iterative operation to recover the nerve delivery rate, then the negative half signal is recovered by the half-wave rectification inverse operation. Finally, the inverse equations describing the inner hair cells and the synaptic model are inversely calculated And Gammatone inverse filtering, constitute the entire auditory model of the inversion method. As an application paper, a noisy speech enhancement method based on inverse transform of auditory model is proposed. Experimental results show that this method has a good noise reduction effect on noisy speech, especially when the SNR is low, which is more effective than the commonly used method. The proposed method of inverse transformation of auditory model can be applied to speech enhancement and other fields.