论文部分内容阅读
为了精确地量化基音周期参数,提高低速率声码器的语音质量,针对低速率声码器算法,考虑到基音周期参数与带通浊音度参数的关系,论文对二者进行了基于码字索引空间分配的联合量化。将码字索引空间按照带通浊音度模式进行细致划分,全浊音模式下的基音周期得到了更多的码矢量,从而进一步提高了基音周期的量化精度。实验结果表明,与单纯矢量量化的算法相比,该算法能够使基音周期参数的量化误差降低20%以上,从而提高合成语音的质量。
In order to accurately quantize the pitch period parameters and improve the speech quality of the low rate vocoder, aiming at the low rate vocoder algorithm, taking into account the relationship between pitch period parameters and bandpass vaporizability parameters, Joint Quantification of Spatial Distribution. The codeword index space is finely divided according to the band-pass turbidity mode, more code vectors are obtained in the pitch cycle in the all-voiced mode, so as to further improve the quantization accuracy of the pitch cycle. The experimental results show that the proposed algorithm can reduce the quantization error of pitch parameters by more than 20% compared with the simple vector quantization algorithm, and improve the quality of synthesized speech.