一种基于MELP的300bps声码器设计  

Design of a 300bps vocoder based on MELP

作　　者：曾镇城 郁文贤[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《信息技术》2014年第10期73-75,81,共4页Information Technology

摘　　要：基于MELP设计了一种300bps声码器。该声码器将8个语音帧组成一个超级帧。子带清浊音、基音周期、短时能量三项参数采用8帧联合的方式进行矢量量化,线谱频率采用4帧联合的方式进行多级矢量量化。通过概率统计的方法得到子带清浊音的量化码本,并利用清浊音信息与其他参数的相关性将编解码分为不同的模式,从而改善量化性能。采用加权的欧氏距离度量基音周期和线谱频率的失真度,以提高合成语音的自然度。合成语音的PESQMOS分数为2.096。A 300bps vocoder based on MELP is designed. The vocoder composes a super-frame of eight frames. The parameters of band pass voicing coefficients （BPVC）, pitch and gain are quantized by 8- frame-joint vector quantization. The parameter of linear spectral frequency（LSF） is quantized by 4-frame- joint multi-stage vector quantization. The BPVC quantization codebook is obtained in the way of probabilistic statistics. In order to improve quantized performance, the vocoder uses the correlation of voiced/unvoiced sound information and other parameters to divide code/decode into different modes. In order to enhance the naturalness of synthetic speech, the vocoder uses the weighted Euclidean distance to measure the distortion of pitch and LSF. The PESQMOS score of synthetic speech quality is 2. 096.

关 键 词：MELP 300bps 声码器 

分 类 号：TN912[电子电信—通信与信息系统]