去除AMR语音编码器前瞻时延的研究

发布时间：2018-02-25 05:27

本文关键词： 语音编码自适应多速率线性预测协方差前瞻　出处：《北京理工大学学报》2014年07期 　论文类型：期刊论文

【摘要】：为降低自适应多速率(AMR)语音编码器的时延,提出了一种去除其前瞻时延的方法.此方法采用协方差算法计算线性预测系数,然后用线性预测系数外推得到新样点.用新推得的样点取代原来的前瞻样点,从而去除了编码器的前瞻时延.进一步对重建语音质量分别进行了客观评价和主观评价,并且计算了算法复杂度.实验结果表明该方法在去除前瞻时延的同时,重建语音质量接近原始AMR的重建语音质量.
[Abstract]:In order to reduce the delay of adaptive multi-rate AMR-speech encoder, a method to eliminate the forward delay is proposed. The covariance algorithm is used to calculate the linear prediction coefficient. Then the new sample points are obtained by extrapolating the linear prediction coefficients, and the original prospective samples are replaced by the new ones, thus the forward delay of the encoder is removed. The objective and subjective evaluation of the reconstructed speech quality is carried out, respectively. The algorithm complexity is calculated and the experimental results show that the reconstructed speech quality is close to that of the original AMR at the same time as the prospective delay is removed.
【作者单位】：北京理工大学信息与电子学院;
【分类号】：TN912.3

【相似文献】