论文部分内容阅读
论文描述了为得到高质量的宽带语音而使用的有效编解码方法。CELP(Code Excited Linear Prediction)技术应用于窄带语音编码时已获得了很高的语音质量,然而直接应用于宽带语音信号编码时不能有效地保持高质量语音,因此需要在CELP模型上添加额外的方法,以使其在宽带信号上亦取得高质量。文章所讨论的提高CELP模型性能的有效技术是感觉加权滤波器,其技术也被用于3GPP所选的AMR-WB(Adaptive Multi-Rate Wideband)[1][2]声码器中。
The paper describes efficient coding and decoding methods used to obtain high quality wideband speech. CELP (Code Excited Linear Prediction) technology has been used in narrow-band speech coding has been very high voice quality, but directly applied to broadband voice signal coding can not effectively maintain high-quality voice, so the CELP model requires the addition of additional methods , So that it also achieved high quality on wideband signals. An effective technique to improve the performance of CELP model discussed in the paper is the perceptual weighting filter, and its technology is also used in AMR-WB (Adaptive Multi-Rate Wideband) [1] [2] vocoder selected by 3GPP.