G.723语音编码介绍及优化 语音编码分成波形编码、参数编码和混合编码.波形编码就是语音信号形成的波形进行处理和加工，参数编码是以语音信号产生的数字模型为基础，求出其模型参数，然后按照这些参数还原数字模型，合成语音。混合编码是综合了波形编码和参数编码的长处，在4－16kbps的速率上进行高质量的语音合成。例如多脉冲激励线性预测编码(MPLPC),码本激励线性预测编码(CELP)都是混合编码。-G.723 speech coding and optimization of voice introduced into coding waveform coding, coding parameters and mixed coding. Waveform voice signal is encoded form of waveform handling and processing, encoding parameters based on the voice signal to digital model based on its model parameters calculated, and then it these parameters reduction digital model, synthetic voice. Mixed coding is integrated waveform coding and coding parameters of the merits of the 4-without compromising the audio on the rate of high-quality speech synthesis. For example, multi-pulse Excited Linear Predictive Coding (MODIFIED), the Code Excited Linear Predictive Coding (CELP) are mixed coding.