论文部分内容阅读
语言编码方法可以划分为两个大类,即波形编码器和声码器。波形编码器努力设法恢复语言波形,它并不企图了解语言信号是如何产生的。这些编码器在比特率高于16kb/s时便能产生高质量的语言。低比特率语言编码的领域通常为声码器及其派生的编码器所占有。声码器在把比特率降低到极低的数值方面效率很高,但于此却必须以较低的语言质量为代价。本文描述了语言编码的某些新方法,旨在弥补波形编码器和声码器之间的间隙而又不受其局限性的损害。这些新方法采用综合分析措施决定最佳传输参数,这样就使在主观上显得重要的原始语言信号和编码语言信号之间的差值最小。
Language coding methods can be divided into two categories, namely, waveform encoder and vocoder. The waveform encoder struggled to recover the speech waveform and did not attempt to understand how the speech signal was generated. These encoders produce high-quality speech at bit rates above 16kb / s. The field of low bit rate speech coding is usually owned by vocoders and their derived encoders. Vocoders are efficient at lowering the bit rate to very low values, but at the cost of lower language quality. This article describes some of the new ways of encoding speech to compensate for the gap between the waveform encoder and vocoder without being compromised by its limitations. These new approaches employ comprehensive analysis to determine the optimal transmission parameters, thus minimizing the difference between the subjectively significant original and coded speech signals.