Analysis-by-Synthesis Speech Coding Based on Relaxed Waveform-MatchingConstraints

摘要

In modern digital communication systems, speech is transmitted in the form of asequence of bits. Often, physical, regulatory, and financial constraints make it desirable to minimize the transmission rate of the bits describing the speech signal. In the thesis, new methods to encode the speech signal at bit rates of around 3000 to 5000 bits/second (b/S) are described. The procedures described have in common a relaxation of the constraints used in existing Linear Prediction-based analysis-by-synthesis speech coders, which does not affect the perceptual quality of the reconstructed speech. As a result, a speech signal of given perceptual quality can be encoded at lower bit rates.