Entropy coded vector quantization with hidden Markov models
Tadashi Yonezaki, Kiyohiro Shikano · 2002
The authors propose a new vector quantization approach, which consists of hidden Markov models (HMMs) and an entropy coding scheme. The entropy coding system is determined depending on the speech status modeled by HMMs, so the proposed approach can adaptively allocate suitable numbers of bits to the codewords. This approach realizes about 0.3[dB] coding gain in cepstrum distance (8 state HMMs). In other words, an 8 bit codebook is represented by about 6.5 bits for average code length. They also research for robustness to the channel error. HMMs and the entropy coding system, which seem to be weak to the channel error, are augmented to be robust, so that the influence of the channel error is decreased into one-third.