This paper proposes a vector quantization scheme which makes it possible to
consider the dynamics of input vectors. In the proposed scheme, a linear t
ransformation is applied to the consecutive input vectors and the resulting
vector is quantized with a distortion measure defined by the statistics. A
t the decoder side, the output vector sequence is determined using the stat
istics associated with the transmitted indices in such a way that a likelih
ood is maximized. To solve the maximization problem, a computationally effi
cient algorithm is derived. The performance of the proposed method is evalu
ated in LSP parameter quantization. It is found that the LSP trajectories a
nd the corresponding spectra change quite smoothly in the proposed method.
It is also shown that the use of the proposed method results in a significa
nt improvement of subjective quality.