Quantization of speech and audio coding parameters using partial information on atypical subsequences

The method and equipment for quantizing a parameter using the partial information about an un-typical partial series are indicated on these specifications. The step according to which this method classifies partially two or more 1st partial series in a target vector into some selected groups in 1 enforcement form, The step which generates the fidelity standard refined about each partial series of two or more 1st partial series based on the information acquired from the classification, It is a step which divides a target vector into two or more 2nd partial series, and a step which codes two or more 2nd partial series, and the step concerned including the process of giving the refined fidelity standard and quantizing two or more 2nd partial series is included. [Selection figure] Fig. 1