Device and method for efficiently encoding quantization parameters of spectral coefficient coding

摘要

This invention introduces apparatus and methods to efficiently encode the quantization parameters of split multi-rate lattice vector quantization. In this invention, by doing spectral analysis on the split multi-rate vector quantized spectrum, the spectrum is split to null vectors region and non-null vectors region. For the null vectors region, instead of transmitting series of indication for null vectors, an indication of null vectors region and the quantized value of index of the ending vector in the null vectors region (or the number of the null vectors in the null vectors region) are transmitted. The indication of null vectors region can be designed in many ways, the only requirement is the indication should be distinguishable in the decoder side. The ending index or the number of null vectors can be quantized by an adaptively designed codebook. By applying of the invented method, some bits can be saved from the codebook indications.

主权项

1. An audio/speech encoding apparatus, comprising:
a band splitting section that splits a spectrum of input signal to a plurality of sub-bands; a vector quantization section that quantizes spectral coefficients in each sub-band; a spectral analysis section that splits the spectrum to null vectors region and non-null vectors region by analyzing on a series of indications of sub-bands generated by vector quantization; and a parameter encoding section that converts the series of indications for null vectors in the null vector region to an indication of null vectors region and a parameter which represents the ending position of the null vectors region, wherein the parameter encoding section comprises: a first parameter encoding section that converts the series of indications for null vectors in the null vector region to an indication of null vectors region and a parameter which represents the ending position of the null vectors region; a reversing section that reverses the series of indications; a second parameter encoding section that converts the reversed series of indications for null vectors; and a selection section that selects between the first parameter encoding section and the second parameter encoding section based on which consumes less bits; and wherein audio/speech is encoded in accordance with the selection by the selection section.