MODEL BASED PREDICTION IN A CRITICALLY SAMPLED FILTERBANK

摘要

The present document describes methods and systems which improve the quality of audio source coding employing prediction in the subband domain of a critically sampled filterbank. The methods and systems may make use of a compact description of subband predictors, wherein the description is based on signal models. Alternatively or in addition, the methods and systems may make use of an efficient implementation of predictors directly in the subband domain. Alternatively or in addition, the methods and systems may make use of cross subband predictor terms, as described in the present document, to allow for a reduction of alias artifacts.