LINEAR PREDICTION BASED AUDIO CODING USING IMPROVED PROBABILITY DISTRIBUTION ESTIMATION

摘要

Linear prediction based audio coding is improved by coding a spectrum composedof a plurality of spectral componentsusing a probability distribution estimation determined for each of theplurality of spectral components from linear predictioncoefficient information. In particular, the linear prediction coefficientinformation is available anyway. Accordingly, it may be usedfor determining the probability distribution estimation at both encoding anddecoding side. The latter determination may beimplemented in a computationally simple manner by using, for example, anappropriate parameterization for the probability distributionestimation at the plurality of spectral components. All together, the codingefficiency as provided by the entropy coding is compatiblewith probability distribution estimations as achieved using context selection,but its derivation is less complex. For example, thederivation may be purely analytically and/or does not require any informationon attributes of neighboring spectral lines such aspreviously coded/decoded spectral values of neighboring spectral lines as isthe case in spatial context selection.