OptimalBitLayering forScalableAudio Compression Using Objective AudioQuality Metrics
Srivatsan Kandadai, D. Creusere · 2007
Perceptual audiocompression usestheideaof auditory maskingtohidecodingdistortion. Theseauditory masking thresholds areobtained frommathematical modelsofthe humanear.Atlowbitrates however, coding noise issignificant andcannotbemaskedbytheaudiocontent. A perceptually scalable audiocompression system, evenatlowbitrates, should generate abitstream withlayers ofaudiofidelity suchthateach layer improves thequality ofthereconstructed audiothatisjust noticeable bythelistener. Inthis paperwedescribe alowbitrate (8- 64kbps), scalable audiocompression systemwhichusesa residual weighted VQ algorithm togenerate ascalable bitstream. Tomodifythisbitstream sothatitisperceptually scalable, we layer thedifferent residual indices generated bythecoderusing objective audioquality metrics developed forevaluating highly impaired audio signals.