Perceptual coding using sinusoidal modeling in the MDCT domain

摘要

MDCT based perceptual audio coders shape the quantization noise according to simple psychoacoustic rules andgeneral behavioral aspects of the audio signal such as stationarity and tonality . As a consequence, the resultingcompressed audio representation has little semantic value making difficult MPEG-7 oriented operations such asfeature extraction and audio modification directly in the compressed domain. First results in this perspective arereported using an enhanced version of an MDCT based perceptual coder that implements sinusoidal modeling andsubtraction directly in the MDCT frequency domain, as well as spectral envelope modeling and normalization. Theimplications on the coding efficiency are also addressed.