| 摘要 |
PROBLEM TO BE SOLVED: To provide a reversible audio codec whose compression performance is optimized according to constraint of the maximum size to respective data units which can be independently decoded.SOLUTION: A reversible audio codec segmentalizes audio data on respective frames so as to raise compression performance according to constraint that respective segments should be made to be completely decoded and made into size less than the maximum size. A codec selects segment duration, and encoding parameters for every frame, and minimizes encoding payloads of the whole frames according to the constraint. The compression performance can be further raised by forming M/2 non-correlated channels to audio of M channels. A combination of three, 'a base, correlation, non-correlation' of the channels provides combination of two pairs, 'the base, the correlation', and 'the base, the non-correlation' which can be considered in segmentalization and optimization of entropy encoding for further raising the compression performance. |
