| 摘要 |
PROBLEM TO BE SOLVED: To enable great lossless compression by changing ways of acquiring difference data on the pitch of a sound waveform according to the kind of sound waveform data, finding a time difference minimizing the difference at the time of bit compression, and finding and using such a constant that data which can be compressed become more. SOLUTION: The inputted sound waveform is converted into a digital value, which is stored temporarily in a sound waveform data file 2-1; and a difference calculation parameter setting part 3-1 finds respective parameters of magnification, a time difference, and calculation order used for difference data calculation and stores them in a difference calculation parameter storage part 2-2. For the magnification parameter, the source such a constant that the source sound waveform data have data which can be represented with bits 1/2<n> as many as that of the source data is found at the time of the calculation of differences. The time difference parameter is effective when the same sound source is picked up by more than one microphone. A difference calculation bit compression part 3-2 combines those parameters, performs the compression so that the total of differences in a time of one pitch becomes minimum, and sends the resulting data to a data storage part 2-3. |
