| 摘要 |
Data to be compressed is sampled at a time interval of a sampled point where an error is at a desired value or smaller between a data value on a straight line for connecting two pieces of sampled data and a sampled data value corresponding to the data value, and a pair of discrete amplitude data D 1 , D 6 , . . . , on each sample point and timing data indicative of a time interval between sample points is obtained as compressed data. Hence, from a large number of pieces of sampled data D 1 , D 2 , . . . , included in data to be compressed, only a pair of amplitude data and timing data can be obtained as compressed data on a sample point where an error from original data is not large even when linear interpolation is performed during expansion. Thus, it is possible to remarkably improve the quality of reproduced data while achieving high compressibility.
|
