| 摘要 |
PROBLEM TO BE SOLVED: To provide the compression method for digital data where an arithmetic error of conversion coding is minimized and signal deterioration attended with compression is reduced. SOLUTION: A round-off circuit 2 applies round-off processing to a 16-bit DCT coefficient into data in 10-bit (9-bit for a DC component), a coring circuit 4 discriminates whether or not the absolute value|D(i)|of DCT coefficients D(i) (i=N1-63) of a high frequency region is smaller than a threshold TH1, and in the case of|D(i)|<TH1, the circuit 4 applies coring processing to replace the DCT coefficients with '0', and variable length coding is applied to data after the coring processing as an object. |
