| 摘要 |
PROBLEM TO BE SOLVED: To attain the lossless encoding processing which has its high efficiency by deciding the optimum number of bits in accordance with the input signal state. SOLUTION: The differential value data (dx) are obtained from the series of linear PCM data (xi) by a differential value generation part 5 and then quantized again by a quantization part 8. The part 8 includes a 1st mode where the series of data (dx) are collected in each frame and the minimum bit number necessary for expression of the data (dx) having the maximum absolute value among those data collected in each frame is decided as the word width, and a 2nd mode where the value of L is decided so as to minimize the bit length of a frame when the data (dx) are divided in each frame and every L bits from the LSB side and a one-bit flag showing the presence or absence of the final word of each value (dx) every L bits is set as the word width. Then the part 8 selects the 1st or 2nd mode that has the shorter bit length of a frame and adds this mode to a mode information frame. |
