| 摘要 |
PURPOSE:To increase the effect of compression, by making the higher (k) bits effective among (l) bits/picture element for both the input picture element value and the predicted error each and at the same time making the (l) bits/picture element effective for the expansion value, the reproduced picture element value and the prediction value each for an operation. CONSTITUTION:The input picture signal X of 6 bits/picture element is supplied to a differential device 1 in the form of the input picture element value X of 8 bits/ picture element. The device 1 delivers the predicted error (e) obtained by subtracting the prediction value from the value X. For this error (e), the higher 6 bits among the 8 bits/picture element are referred to by the quantizer 2 as the effective bits for the error (e). The quantizer 2 converts the error (e) into the quantized value (q) by referring to the error (e) and a table of the quantized value corresponding to the error (e). The value (q) is coded by the coder 3 to be delivered and at the same time expamded by the expander 4 to be approximate to the value of the original error (e). Thus the expansion value (q') is obtained. The value (q') is added 5 with the prediction value to obtain the reproduced picture element value X'. The prediciting device 6 calculates a prediction value to the next picture element value from the several units of value X'. |
