| 摘要 |
PURPOSE:To vary easily a quantized bit number and a quantized step width by obtaining a product between a quantized data and a reciprocal of the quantized step width through the use of a multiplier so as to quicken the calculation of quantization. CONSTITUTION:A quantized data (i) is given by a multiplier as a product between a data (x) and the reciprocal (d) of the quantized step width. Then the (i) is compared with a maximum value max and a minimum value min reached by data. If the (i) is at the outside of the range decided by the max and min, the max or the min is used as the value of (i). The value (i) quantizing the data (x) is obtained through the calculation above. That is, since the calculation is attained only by one multiplication and two condition discriminations, the method offers advantages such as short processing time, only a small memory capacity required because of no table employed, and ease of change of quantized step width and quantized bit number through the change of the reciprocal (d) of the quantized step width, the maximum value max and the minimum value min. |
