| 摘要 |
<p>PURPOSE:To reduce memory capacity by previously storing a quantizing coefficient smaller than the original number of the quantizing coefficient for a signal quantizing matrix and dividing plural DCT coefficients by the same quantizing coefficient. CONSTITUTION:In one quantization matrix, only 13 pieces of Q00 to Q12 are stored in a quantization table 13 as the quantization coefficient constituting the single quantization matrix. Therefor, this compressor is provided with an address encoder 16 storing a quantization coefficient QUV corresponding to a DCT coefficient SUV which is successively outputted from a DCT transform unit 11 in advance and generating an address. Thereby, the size of the quantization table is about 13/64 20% as compared with a conventional manner, so that it is enough to provide only RAM with the memory capacity smaller by the reduced part. Otherwise, more quantization coefficient matrixes can be stored.</p> |
