| 摘要 |
The difference dn between an input signal xn and a predicted value yn is adaptively quantized and encoded. The adaptive quantization is performed so that the reverse quantization value qn of a code Ln for a section where the absolute value of the difference dn is small may be approximately zero.
|
