| 摘要 |
PURPOSE:To realize quantization and efficient compression of a signal in response to the static property of an input signal by preparing plural quantizers having a different characteristic and using the quantizing characteristic in response to the sequence of transformation coefficients and the magnitude of the moving vector. CONSTITUTION:Transformation coefficients 104 and a moving vector 105 obtained by the similar processing to that of a conventional system are inputted to an adaptive quantizing section 2a. The adaptive quantizing section 2a selects a proper quantizing characteristic in response to the sequence of the transformation coefficients 104 and the magnitude of the moving vector 105, and as a result of a threshold level processing based on a buffer storage quantity 111, quantization to a significant transformation coefficients 104 is implemented. In the case of selecting the quantization characteristic, the moving vector 105 is sent to the decoder side and since the sequence of the transformation coefficients is known, it is not required to send overhead information. Thus, the efficient quantization in response to the statistical property of the input signal is attained. |
