| 摘要 |
PURPOSE:To realize a fixed quantizer with a simple circuit and a high speed processing by adding at least two fixed quantizers and selecting the quantization step size of the fixed quantizer to be the power of two. CONSTITUTION:A video signal is inputted from a video signal input 101 for every macro block, inputted to a movement compensation prediction device 107, the signal of the macro block minimizing a predictive error is selected from one-preceding signal recorded already and the signal of the macro block is outputted. The signal passes through a loop filter 108 and a difference from the video input signal is taken and converted by a discrete cosine transformation device 103 for every block. Then the result is quantized by a variable quantizer 104, subjected to Huffman coding and the result 112 of coding is outputted. The quantization characteristic of the variable quantizer 104 is set properly so as to be in matching with the transmission rate. |
