TRANSCODER AND CODED IMAGE CONVERSION METHOD

摘要

A transcoder and a coded image converting method are provided to determine a quantization step value to be allocated to each macroblock on the basis of an evaluation value of indicating a degree of dispersion in pixels within a macroblock, thereby improving image quality by optimizing bit distribution after converting the coded image. A transcoder(1) comprises a decoder(2), a macroblock evaluation value calculating unit(5), a macroblock evaluation average calculating unit and an encoding unit(3). The decoder decodes a first coded image. The macroblock evaluation value calculating unit calculates a macroblock evaluation value of indicating a degree of dispersion in pixel values within a macroblock of an image decoded in the decoder. The macroblock evaluation average calculating unit calculates a macroblock evaluation average by averaging the macroblock evaluation values in a frame. The encoder encodes the image encoded by the decoder to generate a second coded image. The encoder comprises a quantization step average acquisition unit, a weight calculating unit, and a step value calculating unit. The quantization step average acquisition unit acquires a quantization step average by averaging quantization step values within a frame of the first coded image. The weight calculating unit calculates a weight value for each macroblock on the basis of a difference between the macroblock evaluation value and the macroblock evaluation average. The step value calculation unit calculates a quantization step value for conversion for each macroblock of the second coded image on the basis of a value of adding the weight value to the quantization step average.