Transcoder and coded image conversion method
View Patent ↗A quantization step determination part inputs an evaluation value (ACT_MB) indicating the dispersion in a macroblock and its average value (ACT_PIC). A subtracter obtains the difference between these values, and a multiplier multiplies the difference by r aq (<1) to obtain a weighting value. Next, an adder adds the weighting value to an average quantization step value of source data, and finally a multiplier multiplies the sum by a step value adjustment factor α (>1) to obtain a converted quantization step value (Qstep_AVC). This optimizes a bit allocation in accordance with an Activity value of the macroblock, to thereby improve the quality of image.
1. A transcoder comprising:
decoder means for decoding a first coded image;
means for calculating a macroblock evaluation value indicating the degree of dispersion in pixel values inside a macroblock of an image obtained by decoding in said decoder means;
means for obtaining an average macroblock evaluation value by averaging said macroblock evaluation values in a frame; and
encoder means for encoding the image obtained by decoding in said decoder means to generate a second coded image,
wherein said encoder means comprises
means for acquiring an average quantization step value by averaging quantization step values in a frame of said first coded image;
weighting value calculation means for calculating a weighting value for each macroblock on the basis of the difference between said macroblock evaluation value and said average macroblock evaluation value; and
step value calculation means for calculating a converting quantization step value for each macroblock of said second coded image on the basis of a value obtained by adding said weighting value to said average quantization step value.
2. The transcoder according to claim 1 , wherein
said macroblock evaluation value is a differential absolute value sum of an average pixel value in a macroblock and each pixel value in the macroblock.
3. The transcoder according to claim 1 , wherein
said step value calculation means includes means for multiplying a value obtained by adding said weighting value to said average quantization step value by a step value adjustment factor which is larger than 1.
4. The transcoder according to claim 1 , wherein
said weighting value calculation means includes means for multiplying a difference between said macroblock evaluation value and said average macroblock evaluation value by a weighting value adjustment factor which is smaller than 1.
5. The transcoder according to claim 4 , wherein
said weighting value calculation means includes a limiter giving an upper limit to a value obtained by multiplication by said weighting value adjustment factor.
6. The transcoder according to claim 4 , wherein
said weighting value calculation means includes a limiter giving a lower limit to a value obtained by multiplication by said weighting value adjustment factor.
7. The transcoder according to claim 5 , wherein
said weighting value calculation means includes means for calculating an upper limit value of said limiter by multiplying said average quantization step value by a predetermined factor.
8. The transcoder according to claim 6 , wherein
said weighting value calculation means includes means for calculating a lower limit value of said limiter by multiplying said average quantization step value by a predetermined factor.
9. The transcoder according to claim 7 , wherein
said weighting value calculation means includes
means for storing a fixed upper limit value to be set to said limiter; and
means for selecting a smaller one out of the upper limit value obtained by multiplication by said predetermined factor and said fixed upper limit value, as an upper limit value of said limiter.
10. The transcoder according to claim 8 , wherein
said weighting value calculation means includes
means for storing a fixed lower limit value to be set to said limiter; and
means for selecting a larger one out of the lower limit value obtained by multiplication by said predetermined factor and said fixed lower limit value, as a lower limit value of said limiter.
11. The transcoder according to claim 3 , wherein
said step value calculation means includes
means for storing a picture adjustment factor corresponding to a picture frame type; and
means for selecting said picture adjustment factor corresponding to a picture frame type of an object macroblock and utilizing said selected picture adjustment factor as said step value adjustment factor.
12. The transcoder according to claim 1 , further comprising
means for inputting a macroblock motion evaluation value indicating the degree of variation in a time direction, of a macroblock of an image obtained by decoding in said decoder means,
wherein said step value calculation means includes means for multiplying said average quantization step value by a motion adjustment value calculated on the basis of said macroblock motion evaluation value.
13. The transcoder according to claim 12 , wherein
said macroblock motion evaluation value is an interframe differential absolute value sum between a macroblock and a reference image macroblock.
14. The transcoder according to claim 13 , wherein
said macroblock motion evaluation value calculated in an immediately preceding P frame is substituted for said macroblock motion evaluation value of an I frame which can not be calculated.
15. The transcoder according to claim 13 , wherein
said macroblock motion evaluation value calculated in an immediately preceding macroblock is substituted for said macroblock motion evaluation value of an intra macroblock, whether a P frame or a B frame, which can not be calculated.
16. The transcoder according to claim 13 , wherein
said macroblock motion evaluation value calculated in a neighborhood macroblock is substituted for said macroblock motion evaluation value of an intra macroblock, whether a P frame or a B frame, which can not be calculated.
17. The transcoder according to claim 12 , wherein
said step value calculation means includes means for calculating said motion adjustment value in accordance with the ratio between said macroblock motion evaluation value and said macroblock evaluation value.
18. The transcoder according to claim 3 , wherein
said step value calculation means includes
means for calculating a picture adjustment factor corresponding to a picture frame type on the basis of a source bit rate of said first coded image and a target bit rate of said second coded image; and
means for selecting said picture adjustment factor corresponding to a picture frame type of an object macroblock and utilizing said selected picture adjustment factor as said step value adjustment factor.
19. The transcoder according to claim 18 , wherein
said picture adjustment factor is recalculated when said source bit rate changes.
20. The transcoder according to claim 18 , wherein
said target bit rate is changed in accordance with the capacity of a storage medium for storing said second coded image and said picture adjustment factor is recalculated in accordance with the change.
21. The transcoder according to claim 18 , wherein
said target bit rate is changed in accordance with a free space of a storage medium for storing said second coded image and a rate control factor is recalculated in accordance with the change.
22. The transcoder according to claim 3 , wherein
said step value calculation means includes
rate control factor calculation means for calculating a rate control factor on the basis of a target bit rate of said second coded image and a current bit rate of said second coded image; and
means for multiplying a value obtained by multiplication by said step value adjustment factor, by said rate control factor.
23. The transcoder according to claim 22 , wherein
said rate control factor calculation means includes means for calculating said rate control factor on the basis of a differential value between said target bit rate and said current bit rate.
24. The transcoder according to claim 22 , wherein
said rate control factor calculation means includes means for calculating said rate control factor on the basis of a differential value between a logarithmic value of said target bit rate and a logarithmic value of said current bit rate.
25. The transcoder according to claim 22 , wherein
said rate control factor calculation means includes a look-up table for inputting said target bit rate and said current bit rate and outputting a rate control characteristic value, and
said rate control factor calculation means calculates said rate control factor on the basis of said rate control characteristic value.
26. The transcoder according to any one of claims 23 to 25 , wherein
said rate control factor calculation means includes a limiter for giving an upper limit to said rate control factor.
27. The transcoder according to any one of claims 23 to 25 , wherein
said rate control factor calculation means includes a limiter for giving a lower limit to said rate control factor.
28. The transcoder according to claim 22 , wherein
said target bit rate is changed in accordance with the capacity of a storage medium for storing said second coded image and said rate control factor is recalculated in accordance with the change.
29. The transcoder according to claim 22 , wherein
said target bit rate is changed in accordance with a free space of a storage medium for storing said second coded image and said rate control factor is recalculated in accordance with the change.
30. The transcoder according to claim 18 , wherein
said means for calculating said picture adjustment factor is a software processing.
31. The transcoder according to claim 22 , wherein
said rate control factor calculation means is a software processing.
32. A coded image conversion method comprising:
decoding, with a decoder of a transcoding apparatus, a first coded image;
calculating, with a processor of the transcoding apparatus, a macroblock evaluation value indicating the degree of dispersion in pixel values inside a macroblock of an image obtained by said decoding;
obtaining, with the processor of the transcoding apparatus, an average macroblock evaluation value by averaging said macroblock evaluation values in a frame; and
encoding, with an encoder of the transcoding apparatus, the image obtained by said decoding to generate a second coded image,
wherein said encoding includes,
acquiring an average quantization step value by averaging quantization step values in a frame of said first coded image,
calculating a weighting value for each macroblock on the basis of the difference between said macroblock evaluation value and said average macroblock evaluation value, and
calculating a converting quantization step value for each macroblock of said second coded image on the basis of a value obtained by adding said weighting value to said average quantization step value.
33. A non-transitory computer readable storage medium encoded with instructions, which when executed by a transcoding apparatus causes the transcoding apparatus to implement a method comprising:
decoding, with a decoder of a transcoding apparatus, a first coded image;
calculating, with a processor of the transcoding apparatus, a macroblock evaluation value indicating the degree of dispersion in pixel values inside a macroblock of an image obtained by said decoding;
obtaining, with the processor of the transcoding apparatus, an average macroblock evaluation value by averaging said macroblock evaluation values in a frame; and
encoding, with an encoder of the transcoding apparatus, the image obtained by said decoding to generate a second coded image,
wherein said encoding includes,
acquiring an average quantization step value by averaging quantization step values in a frame of said first coded image,
calculating a weighting value for each macroblock on the basis of the difference between said macroblock evaluation value and said average macroblock evaluation value, and
calculating a converting quantization step value for each macroblock of said second coded image on the basis of a value obtained by adding said weighting value to said average quantization step value.