Method and apparatus for QP modulation based on perceptual models for picture encoding
View Patent ↗A method for encoding a picture is disclosed. The method generally includes the steps of (A) generating at least one respective macroblock statistic from each of a plurality of macroblocks in the picture, (B) generating at least one global statistic from the picture and (C) generating a respective macroblock quantization parameter for each of the macroblocks based on both (i) the at least one respective macroblock statistic and (ii) said at least one global statistic.
1. A method for encoding a picture, comprising the steps of:
(A) generating at least one respective macroblock statistic from each of a plurality of macroblocks in said picture, the at least one respective macroblock statistic being associated with characteristics of the Human Visual System (HVS);
(B) generating at least one global statistic from said picture, the at least one global statistic being an average value generated by averaging the at least one respective macroblock statistic for the plurality of macroblocks over the picture; and
(C) generating a respective macroblock quantization parameter for each of said macroblocks based on both (i) said at least one respective macroblock statistic and (ii) said at least one global statistic,
wherein each of said respective macroblock quantization parameters comprises a respective delta quantization parameter generated only from said at least one respective macroblock statistic and said at least one global statistic.
2. The method according to claim 1 , wherein (i) said at least one respective macroblock statistic comprises a luminance motion value and (ii) said at least one global statistic comprises an average luminance motion value for said macroblocks averaged over said picture.
3. The method according to claim 1 , wherein (i) said at least one respective macroblock statistic comprises a luminance DC value and (ii) said at least one global statistic comprises an average luminance DC value for said macroblocks averaged over said picture.
4. The method according to claim 1 , wherein (i) said at least one respective macroblock statistic comprises a luminance high-frequency value and (ii) said at least one global statistic comprises an average luminance high-frequency value for said macroblocks averaged over said picture.
5. The method according to claim 1 , wherein (i) said at least one respective macroblock statistic comprises a spatial edge strength value and (ii) said at least one global statistic comprises an average spatial edge strength value for said macroblocks averaged over said picture.
6. The method according to claim 1 , wherein (i) said at least one respective macroblock statistic comprises a temporal edge strength value and (ii) said at least one global statistic comprises an average temporal edge strength value for said macroblocks averaged over said picture.
7. The method according to claim 1 , wherein each of said respective macroblock quantization parameters is further based on a picture quantization parameter.
8. The method according to claim 1 , further comprising the step of:
generating a respective modulation factor for each of said macroblocks based on both (i) said at least one respective macroblock statistic and (ii) said at least one global statistic, wherein in step (C) said respective macroblock quantization parameter is generated based on said respective modulation factor.
9. The method according to claim 8 , wherein the step of generating said respective modulation factor is calculated using (i) a first equation if said at least one respective macroblock statistic is greater than said at least one global statistic and (ii) a second equation if said at least one respective macroblock statistic is less than said at least one global statistic.
10. The method according to claim 8 , further comprising the step of:
spatially filtering said respective modulation factors prior to generating said respective macroblock quantization parameters.
11. The method according to claim 8 , further comprising the step of:
temporally filtering said respective modulation factors prior to generating said respective macroblock quantization parameters.
12. The method according to claim 1 , further comprising the step of:
spatially filtering said respective macroblock quantization parameters.
13. The method according to claim 1 , further comprising the step of:
temporally filtering said respective macroblock quantization parameters.
14. The method according to claim 1 , wherein at least one of said at least one global statistics comprises a fixed value.
15. The method according to claim 1 , further comprising the step of:
spatially processing said picture before generating said at least one respective macroblock statistic.
16. The method according to claim 1 , wherein said picture forms a portion of a video stream.
17. The method according to claim 1 , wherein said picture comprises a still image.
18. A system comprising:
a processor configured to (i) generate at least one respective macroblock statistic from each of a plurality of macroblocks in a picture and (ii) generate at least one global statistic from said picture, the at least one respective macroblock statistic is associated with characteristics of the Human Visual System (HVS), and the at least one global statistic is an average value generated by averaging the at least one respective macroblock statistic for the plurality of macroblocks over the picture; and
an encoder configured to generate a respective macroblock quantization parameter for each of said macroblocks based on both (i) said at least one respective macroblock statistic and (ii) said at least one global statistic,
wherein each of said respective macroblock quantization parameters comprises a respective delta quantization parameter generated only from said at least one respective macroblock statistic and said at least one global statistic.
19. The system according to claim 18 , wherein said encoder is further configured to generate a respective macroblock quantization step size for each of said macroblocks based on said respective macroblock quantization parameter.
20. The system according to claim 19 , further comprising a memory configured to store a lookup table for converting said respective macroblock quantization step sizes to said respective macroblock quantization parameters.
21. The system according to claim 18 , wherein said at least one respective macroblock statistic comprises a texture value.
22. The system according to claim 18 , wherein said at least one respective macroblock statistic comprises at least one of (i) a luminance motion value, (ii) a luminance DC value, (iii) a luminance high-frequency value, (iv) a spatial edge strength value and (v) a temporal edge strength value.
23. The system according to claim 18 , wherein said encoder is further configured to encode said picture into an output signal compliant with at least one of an International Organization for Standardization/International Electrotechnical Commission 14496-10 standard and an International Telecommunication Union-Telecommunications Revised Recommendation H.264.
24. A system comprising:
means for generating at least one respective macroblock statistic from each of a plurality of macroblocks in a picture the at least one respective macroblock statistic is associated with characteristics of the Human Visual System (HVS),
means for generating at least one global statistic from said picture, and the at least one global statistic is an average value generated by averaging the at least one respective macroblock statistic for the plurality of macroblocks over the picture; and
means for generating a respective macroblock quantization parameter for each of said macroblocks based on both (i) said at least one respective macroblock statistic and (ii) said at least one global statistic,
wherein each of said respective macroblock quantization parameters comprises a respective delta quantization parameter generated only from said at least one respective macroblock statistic and said at least one global statistic.