Techniques for calculation of quantization matrices in video coding

1. A method, comprising:

defining a first quantization matrix from a second quantization matrix by scaling quantization values of the second quantization matrix by scaling parameters that vary proportionately with camera exposure value of a capture operation that generated an image,

communicating values of the first quantization matrix to a decoder, and

predictively coding a portion of the image, wherein the coding comprises:

transforming pixel data of the image to a plurality of frequency domain coefficients, and

quantizing the frequency domain coefficients according to the first quantization matrix.

2. The method of claim 1 , wherein the exposure value is an image-level exposure value of the image.

3. The method of claim 1 , wherein the exposure value is an exposure value of a region of the image, smaller than the image.

4. The method of claim 1 , wherein the first quantization matrix definition is further based on an estimated signal-to-noise ratio of the image.

5. The method of claim 4 , wherein the estimated signal-to-noise ratio is an image-level signal-to-noise value.

6. The method of claim 4 , wherein the estimated signal-to-noise ratio is a signal-to-noise ratio of a region of the image, smaller than the image.

7. The method of claim 1 , wherein the first quantization matrix definition is further based on a resolution of the image as it is coded.

8. The method of claim 1 , wherein the scaling parameters increase according to a distance between each matrix position and a matrix origin.

9. The method of claim 1 , wherein values of the first quantization matrix are communicated in a picture parameter dataset message of a governing coding protocol.

10. The method of claim 1 , wherein values of the first quantization matrix are communicated differentially with respect to the second quantization matrix.

11. The method of claim 1 , wherein the image is a part of a multi-frame video sequence and the method repeats for a plurality of frames of the video sequence.

12. The method of claim 11 , wherein the first quantization matrix definition is further based on a frame rate of the video sequence as it is coded.

13. The method of claim 11 , wherein values of the first quantization matrix are communicated during a negotiation phase of a coding session.

14. The method of claim 11 , wherein values of the first quantization matrix are communicated during a video exchange phase of a coding session.

15. The method of claim 11 , wherein

the defining and communicating are repeated for a plurality of quantization matrices that are communicated to the decoder, and

coded video data of different coded portions of the video sequence include identifiers of the quantization matrices that were used, respectively, for the quantizing of the different portions of the video sequence.

16. A non-transitory computer readable medium having stored thereon program instructions that, when executed by a processing device, cause the device to:

define a first quantization matrix from a second quantization matrix by scaling quantization values of the second quantization matrix by scaling parameters that vary proportionately with camera exposure value of a capture operation that generated an image,

communicate values of the first quantization matrix to a decoder, and

predictively code a portion of the image, wherein the coding comprises:

transforming pixel data of the image to a plurality of frequency domain coefficients, and

quantizing the frequency domain coefficients according to the first quantization matrix.

17. The medium of claim 16 , wherein the first quantization matrix definition is further based on an estimated signal-to-noise ratio of the image.

18. The medium of claim 16 , wherein the first quantization matrix definition is further based on a resolution of the image as it is coded.

19. The medium of claim 16 , wherein the image is part of a multi-frame video sequence and the first quantization matrix definition is further based on a frame rate of the video sequence as it is coded.

20. The medium of claim 16 , wherein the scaling parameters increase according to a distance between each matrix position and a matrix origin.

21. The medium of claim 16 , wherein values of the first quantization matrix are communicated in a picture parameter dataset message of a governing coding protocol.

22. The medium of claim 16 , wherein values of the first quantization matrix are communicated differentially with respect to the second quantization matrix.

23. Apparatus, comprising:

a pixel block coder, comprising a transform unit having an input for pixel block data of an input image and an output for transform-domain coefficients, and a quantizer, having inputs for the transform-domain coefficients and for values of a quantization matrix, and an output for quantized transform-domain coefficients;

a controller to define a first quantization matrix from a second quantization matrix by scaling quantization values of the second quantization matrix by scaling parameters that vary proportionately with camera exposure value of a capture operation that generated an image.

24. The method of claim 1 , where elements of the second quantization matrix are defined by

H i,j ′=A i,j ·H i,j

where (i, j) indicates which element of a matrix, H′ is the first quantization matrix, H is the second quantization matrix, and A is a matrix of the linear scaling parameters.

25. The method of claim 1 , wherein the characteristic is a camera exposure value of a capture operation that generated the image, and the linear scaling parameters vary proportionally with the camera exposure value.

26. The method of claim 1 , wherein the characteristic is an estimated signal-to-noise ratio of the image, and the linear scaling parameters vary inversely proportionally with the estimated signal-to-noise ratio.

27. The method of claim 1 , wherein the scaling parameters are linear scaling parameters.