Method and system of transform block processing according to quantization matrix in video coding

Patent metadata

Field	Value
Publication number	US-9560347-B2
Application number	US-201214234505-A
Country	US
Kind code	B2
Filing date	Aug 21, 2012
Priority date	Aug 23, 2011
Publication date	Jan 31, 2017
Grant date	Jan 31, 2017

Abstract

A method and system process transform blocks according to quantization matrices in a video coding system. Embodiments of the present invention derive one or more derived quantization matrices from one or more initial quantization matrices or from one previously derived quantization matrix. In one embodiment, the initial quantization matrices include a 4×4 and 8×8 quantization matrices, which can be either default or user-defined. All quantization matrices larger than 8×8 can be derived from the 4×4 and 8×8 initial quantization matrices. Non-square quantization matrices can be derived from at least one initial square quantization matrix or at least one derived square quantization matrix. Individual initial quantization matrices may be used to derive respective larger quantization matrices. Furthermore, the individual initial quantization matrices may be derived from larger quantization matrices designed for corresponding transform sizes. Syntax design also enables the quantization matrix representation.

First claim

The invention claimed is:

1. A method for processing transform blocks according to quantization matrices in a video coding system, the method comprising:
receiving, via a receiving circuit, one or more initial quantization matrices having first sizes;
deriving, using a circuit, one or more derived quantization matrices having second sizes from one of said one or more initial quantization matrices, wherein the second size is different than the first size;
receiving, via a receiving circuit, a transform block having a block size, where the transform block is associated with a picture, and wherein the block size is equal to one of the first sizes or one of the second sizes;
processing, using a processing circuit, the transform block according to respective one of said one or more initial quantization matrices if the block size is equal to said one of the first sizes; and
processing, using a processing circuit, the transform block according to respective one of said one or more derived quantization matrices if the block size is equal to said one of the second sizes;
wherein said one or more initial quantization matrices comprise one or more square initial quantization matrices, wherein said one or more derived quantization matrices comprise one or more non-square, rectangular transform block sizes.

2. The method of claim 1 , wherein said one or more initial quantization matrices comprise a 4×4 quantization matrix and an 8×8 quantization matrix, wherein the 4×4 quantization matrix and the 8×8 quantization matrix are default quantization matrices, and wherein said one or more derived quantization matrices correspond to transform block sizes larger than 8×8.

3. The method of claim 1 , wherein said one or more initial quantization matrices comprise a 4×4 quantization matrix and an 8×8 quantization matrix, wherein the 4×4 quantization matrix and the 8×8 quantization matrix are user-defined quantization matrices, and wherein said one or more derived quantization matrices correspond to transform block sizes larger than 8×8.

4. The method of claim 1 , wherein said one or more square initial quantization matrices are default quantization matrices or user-defined quantization matrices.

5. The method of claim 1 , wherein a flag is incorporated in a sequence level to indicate whether default quantization matrices are used for a respective sequence.

6. The method of claim 1 , wherein a flag is incorporated in a picture level to indicate whether sequence-level quantization matrices are used for a current picture.

7. The method of claim 1 , wherein a flag is incorporated in an adaptation level to indicate whether sequence-level quantization matrices are used for a current picture.

8. The method of claim 1 , wherein a flag is incorporated in a picture level to indicate whether sequence-level quantization matrices are derived from said one or more initial quantization matrices, and wherein said one or more initial quantization matrices comprise a 4×4 quantization matrix and an 8×8 quantization matrix.

9. The method of claim 1 , wherein a flag is incorporated in an adaptation level to indicate whether sequence-level quantization matrices are derived from said one or more initial quantization matrices, and wherein said one or more initial quantization matrices comprise a 4×4 quantization matrix and an 8×8 quantization matrix.

10. The method of claim 1 , wherein said one or more initial quantization matrices comprise a 4×4 quantization matrix and a 8×8 quantization matrix, and wherein said one or more derived quantization matrices comprise a 16×16 quantization matrix and a 32×32 quantization matrix.

11. The method of claim 1 , wherein said one or more derived quantization matrices comprise 16×4, 4×16, 32×8, and 8×32 quantization matrices.

12. The method of claim 1 , wherein syntax information is incorporated in a video bitstream to indicate that said second sizes include all rectangular sizes associated with the block size.

13. The method of claim 1 , wherein syntax information is incorporated in a video bitstream to indicate that said second sizes include all sizes associated with the block size larger than 8×8.

14. The method of claim 1 , wherein said deriving one or more derived quantization matrices converts one square initial quantization matrix to one square derived quantization matrix by either spatial frequency up-sampling and linear interpolation/extrapolation, or spatial frequency up-sampling and sample repetition, and wherein said one square initial quantization matrix is smaller than said one square derived quantization matrix.

15. The method of claim 14 , wherein the spatial frequency up-sampling comprises end-point aligned spatial frequency up-sampling.

16. The method of claim 1 , wherein said deriving one or more derived quantization matrices maps one square initial quantization matrix to low-frequency entries of one square derived quantization matrix by aligning an entry corresponding to a lowest spatial frequency of said one square initial quantization matrix and said one square derived quantization matrix, said deriving one or more derived quantization matrices generates remaining entries of said one square derived quantization matrix using data extrapolation, and wherein said one square initial quantization matrix is smaller than said one square derived quantization matrix.

17. The method of claim 1 , wherein said one or more derived rectangular quantization matrices are derived from respective square derived quantization matrix using spatial frequency decimation.

18. The method of claim 1 , wherein said one or more initial quantization matrices comprise a first 8×8 square initial quantization matrix and a second 8×8 square initial quantization matrix, wherein the first 8×8 square initial quantization matrix is converted from a 16×16 square initial quantization matrix, and wherein the second 8×8 square initial quantization matrix is converted from a 32×32 square initial quantization matrix.

19. The method of claim 1 , wherein said one or more square initial quantization matrices are converted from said respective one or more larger square initial quantization matrices based on spatial frequency decimation.

20. The method of claim 1 , wherein said one or more square initial quantization matrices are converted from said respective one or more larger square initial quantization matrices based on mapping low-frequency entries of said respective one or more larger square initial quantization matrices to said one or more square initial quantization matrices with an entry corresponding to a lowest spatial frequency of said one or more square initial quantization matrices and said respective one or more larger square initial quantization matrices aligned.

21. A video coding system, wherein transform blocks are processed according to quantization matrices, the system comprising:
a first electronic circuit configured to receive one or more initial quantization matrices having first sizes;
means deriving circuit for deriving one or more derived quantization matrices having second sizes from one of said one or more initial quantization matrices;
receiving circuit for receiving receive a transform block having a block size, wherein the block size is equal to one of the first sizes or one of the second sizes;
processing circuit for processing the transform block according to respective one of said one or more initial quantization matrices if the block size is equal to said one of the first sizes; and
second processing ci

Assignees

• Zhang Ximin
• Liu Shan
• Hfi Innovation Inc

Inventors

• Zhang Ximin
• Liu Shan

Classifications

• H04N19/126 Primary

  Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers · CPC title

• H04N19/122

  Selection of transform size, e.g. 8x8 or 2x4x8 DCT; Selection of sub-band transforms of varying structure or type · CPC title

• H04N19/70

  characterised by syntax aspects related to video coding, e.g. related to compression standards · CPC title

• H04N19/18

  the unit being a set of transform coefficients · CPC title

• H04N19/80

  Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation (H04N19/635, H04N19/86 take precedence) · CPC title