Content and quality adaptive wavefront split for parallel video coding

Patent metadata

Field	Value
Publication number	US-12101479-B2
Application number	US-202016828650-A
Country	US
Kind code	B2
Filing date	Mar 24, 2020
Priority date	Mar 24, 2020
Publication date	Sep 24, 2024
Grant date	Sep 24, 2024

Abstract

Techniques related to parallel block video coding using an adaptive wavefront split are discussed. Such techniques include adaptively applying a wavefront split in a video frame based on a frame level quantization parameter of the video frame, an encode duration of a previous frame, and/or image content of the video frame.

First claim

What is claimed is:

1. A device for video coding comprising:
a memory to store video; and
one or more processors coupled to the memory, the one or more processors to:
determine a first frame level quantization parameter (QP) for a first frame of the video and a second frame level QP for a second frame of the video, wherein the second frame level QP is less than the first frame level QP;
apply, during encode of the first and second frames, a first wavefront split to divide the first frame into first and second regions for parallel encode and a second wavefront split to divide the second frame into third and fourth regions for parallel encode, respectively, wherein the fourth region is larger than the second region in response to the second frame level QP being less than the first frame level QP; and
output a bitstream comprising the encoded first and second frames.

2. The device of claim 1 , wherein the first and second regions are first pre- and post-wavefront split regions of the first frame and the third and fourth regions are second pre- and post-wavefront split regions of the second frame.

3. The device of claim 1 , the one or more processors to:
apply, during the encode of the first and second frames, a third wavefront split to further divide the first region into fifth and sixth regions for parallel encode and a fourth wavefront split to further divide the third region into seventh and eighth regions for parallel encode, wherein the seventh region is larger than the fifth region in response to the second frame level QP being less than the first frame level QP.

4. The device of claim 3 , wherein the sixth region is between the fifth and second regions and the eighth region is between the seventh and fourth regions.

5. The device of claim 3 , wherein an area of the eighth region is within two percent of a sum of areas of the seventh and fourth regions in response to the second frame level QP being less than a threshold QP.

6. The device of claim 1 , the one or more processors to:
disable, during encode of a third frame of the video, application of wavefront splitting to the third frame in response to a third frame level QP of the third frame being greater than a threshold QP.

7. The device of claim 1 , wherein the third and fourth regions have the same size in response to the second frame level QP being less than a threshold QP.

8. The device of claim 1 , the one or more processors to:
determine a duration of the encode of the first frame is outside of a range between a lower frame encode duration threshold and an upper frame encode duration threshold; and
apply, during encode of a third frame subsequent to the second frame, a third wavefront split to divide the third frame into fifth and sixth regions for parallel encode, wherein the sixth region is larger than the second region in response to the duration being greater than the upper frame encode duration threshold or smaller than the second region in response to the duration being less than the lower frame encode duration threshold.

9. The device of claim 1 , the one or more processors to:
detect, in a third frame, a boundary between a low complexity content region and a high complexity content region; and
apply, during encode of the third frame, a third wavefront split along the detected boundary to divide the third frame into fifth and sixth regions for parallel encode.

10. The device of claim 9 , wherein the boundary is detected based on comparisons of complexity measures between neighboring blocks of the third frame.

11. The device of claim 1 , the one or more processors to:
perform between block comparisons of complexity measures of neighboring blocks of the first frame to determine all neighboring block complexity measurement differences of the first frame are less than a threshold, wherein the first wavefront split is applied responsive to all neighboring block complexity measurement differences being less than the threshold.

12. A method for video coding, comprising:
determining a first frame level quantization parameter (QP) for a first frame of video and a second frame level QP for a second frame of the video, wherein the second frame level QP is less than the first frame level QP;
applying, during encode of the first and second frames, a first wavefront split to divide the first frame into first and second regions for parallel encode and a second wavefront split to divide the second frame into third and fourth regions for parallel encode, respectively, wherein the fourth region is larger than the second region in response to the second frame level QP being less than the first frame level QP; and
outputting a bitstream comprising the encoded first and second frames.

13. The method of claim 12 , further comprising:
applying, during the encode of the first and second frames, a third wavefront split to further divide the first region into fifth and sixth regions for parallel encode and a fourth wavefront split to further divide the third region into seventh and eighth regions for parallel encode, wherein the seventh region is larger than the fifth region in response to the second frame level QP being less than the first frame level QP.

14. The method of claim 12 , further comprising:
disabling, during encode of a third frame of the video, application of wavefront splitting to the third frame in response to a third frame level QP of the third frame being greater than a threshold QP.

15. The method of claim 12 , further comprising:
determining a duration of the encode of the first frame is outside of a range between a lower frame encode duration threshold and an upper frame encode duration threshold; and
applying, during encode of a third frame subsequent to the second frame, a third wavefront split to divide the third frame into fifth and sixth regions for parallel encode, wherein the sixth region is larger than the second region in response to the duration being greater than the upper frame encode duration threshold or smaller than the second region in response to the duration being less than the lower frame encode duration threshold.

16. The method of claim 12 , further comprising:
detecting, in a third frame, a boundary between a low complexity content region and a high complexity content region; and
applying, during encode of the third frame, a third wavefront split along the detected boundary to divide the third frame into fifth and sixth regions for parallel encode.

17. At least one non-transitory machine readable medium comprising a plurality of instructions that, in response to being executed on a computing device, cause the computing device to code video by:
determining a first frame level quantization parameter (QP) for a first frame of video and a second frame level QP for a second frame of the video, wherein the second frame level QP is less than the first frame level QP;
applying, during encode of the first and second frames, a first wavefront split to divide the first frame into first and second regions for parallel encode and a second wavefront split to divide the second frame into third and fourth regions for parallel encode, respectively, wherein the fourth region is larger than the second region in response to the second frame level QP being less than the first frame level QP; and
outputting a bitstream comprising the encoded first and second frames.

18. The non-transitory machine readable medium of claim 17 , further comprising instructions that, in response to being executed on the computing device, cause the computing device to code video by:

Assignees

• Intel Corp

Inventors

• Tanner Jason
• Plotnikov Sergei
• Buxton Mark

Classifications

• H04N19/172

  the region being a picture, frame or field · CPC title

• H04N19/14

  Coding unit complexity, e.g. amount of activity or edge presence estimation (H04N19/146 takes precedence) · CPC title

• H04N19/176

  the region being a block, e.g. a macroblock · CPC title

• H04N19/124 Primary

  Quantisation · CPC title

• H04N19/436

  using parallelised computational arrangements · CPC title