Subjective visual quality enhancement for high spatial and temporal complexity video encode

Patent metadata

Field	Value
Publication number	US-11856205-B2
Application number	US-201916692516-A
Country	US
Kind code	B2
Filing date	Nov 22, 2019
Priority date	Nov 22, 2019
Publication date	Dec 26, 2023
Grant date	Dec 26, 2023

Abstract

Techniques related to improved visual quality for high spatial and temporal complexity video encoding are discussed. Such techniques include ranking candidate coding structures based on rate distortion values generated using a first distortion measurement technique, detecting candidate coding structures with large coding unit and transform sizes, and disabling detected candidate coding structures with a distortion, generated using a second distortion measurement technique, that meets or exceeds a threshold.

First claim

What is claimed is:

1. A device for video coding comprising:
a memory to store a video; and
a processor coupled to the memory, the processor to:
determine a first candidate coding structure for a portion of a video frame from a plurality of candidate coding structures based on a first rate distortion value corresponding to the first candidate coding structure being a minimum rate distortion value of a plurality of rate distortion values each corresponding to one of the plurality of candidate coding structures, wherein the first candidate coding structure comprises a coding unit having a coding unit size, a coding mode for the coding unit, and a transform unit size for the coding unit, and wherein each of the plurality of rate distortion values comprises a distortion value determined by application of a first distortion measurement technique to the portion, including a first distortion value corresponding to the first candidate coding structure;
compare, in response to the coding unit size meeting or exceeding a coding unit size threshold and the transform unit size meeting or exceeding a transform unit size threshold, a second distortion value to a threshold distortion value, wherein the second distortion value corresponds to a second distortion measurement technique applied to the portion based on the first candidate coding structure;
disable the first candidate coding structure in response to the second distortion value exceeding the threshold distortion value; and
encode the portion using a second candidate coding structure to generate a bitstream, wherein the second candidate coding structure has a second rate distortion value that is greater than the first rate distortion value.

2. The device of claim 1 , wherein the first distortion measurement technique comprises a sum of squared errors applied to a residual block corresponding to the portion and the second distortion measurement technique comprises one of a sum of absolute differences or a sum of absolute transform differences applied to the residual block.

3. The device of claim 1 , wherein the coding unit size threshold comprises a 32×32 coding unit size and the transform unit size threshold comprises a 16×16 transform unit size.

4. The device of claim 1 , wherein the second candidate coding structure comprises the coding unit and a second transform unit size that is smaller than the transform unit size responsive to the second distortion value exceeding the threshold distortion value.

5. The device of claim 1 , wherein the second candidate coding structure comprises a second coding unit having a second coding unit size that is smaller than the coding unit size responsive to the second distortion value exceeding the threshold distortion value.

6. The device of claim 1 , wherein the processor to determine the first candidate coding structure comprises the processor to:
generate a plurality of candidate coding structures for the portion and a corresponding plurality of rate distortion values based on the first distortion measurement technique; and
select the first candidate coding structure in response to the first rate distortion value being a minimum of the plurality of rate distortion values.

7. The device of claim 1 , wherein the coding mode comprises an inter coding mode and the processor is further to:
generate, for the portion, a third candidate coding structure comprising a second coding unit having the coding unit size, an intra coding mode for the second coding unit, and the transform unit size for the second coding unit using a third distortion value for the third candidate coding structure that corresponds to the first distortion measurement technique applied to the portion based on the third candidate coding structure;
compare, in response to the coding unit size meeting or exceeding the coding unit size threshold and the transform unit size meeting or exceeding the transform unit size threshold for the third candidate coding structure, a fourth distortion value for the third candidate coding structure to a second threshold distortion value, wherein the fourth distortion value corresponds to the second distortion measurement technique applied to the portion based on the third candidate coding structure; and
disable the third candidate coding structure in response to the fourth distortion value exceeding the second threshold distortion value, wherein the second threshold distortion value is less than the threshold distortion value in response to the second coding unit having the intra coding mode and the coding unit having the inter coding mode.

8. The device of claim 1 , wherein the portion consists of the coding unit and the processor is further to:
generate, for the portion, a third candidate coding structure comprising a second coding unit having a second coding unit size less than the coding unit size, a second coding mode for the second coding unit, and a second transform unit size for the second coding unit using a third distortion value for the third candidate coding structure that corresponds to the first distortion measurement technique applied to the portion based on the third candidate coding structure;
compare, in response to the second coding unit size meeting or exceeding the coding unit size threshold and the second transform unit size meeting or exceeding the transform unit size threshold, a fourth distortion value for the third candidate coding structure to a second threshold distortion value, wherein the fourth distortion value corresponds to the second distortion measurement technique applied to the portion based on the third candidate coding structure; and
disable the third candidate coding structure in response to the fourth distortion value exceeding the second threshold distortion value, wherein the second threshold distortion value is greater than the threshold distortion value in response to the second coding unit size being less than the coding unit size.

9. The device of claim 1 , wherein the portion consists of the coding unit and the processor is further to:
generate, for the portion, a third candidate coding structure comprising a second coding unit having the coding unit size, a second coding mode for the second coding unit, and a second transform unit size for the second coding unit that is less than the transform unit size using a third distortion value for the third candidate coding structure that corresponds to the first distortion measurement technique applied to the portion based on the third candidate coding structure;
compare, in response to the coding unit size meeting or exceeding the coding unit size threshold and the second transform unit size meeting or exceeding the transform unit size threshold, a fourth distortion value for the third candidate coding structure to a second threshold distortion value, wherein the fourth distortion value corresponds to the second distortion measurement technique applied to the portion based on the third candidate coding structure; and
disable the third candidate coding structure in response to the fourth distortion value exceeding the second threshold distortion value, wherein the second threshold distortion value is greater than the threshold distortion value in response to the second transform unit size being less than the transform unit size.

10. The device of claim 9 , the processor further to:
generate, for the portion, a fourth candidate coding structure comprising a third coding unit having a second coding unit size less than the coding unit size, a third coding mode for the third coding unit, and the second transform unit size for the third coding unit using a fifth distortion value for the fourth candidate coding structure that corresponds to the first

Assignees

• Intel Corp

Inventors

• Tanner Jason

Classifications

• H04N19/147 Primary

  according to rate distortion criteria (rate-distortion as a criterion for motion estimation H04N19/567) · CPC title

• H04N19/103 Primary

  Selection of coding mode or of prediction mode · 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

• H04N19/14

  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