Granted Patent B2

US 12,339,902 · App. 17/493,877 · Granted Jun 24, 2025

Inventors: Ohad Markus (Haifa, IL); Assaf Weissman (Moreshet, IL); Dotan David Levi (Kiryat Motzkin, IL); Eyal Frishman (Hod Hasharon, IL)

Assignee: Mellanox Technologies, Ltd

G06F16/71G06F9/5027G06T9/00

Abstract

Claims (85)

1. A system comprising:

a hardware accelerator to:

receive video data of multiple video frames;

divide each of the video frames into respective blocks;

compute encoding assist data including at least one video encoding parameter type for each of the respective blocks of each of the video frames; and

store respective portions of the encoding assist data across respective block-size-specific database tables such that one block-size-specific database table is to store contiguously the encoding assist data of first size blocks, and another block-size-specific database table is to store contiguously encoding assist data of second size blocks, the first size blocks being a different size to the second size blocks; and

an interface to provide the respective database tables to video encoding software running on a processor, wherein the processor is to:

load the one block-size-specific database table storing the encoding assist data of the first size blocks into memory and perform a first part of an encoding process responsively to the encoding assist data of the first size blocks;

load the other block-size-specific database table storing the encoding assist data of the second size blocks into memory and perform a second part of the encoding process responsively to the encoding assist data of the second size blocks; and

compress the video data into encoded video frames responsively to the encoding assist data included in the respective block-size-specific database tables.

2. The system according to claim 1 , wherein the at least one video encoding parameter type includes any one or more of the following: a motion vector; a measure of block matching; a motion vector cost; a rate distortion optimization score; an intra prediction cost; an intra prediction direction; a block distance metric; or a weighted cost score.

3. The system according to claim 1 , wherein the hardware accelerator is configured to store the respective portions of the encoding assist data across the respective database tables responsively to respective video encoding parameter types.

4. The system according to claim 1 , wherein the hardware accelerator is configured to store the respective portions of the encoding assist data across the respective database tables responsively to respective block sizes.

5. The system according to claim 1 , wherein the hardware accelerator is configured to store the respective portions of the encoding assist data across the respective database tables responsively to respective video encoding parameter types and respective block sizes.

6. The system according to claim 1 , wherein the hardware accelerator is configured to:

scan the video frames responsively to a recursive z-shaped scan pattern; and

store the encoding assist data ordered in the respective database tables responsively to the recursive z-shaped scan pattern.

7. The system according to claim 6 , wherein the recursive z-shaped scan pattern is a nested z-shaped scan pattern.

8. The system according to claim 1 , wherein:

the first table includes respective video encoding parameters of a first video encoding parameter type for selected blocks of the respective blocks of at least one of the video frames; and

the second table includes respective video encoding parameters of a second video encoding parameter type, different to the first video encoding parameter type, for the selected blocks.

9. The system according to claim 8 , wherein:

the first video encoding parameter type of the first table is a motion vector; and

the first part of the encoding process includes performing rate control responsively to motion vectors in the first table.

10. The system according to claim 8 , wherein:

the second video encoding parameter type of the second table is a measure of block matching; and

the second part of the encoding process includes identifying a scene cut responsively to measures of block matching in the second table.

11. The system according to claim 1 , wherein:

the encoding software is configured to load a third table of the respective database tables into memory and perform a third part of an encoding process responsively to data of the loaded third table;

the encoding software is configured to load a fourth table of the respective database tables into memory and perform a fourth part of the encoding process responsively to data of the loaded fourth table;

the first table includes respective video encoding parameters of a first video encoding parameter type for first selected blocks of the respective blocks of at least one of the video frames of a first block size;

the second table includes respective video encoding parameters of a second video encoding parameter type, different to the first video encoding parameter type, for the first selected blocks of the first block size;

the third table includes respective video encoding parameters of the first video encoding parameter type for second selected blocks of the respective blocks of the at least one of the video frames of a second block size, different to the first block size; and

the fourth table includes respective video encoding parameters of the second video encoding parameter type for the second selected blocks of the second block size.

12. The system according to claim 1 , wherein:

the first table includes first selected blocks of the respective blocks of at least one of the video frames of a first block size; and

the second table includes second selected blocks of the respective blocks of the at least one of the video frames of a second block size, different to the first block size.

13. The system according to claim 12 , wherein:

the first block size is larger than the second block size; and

the encoding software is configured to:

load the first table and analyze data of the first selected blocks of the first block size; and then

load the second table and analyze data of the second selected blocks of the second block size.

14. A method comprising:

receiving video data of multiple video frames;

dividing each of the video frames into respective blocks;

computing encoding assist data including at least one video encoding parameter type for each of the respective blocks of each of the video frames;

storing respective portions of the encoding assist data across respective block-size-specific database tables such that one blocks-size specific database table stores contiguously the encoding assist data for first size blocks, and another block-size-specific database table stores contiguously encoding assist data for second size blocks, the first size blocks being a different size to the second size blocks;

providing the respective database tables to video encoding software;

loading the one block-size-specific database table storing the encoding assist data of the first size blocks into memory and perform a first part of an encoding process responsively to the encoding assist data of the first size blocks;

loading the other block-size-specific database table storing the encoding assist data of the second size blocks into memory and perform a second part of the encoding process responsively to the encoding assist data of the second size blocks; and

compressing the video data into encoded video frames responsively to the encoding assist data included in the respective block-size-specific database tables.

15. The method according to claim 14 , wherein the at least one video encoding parameter type is selected from any one or more of the following: a motion vector; a rate distortion optimization score; an intra prediction cost; an intra prediction direction; a block distance metric; or a weighted cost score.

16. The method according to claim 14 , wherein the storing includes storing the respective portions of the encoding assist data across the respective database tables responsively to respective video encoding parameter types.

17. The method according to claim 14 , wherein the storing includes storing the respective portions of the encoding assist data across the respective database tables responsively to respective block sizes.

18. The method according to claim 14 , wherein the storing includes storing the respective portions of the encoding assist data across the respective database tables responsively to respective video encoding parameter types and respective block sizes.

19. The method according to claim 14 , wherein:

the first table includes respective video encoding parameters of a first video encoding parameter type for selected blocks of the respective blocks of at least one of the video frames; and

the second table includes respective video encoding parameters of a second video encoding parameter type, different to the first video encoding parameter type, for the selected blocks.

20. The method according to claim 14 , further comprising:

loading a third table of the respective database tables into memory and performing a third part of an encoding process responsively to data of the loaded third table; and

loading a fourth table of the respective database tables into memory and performing a fourth part of the encoding process responsively to data of the loaded fourth table, and wherein:

the first table includes respective video encoding parameters of a first video encoding parameter type for first selected blocks of the respective blocks of at least one of the video frames of a first block size;

the second table includes respective video encoding parameters of a second video encoding parameter type, different to the first video encoding parameter type, for the first selected blocks of the first block size;

the third table includes respective video encoding parameters of the first video encoding parameter type for second selected blocks of the respective blocks of the at least one of the video frames of a second block size, different to the first block size; and

the fourth table includes respective video encoding parameters of the second video encoding parameter type for the second selected blocks of the second block size.

21. The method according to claim 14 , wherein:

the first table includes first selected blocks of the respective blocks of at least one of the video frames of a first block size; and

the second table includes second selected blocks of the respective blocks of the at least one of the video frames of a second block size, different to the first block size.

22. The method according to claim 21 , wherein the first block size is larger than the second block size, the method further comprising:

loading the first table and analyzing data of the first selected blocks of the first block size; and then

loading the second table and analyzing data of the second selected blocks of the second block size.

23. The method according to claim 14 , further comprising:

scanning the video frames responsively to a recursive z-shaped scan pattern; and

storing the encoding assist data ordered in the respective database tables responsively to the recursive z-shaped scan pattern.

24. The method according to claim 23 , wherein the recursive z-shaped scan pattern is a nested z-shaped scan pattern.

25. A system comprising:

a hardware accelerator to:

receive video data of multiple video frames;

divide each of the video frames into respective blocks;

compute encoding assist data including at least one video encoding parameter type for each of the respective blocks of each of the video frames; and

store respective portions of the encoding assist data across respective video-encoding-parameter-type-specific database tables such that one video-encoding-parameter-type-specific database table is to store contiguously the encoding assist data of a first video encoding parameter type, and another video-encoding-parameter-type-specific database table is to store contiguously encoding assist data of a second video encoding parameter type, the first video encoding parameter type being different from the second video encoding parameter type; and

an interface to provide the respective database tables to video encoding software running on a processor, wherein the processor is to:

load the one video-encoding-parameter-type-specific database table storing the encoding assist data of the first video encoding parameter type into memory and perform a first part of an encoding process responsively to the encoding assist data of the first video encoding parameter type;

load the other video-encoding-parameter-type-specific database table storing the encoding assist data of the video encoding parameter type into memory and perform a second part of the encoding process responsively to the encoding assist data of the video encoding parameter type; and

compress the video data into encoded video frames responsively to the encoding assist data included in the respective video-encoding-parameter-type-specific database tables.

Assignments (1)

Continuity (1)

References Cited (131)