Control of video compression based on file size constraint
View Patent ↗A method for processing data includes receiving an image sequence including a succession of image frames for encoding in an image sequence file. A target size of the image sequence file is specified. The image sequence file is generated by encoding the succession of the image frames using at least one variable compression parameter, which is adjusted during the encoding responsively to the target size. The image sequence file is outputted for transmission over a communication link.
1. A method for processing data, comprising:
receiving an image sequence comprising a succession of image frames for encoding in an image sequence file;
specifying a target size of the image sequence file;
generating the image sequence file by performing for each current frame of the succession:
monitoring a cumulative data volume of the compressed image frames preceding the current frame;
calculating a budget for the current frame responsively to the cumulative data volume and the target size of the image sequence file;
selecting at least one variable compression parameter for the current frame, responsively to the budget calculated for the current frame; and
encoding the current frame responsively to the at least one selected variable compression parameter; and
outputting the image sequence file for transmission over a communication link.
2. The method according to claim 1 , wherein the at least one variable compression parameter comprises a quantization parameter.
3. The method according to claim 2 , wherein encoding the succession of the image frames comprises dividing each of the image frames into a plurality of basic units, and setting the quantization parameter respectively for each of the basic units in each of the image frames.
4. The method according to claim 1 , wherein the at least one variable compression parameter comprises a skip frame parameter, which is set so as to cause at least one of the image frames to be dropped from the image sequence file.
5. The method according to claim 1 , wherein encoding the succession of the image frames comprises:
defining a virtual buffer size responsively to the target size;
defining a buffer overflow curve, which decreases monotonically over at least a final portion of the succession of the image frames, indicating an amount by which a data volume at each frame in the image sequence file is permitted to exceed a working point indicated by the virtual buffer size; and
controlling the variable compression parameter while encoding the data so that the data volume remains within a bound indicated by the buffer overflow curve.
6. The method according to claim 5 , wherein controlling the variable compression parameter comprises dropping a given frame from the image sequence file upon determining that inclusion of the given frame in the image sequence file will cause the data volume to exceed the bound indicated by the buffer overflow curve for the given frame.
7. The method according to claim 1 , wherein encoding the succession of the image frames comprises compressing the image sequence in a single pass through the succession of the image frames.
8. The method according to claim 1 , wherein the file size is no less than 98% of the target size.
9. The method according to claim 1 , wherein said image sequence file having a file size that is no less than 90% of said target size, and, responsive to said image sequence file being greater than said target size, discarding at least one of said encoded image frames of said image sequence file thereby reducing the size of said image sequence file to be less than said target size.
10. The method according to claim 1 , wherein calculating the budget for the current frame comprises calculating a cumulative deviation of the cumulative data volume from a nominal frame budget determined as the file target size divided by a number of images in the succession of frames.
11. The method according to claim 1 , wherein outputting the image sequence file comprises outputting after compressing the entire succession of image frames.
12. Apparatus for processing data, comprising:
a memory, which is configured to receive an image sequence comprising a succession of image frames for encoding in an image sequence file; and
a processor, which is coupled to generate the image sequence file by performing for each current frame of the succession:
monitor a cumulative data volume of the compressed image frames preceding the current frame;
calculate a budget for the current frame responsively to the cumulative data volume and the target size of the image sequence file;
select at least one variable compression parameter for the current frame, responsively to the budget calculated for the current frame; and
encode the current frame responsively to the at least one selected variable compression parameter, and
to output the image sequence file for transmission over a communication link.
13. The apparatus according to claim 12 , wherein the at least one variable compression parameter comprises a quantization parameter.
14. The apparatus according to claim 13 , wherein the processor is configured to divide each of the image frames into a plurality of basic units, and to set the quantization parameter respectively for each of the basic units in each of the image frames.
15. The apparatus according to claim 12 , wherein the at least one variable compression parameter comprises a skip frame parameter, which is set by the processor so as to cause at least one of the image frames to be dropped from the image sequence file.
16. The apparatus according to claim 12 , wherein the processor is configured to encode the succession of the image frames by defining a virtual buffer size responsively to the target size, defining a buffer overflow curve indicating an amount by which a data volume at each frame in the image sequence file is permitted to exceed a working point indicated by the virtual buffer size, and controlling the variable compression parameter while encoding the data so that the data volume remains within a bound indicated by the buffer overflow curve.
17. The apparatus according to claim 16 , wherein the bound indicated by the buffer overflow curve decreases monotonically over at least a final portion of the succession of the image frames.
18. The apparatus according to claim 16 , wherein the processor is configured to drop a given frame from the image sequence file upon determining that inclusion of the given frame in the image sequence file will cause the data volume to exceed the bound indicated by the buffer overflow curve for the given frame.
19. The apparatus according to claim 12 , wherein the processor is configured to encode the succession of the image frames in a single pass through the succession of the image frames.
20. The apparatus according to claim 12 , wherein the file size is no less than 98% of the target size.
21. A computer software product, comprising a non-transitory computer-readable medium in which program instructions are stored, which instructions, when read by a computer, cause the computer to receive an image sequence comprising a succession of image frames for encoding in an image sequence file, and to generate the image sequence file by performing for each current frame of the succession:
monitoring a cumulative data volume of the compressed image frames preceding the current frame;
calculating a budget for the current frame responsively to the cumulative data volume and the target size of the image sequence file;
selecting at least one variable compression parameter for the current frame, the budget calculated for the current frame, and
encoding the current frame responsively to the at least one selected variable compression parameter; and
to output the image sequence file for transmission over a communication link.
22. The product according to claim 21 , wherein the at least one variable compression parameter comprises a quantization parameter.
23. The product according to claim 21 , wherein the at least one variable compression parameter comprises a skip frame parameter, which is set by the computer so as to cause at least one of the image frames to be dropped from the image sequence file.