Memory optimization for video processing
View Patent ↗Memory storage requirements for digital signal processing operations, for example, motion-compensated video scan rate conversion, that produce intermediate output data, which is then used as an input to the operation, are reduced by reordering operations and organizing memory allocations in a special manner to allow intermediate output at a particular execution time, to substantially share the same memory space as the intermediate output of a previous execution time. Such a reduction in the amount of memory required for processing operations advantageously reduces cost and power consumption.
1. A method of digital signal processing, comprising: reading a first set of input data from a first memory array; processing the first set of input data to produce a first set of intermediate data; writing the first set of intermediate data to a second memory array; wherein a first portion of the second memory array overlaps a first portion of the first memory array, a second portion of the second memory array does not overlap the first memory array, and the first set of intermediate data is written to a portion of the second memory array that overlaps a portion of the first memory array from which at least a portion of the first set of input data was read.
2. The method of claim 1 , wherein the input data represents video data, and processing the first set of input data comprises performing a motion-compensation operation.
3. The method of claim 1 , wherein the processing is performed by circuitry that is co-located with the first and second memory arrays on a single chip.
4. The method of claim 1 , wherein the processing is performed by circuitry that is located within a first integrated circuit chip and the memory is located within at least a second integrated circuit chip.
5. The method of claim 2 , further comprising transferring the first set of input data to a buffer memory prior to processing.
6. A method of storing and retrieving intermediate data for processing that includes recursive use of the intermediate data, comprising: determining a plurality of memory addresses for a video field stored in a memory; determining a plurality of memory addresses for each of a plurality of input columns, the input columns having a first width, the input columns at least partially overlapping the video field; determining a plurality of memory addresses for each of a plurality of output columns, the output columns having a second width, the output columns at least partially overlapping the video field; reading a first set of input data from a first input column; producing a first set of output data based at least in part upon the first set of input data; and writing the first set of output data to the first output column; wherein the width of the second width is less than the first width, and the first output column overlaps a portion of the first input column.
7. The method of claim 6 , wherein producing a first set of output data comprises performing a motion-compensation operation.
8. The method of claim 6 , wherein the first set of input data comprises video data.
9. The method of claim 8 , wherein the first set of output data comprises motion-compensated video data.
10. A system for motion-compensated video scan rate conversion, comprising: processing circuitry; a memory coupled to the processing circuitry; a first memory array disposed within the memory; a second memory array disposed within the memory, the second memory array partially overlapping the first memory array; wherein the processing circuitry is adapted to read a first set of input data from a first portion of the first memory array, produce a first set of intermediate output data, and write the first set of intermediate output data to a first portion of the second memory array such that the first set of intermediate output data partially overlaps the first portion of the first memory array.
11. The system of claim 10 , further comprising: a third memory array disposed within the memory; a fourth memory array disposed within the memory, the fourth memory array partially overlapping the third memory array; wherein the processing circuitry is adapted to read a first set of input data from a first portion of the third memory array, produce a second set of intermediate output data, and write the second set of intermediate output data to a first portion of the fourth memory array such that the second set of intermediate output data partially overlaps the first portion of the third memory array.
12. The system of claim 11 , wherein the memory is external to the processing circuitry, and further comprising a buffer memory coupled between the memory and the processing circuitry.