Methods for fixed rate block based compression of image data

摘要

A set of methods, techniques and hardware is described for compressing image data for memory bandwidth and memory storage reduction in graphics processing systems. The disclosed technology can be used for compressing image data sent to the frame buffer and/or image data residing in the frame buffer. The compression process can be based on an adaptive number of base color points and an adaptive number of quantized color points. An adaptive technique for compressing alpha values based on pre-calculated maps or using an estimated alpha value based on thresholds is also disclosed. An implementation of the disclosed methods has, for example, a low hardware overhead, low buffering requirements, and low and predefined compression latency. Also, the disclosed methods allow, for example, random accesses to compressed image data.

主权项

1. In a graphics processing system, a method for compressing source image data comprising:
dividing the source image into image blocks of a fixed number of pixels, each pixel having an associated color and alpha information; encoding each image block in a compressed form of the image block including:
selecting an adaptive number of representative color components, the representative color components encoded with at least two precisions in the compressed form, the representative color components associated with pixels of the image block,calculating an adaptive number of additional color components derived from the selected representative color components,generating an index for each pixel in the image block,associating the index with one representative color component or with one color component derived from the representative color components, wherein the index may be split in two partitions, and one partition may be used to select two of the at least two representative color components, andselecting an adaptive number of alpha values for the image block wherein either a unified, computed alpha value is assigned to each pixel in the image block, or different pre-calculated alpha values are assigned to at least two pixels in the image block; andcomposing the compressed forms of all blocks comprising the source image to generate a compressed image.