Video compression system
An improved video compression system in which the coefficient transform is sped up via folding transposition of coefficients into the token extraction phase. This is accomplished by filling a group of blocks coefficient buffers with 0's before the start of coefficient decoding, extracting the token from the bitstream and placing any extracted coefficient value which is non zero into the transposed order that best suits the target processor.
1. An encoder for compressing digital video signals comprising:
a motion estimator;
a block based discrete cosine transform system connected to said motion estimator;
a quantizer connected to said discrete transform system, said quantizer producing separate image artifacts;
a variable length encoder connected to said quantizer;
a reconstruction buffer connected to said variable length encoder; and
a loop filter connected to said reconstruction buffer for smoothing block edges therein,
wherein said quantizer includes separate but functionally related filters which remove said separate image artifacts, said separate but functionally related filters including:
a de-blocking filter reducing blocking artifacts which are produced by said quantizer; and
an edge-enhancement and de-ringing filter to reduce ringing artifacts produced by said quantizer, said edge-enhancement and de-ringing filter sharpening said text images.
2. A system of separate but functionally related filters which remove separate image artifacts left over from a video compression and decompression process, comprising:
a de-blocking filter reducing blocking artifacts which are produced by a quantizer in a reconstruction frame; and
an edge-enhancement and deranging filter that reduces ringing artifacts produced by the quantizer, said edge-enhancement and de-ringing filter sharpening said images,
wherein one of said filters utilizes information determined by the other of said filters.
3. A system according to claim 2 , wherein said edge-enhancement filter utilizes information determined by said de-blocking filter, said utilized information comprising a block's adaptive strength value determined by said de-blocking filter.
4. A system according to claim 2 , wherein said de-blocking filter utilizes a value calculated by said edge-enhancement filter.
5. An improved video decompression process comprising the steps of:
filling substantially all of a video fiarne's coefficient buffers with 0's before starting coefficient decoding;
extracting a coefficient of a transform for a token from a bitstream in zig zag order;
placing coefficient values which are non-zero into a ordered coefficient buffer for case of use in an inverse transform, the order of the non-zero coefficient values dependent upon a type of processor that will be reconstructing the video frame.
6. An improved video decompression process according to claim 5 , wherein the order of the non-zero coefficient values comprises one of raster order, transposed order, or transposed raster order.
7. An improved decompression process according to claim 6 , wherein the order of the non-zero-coefficients is a transposed order, wherein an additional transpose is not required in order to speed up vector processing.
8. An improved video decompression process according to claim 6 , wherein the order of the non-zero coefficient values is a transposed raster order determined by the type of processor used, further comprising the step of minimizing a number of operations of the decompression process by localizing transpositions.
9. An improved video decompression process according to claim 8 , wherein
if the processor is not a vector processor, the order of the non-zero coefficient values is a raster order.
10. An improved video decompression process according to claim 8 , wherein
if the processor is a 2-element vector processor, the order of the non-zero coefficient values is a raster order with each 2×2 block transposed.
11. An improved video decompression process according to claim 8 , wherein
if the processor is 4-element vector processor, the order of the non-zero coefficient values is a raster order with each 4×4 block transposed.
12. An improved video decompression process according to claim 8 , wherein
if the processor is an 8-element vector processor, the order of the non-zero coefficient values is a raster order with each 8×8 block transposed.
13. An improved video decompression process according to claim 8 , wherein
if the processor is an n-element vector processor, the order of the non-zero coefficient values is a raster order with each n×n block transposed.
14. An improved video decompression process according to claim 5 , further comprising the steps of:
performing an inverse transform on the coefficient buffer to convert the coefficients to one or more of pixel values or error signal values; and
reconstructing the image from a predictor block and said pixel values or error signal values.