Flexible quantization
A digital media encoder/decoder uses a flexible quantization technique that provides the ability to vary quantization along various dimensions of the encoded digital media data, including spatial, frequency sub bands and color channels. The codec utilizes a signaling scheme to signal various permutations of flexible quantization combinations efficiently for primary usage scenarios. When a choice of quantizer is available, the codec efficiently encodes the current quantizer by defining a subset of quantizers and indexes the current quantizer from the set.
1. A computer system comprising one or more processing units and memory, wherein the computer system implements a video decoder configured to perform operations comprising:
receiving encoded data in a bitstream, wherein the encoded data includes:
one or more syntax elements that control variation of quantization parameters along a spatial dimension within a frame of pixel values, the pixel values of the frame being organized spatially into multiple areas within the frame, the multiple areas being further organized spatially into groups, the one or more syntax elements that control variation of the quantization parameters along the spatial dimension including at least one syntax element that indicates whether variation of the quantization parameters is possible between the multiple areas within the respective groups;
one or more syntax elements that control variation of the quantization parameters along a color channel dimension, the pixel values of the frame also being organized by first, second, and third color channels, the one or more syntax elements that control variation of the quantization parameters along the color channel dimension including a syntax element that indicates whether, for a given group of the groups, the encoded data includes syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group; and
for the given group, the syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group; and
decoding the encoded data to reconstruct the frame, wherein the decoding includes:
determining the quantization parameters using the one or more syntax elements that control variation of the quantization parameters along the spatial dimension, the one or more syntax elements that control variation of the quantization parameters along the color channel dimension, and the syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group;
for blocks of at least some of the multiple areas of the frame, inverse quantizing transform coefficients using the determined quantization parameters; and
for the blocks of at least some of the multiple areas of the frame, performing inverse frequency transform operations on the inverse-quantized transform coefficients.
2. The computer system of claim 1 , wherein the multiple areas are macroblocks, each of the macroblocks having one or more blocks for the first color channel, one or more blocks for the second color channel, and one or more blocks for the third color channel.
3. The computer system of claim 2 , wherein each of the macroblocks represents a 16×16 unit of the frame for the first color channel.
4. The computer system of claim 1 , wherein the determined quantization parameters include different sets of quantization parameters for the first color channel, the second color channel, and the third color channel, respectively, for the given group.
5. The computer system of claim 1 , wherein the determined quantization parameters include different sets of quantization parameters for the first color channel, the second color channel, and the third color channel, respectively, for one of the multiple areas.
6. The computer system of claim 1 , wherein the determined quantization parameters include different sets of quantization parameters for at least some different groups among the groups of multiple areas of the frame.
7. The computer system of claim 1 , wherein the determined quantization parameters include different sets of quantization parameters for at least some different areas among the multiple areas of the frame.
8. One or more memory or storage devices having stored thereon encoded data in a bitstream, wherein the encoded data:
includes one or more syntax elements that control variation of quantization parameters along a spatial dimension within a frame of pixel values, the pixel values of the frame being organized spatially into multiple areas within the frame, the multiple areas being further organized spatially into groups, the one or more syntax elements that control variation of the quantization parameters along the spatial dimension including at least one syntax element that indicates whether variation of the quantization parameters is possible between the multiple areas within the respective groups;
includes one or more syntax elements that control variation of the quantization parameters along a color channel dimension, the pixel values of the frame also being organized by first, second, and third color channels, the one or more syntax elements that control variation of the quantization parameters along the color channel dimension including a syntax element that indicates whether, for a given group of the groups, the encoded data includes syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group;
includes, for the given group, the syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group; and
is produced by encoding, using a computer-implemented video encoder comprising one or more processing units, with operations that include:
for blocks of at least some of the multiple areas of the frame, performing frequency transform operations, thereby producing transform coefficients;
for the blocks of at least some of the multiple areas of the frame, quantizing the transform coefficients; and
setting the one or more syntax elements that control variation of quantization parameters along the spatial dimension within the frame, the one or more syntax elements that control variation of the quantization parameters along the color channel dimension, and the syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group.
9. The one or more memory or storage devices of claim 8 , wherein the multiple areas are macroblocks, each of the macroblocks having one or more blocks for the first color channel, one or more blocks for the second color channel, and one or more blocks for the third color channel.
10. The one or more memory or storage devices of claim 9 , wherein each of the macroblocks represents a 16×16 unit of the frame for the first color channel.
11. The one or more memory or storage devices of claim 8 , wherein the determined quantization parameters include different sets of quantization parameters for the first color channel, the second color channel, and the third color channel, respectively, for the given group.
12. The one or more memory or storage devices of claim 8 , wherein the determined quantization parameters include different sets of quantization parameters for the first color channel, the second color channel, and the third color channel, respectively, for one of the multiple areas.
13. The one or more memory or storage devices of claim 8 , wherein the determined quantization parameters include different sets of quantization parameters for at least some different groups among the groups of multiple areas of the frame.
14. The one or more memory or storage devices of claim 8 , wherein the determined quantization parameters include different sets of quantization parameters for at least some different areas among the multiple areas of the frame.
15. In a computer system, a method comprising:
encoding a frame of pixel values, the pixel values of the frame being organized spatially into multiple areas within the frame, the multiple areas being further organized spatially into groups, the pixel values of the frame also being organized by first, second, and third color channels, wherein the encoding includes:
for blocks of at least some of the multiple areas of the frame, performing frequency transform operations, thereby producing transform coefficients;
for the blocks of at least some of the multiple areas of the frame, quantizing the transform coefficients; and
setting one or more syntax elements that control variation of quantization parameters along a spatial dimension within the frame, one or more syntax elements that control variation of the quantization parameters along a color channel dimension, and, for a given group of the groups, syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group; and
outputting encoded data as part of a bitstream, wherein the encoded data includes:
the one or more syntax elements that control variation of the quantization parameters along the spatial dimension, including at least one syntax element that indicates whether variation of the quantization parameters is possible between the multiple areas within the respective groups;
the one or more syntax elements that control variation of the quantization parameters along the color channel dimension, including a syntax element that indicates whether, for the given group, the encoded data includes the syntax elements that indicate how to vary the quantization parameters for the second and third colors channels, respectively, for the given group; and
for the given group, the syntax elements that indicate how to vary the quantization parameters for the second and third color channels, respectively, for the given group.
16. The method of claim 15 , wherein the multiple areas are macroblocks, each of the macroblocks having one or more blocks for the first color channel, one or more blocks for the second color channel, and one or more blocks for the third color channel.
17. The method of claim 15 , wherein the determined quantization parameters include different sets of quantization parameters for the first color channel, the second color channel, and the third color channel, respectively, for the given group.
18. The method of claim 15 , wherein the determined quantization parameters include different sets of quantization parameters for the first color channel, the second color channel, and the third color channel, respectively, for one of the multiple areas.
19. The method of claim 15 , wherein the determined quantization parameters include different sets of quantization parameters for at least some different groups among the groups of multiple areas of the frame.
20. The method of claim 15 , wherein the determined quantization parameters include different sets of quantization parameters for at least some different areas among the multiple areas of the frame.