Fractional quantization step sizes for high bit rates

• US 8,218,624 B2
• Filed: 07/17/2004
• Issued: 07/10/2012
• Est. Priority Date: 07/18/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Holcomb, Thomas W., Srinivasan, Sridhar, Hsu, Pohsiang, Lin, Chih-Lung
• Primary Examiner(s): Kelley, Christopher S
• Assistant Examiner(s): Anyikire, Chikaodili E
• Application Number:
  US10/893,167
• Publication Number:
  US 20050238096A1
• Time in Patent Office: 2,915 Days
• Field of Search: 375/240.03, 375/240.25
• US Class Current: 375/240.03
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/126 Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
• CPC Class Codes: H04N 19/136 Incoming video signal characteristics or properties
• CPC Class Codes: H04N 19/137 Motion inside a coding unit, e.g. average field, frame or block difference
• CPC Class Codes: H04N 19/17 the unit being an image region, e.g. an object
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/18 the unit being a set of transform coefficients
• CPC Class Codes: H04N 19/46 Embedding additional information in the video signal during the compression process H04N19/517, H04N19/68, H04N19/70 take precedence
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/70 characterised by syntax aspects related to video coding, e.g. related to compression standards

Abstract

At high bit rates, the reconstruction error of compressed video is generally proportional to the squared value of quantization step size, such that full quantization step size increments at high bit rates can lead to significant change in the reconstruction error and/or bit rate of the compressed video. A video codec uses fractional increments of quantization step size at high bit rates to permit a more continuous variation of quality and/or bit rate as the quantization scale changes. For high bit rate scenarios, the bit stream syntax includes an additional syntax element to specify fractional step increments (e.g., half step) of the normal quantizer scale step sizes.

Claims

1. A computer-implemented method of coding video using a video encoder, the method comprising:

with the video encoder, encoding video to produce encoded data, including;

selecting a scale for a quantizer from a set of values expressable via an encoding syntax, wherein step size increment between the expressable values varies within the set of expressable values depending at least in part on quantization step size, and wherein the selecting the scale for the quantizer includes;

when coding at a high quantization step size, selecting the scale for the quantizer from among the expressable values at a full step size increment; and

when coding at a low quantization step size, selecting the scale for the quantizer from among the expressable values at a step size increment that is a fraction of the full step size increment; and

processing transform coefficients for blocks based at least in part on the quantizer with the selected scale; and

outputting the encoded data in a bit stream, including;

signaling the selected scale for the quantizer using a first syntax element that represents the full step size increment;

comparing the selected scale to a comparison value to determine whether or not to signal a second syntax element that represents the fraction of the full step size increment, wherein the comparison value is 8; and

if the comparison indicates the second syntax element is to be signaled, signaling the second syntax element that represents the fraction of the full step size increment.

2. The method of claim 1 wherein the first syntax element representing the full step size increment is a quantizer scale index element.

3. The method of claim 1 wherein the fraction is one half.

4. A computer-implemented method of decoding video using a video decoder, the method comprising:

receiving encoded data in a bit stream; and

with the video decoder, decoding video using the encoded data, including;

decoding a first syntax element indicating a value of quantization index for a quantizer at a full step size increment;

comparing the value of the quantization index to a comparison value to determine whether or not to decode a second syntax element indicating a fractional step size increment, wherein the comparison value is 8;

for lower values but not higher values of the quantization index relative to the comparison value, decoding the second syntax element indicating the fractional step size increment, wherein syntax of the bit stream allows different step size increments for the quantizer for different ranges of values of the quantization index, such that the first and second syntax elements specify fractional step size increments of the quantizer for a range of the lower values of the quantization index, and such that the first syntax element specifies full step size increments of the quantizer for a range of the higher values of the quantization index; and

inverse quantizing quantized transform coefficients using the quantizer.

5. The method of claim 4 wherein the first and second syntax elements are signaled at frame level to specify the quantizer for an individual frame.

6. The method of claim 4 wherein high bit rate coding is characterized by the range of the lower values of the quantization index, and wherein low bit rate coding is characterized by the range of the higher values of the quantization index.

7. The method of claim 4 wherein the first syntax element is a fixed length code signaled at picture level.

8. The method of claim 4 wherein the second syntax element is a single bit signaled at picture level.

9. The method of claim 4 further comprising, for the quantizer, adding the fractional step size increment to scale of the quantizer at the full step size increment, wherein the value of quantization index indicates the scale of the quantizer at the full step size increment.

10. A computer-implemented method of decoding video using a video decoder, the method comprising:

receiving encoded data in a bit stream and, with the video decoder, decoding video using the encoded data, including;

receiving a first syntax element that indicates a value of a quantization index;

decoding the first syntax element to determine the value of the quantization index;

comparing the value of the quantization index to a comparison value, wherein the comparison value is 8;

responsive to the comparing;

receiving a second syntax element that indicates a fractional step size increment, wherein syntax of the bit stream allows different step size increments for different ranges of values of the quantization index, such that the second syntax element specifies the fractional step size increment for a first range of values of the quantization index but not for a second range of values of the quantization index;

decoding the second syntax element to determine the fractional step size increment;

calculating a quantization step size based on the value of the quantization index and the fractional step size increment from a set of values expressable for the quantization step size via the syntax of the bit stream; and

inverse quantizing quantized transform coefficients using the calculated quantization step size.

11. The method of claim 10 wherein the comparing comprises determining whether the value of the quantization index is less than or equal to the comparison value, wherein the first syntax element is a fixed length code signaled at picture level, and wherein the second syntax element is a single bit signaled at picture level.

12. The method of claim 10 wherein the first and second syntax elements are for a first video picture, the method further comprising, for a second video picture:

receiving a first syntax element for the second picture that indicates a value of a quantization index for the second picture;

decoding the first syntax element for the second picture to determine the value of the quantization index for the second picture;

comparing the value of the quantization index value for the second picture to the comparison value; and

responsive to the comparing, determining that the bit stream includes no second syntax element indicating a fractional step size increment for the second picture.

13. A computer-implemented method of encoding video using a video encoder, the method comprising:

with the video encoder, encoding video to produce encoded data, including;

determining a quantization step size;

quantizing transform coefficients based at least in part on the quantization step size;

using a first syntax element to represent a quantization index for a full-step increment of the quantization step size;

comparing the quantization index to a comparison value to determine whether or not to use a second syntax element that represents a fractional offset to the full-step increment of the quantization step size, wherein the comparison value is 8; and

conditionally using the second syntax element to represent the fractional offset to the full-step increment of the quantization step size, wherein bit stream syntax allows different step size increments for different ranges of values of the quantization index, such that the second syntax element specifies the fractional offset for a first range of values of the quantization index but not for a second range of values of the quantization index, wherein the first syntax element expresses a selection of the quantization step size from a set of values having the full-set increment for the second range, and wherein the first syntax element and the second syntax element in combination express a selection of the quantization step size from a set of values having a fractional increment relating to the fractional offset for the first range; and

outputting the encoded data in a bit stream, including signaling the first syntax element and conditionally signaling the second syntax element when the second syntax element is used.

14. The method of claim 13 wherein the first syntax element is a fixed length code signaled at picture level, and wherein the second syntax element is a single bit signaled at picture level.