Adaptive B-picture quantization control

• US 7,974,340 B2
• Filed: 04/07/2006
• Issued: 07/05/2011
• Est. Priority Date: 04/07/2006
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Lin, Chih-Lung, Xia, Minghui
• Primary Examiner(s): Liu; Shuwang
• Assistant Examiner(s): Neff; Michael R
• Application Number:
  US11/400,744
• Publication Number:
  US 20070237222A1
• Time in Patent Office: 1,915 Days
• Field of Search: 375/240.03, 375/240.26
• US Class Current: 375/240.03
• CPC Class Codes: H04N 19/126 Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

The present application is directed to techniques and tools for adapting the quantizer as well as the quantization step size used during video encoding. For example, a video encoder uses a non-uniform quantizer (having a relatively large dead zone ratio) for B-picture types, but uses a uniform quantizer for other types of pictures. In many encoding scenarios, this helps the encoder control bit rate for the B-picture types instead of resorting to higher quantization step sizes, which improves overall quality for the B-picture types.

Claims

1. In a video encoder, a method comprising:

selecting a picture type for a current picture;

using the selected picture type when selecting between multiple available quantizers for the current picture, each of the multiple available quantizers having a different dead zone ratio, the selected quantizer for the current picture having a larger dead zone ratio, when the selected picture type is a B-picture type, than one or more other quantizers among the multiple available quantizers, wherein a larger dead zone ratio results in larger dead zone regions that each have a size that is a function of a quantization step size (OP) and wherein the size is approximately 0.8* QP on each side of zero for each OP of the selected quantizer;

applying the selected quantizer to the current picture; and

outputting encoded data for the current picture.

2. The method of claim 1 wherein the selected quantizer is a non-uniform quantizer.

3. The method of claim 1 wherein the larger dead zone ratio of the selected quantizer for the current picture applies for AC coefficients but not for DC coefficients.

4. The method of claim 1, wherein the larger dead zone regions are for AC coefficients and the size is approximately 0.8*QP on each side of zero for each QP of the selected quantizer, and wherein dead zone regions for DC coefficients have a size that is equal to 0.5*QP on each side of zero for each QP of the quantizer.

5. One or more computer-readable media having stored thereon computer executable instructions to cause a computer to perform the method of claim 1.

6. In a video encoder, a method comprising:

getting a current picture of a video sequence;

if the current picture is a P-picture or an I-picture;

selecting a uniform quantizer with uniform dead zone for AC coefficients for the current picture; and

applying the selected uniform quantizer to the current picture;

if the current picture is a B-picture;

selecting a non-uniform quantizer with a non-uniform dead zone for AC coefficients for the current picture, wherein the non-uniform dead zone is larger than the uniform dead zone of the uniform quantizers; and

applying the selected non-uniform quantizer to the current picture; and

outputting encoded data for the current picture.

7. The method of claim 6 wherein the non-uniform dead zone for AC coefficients for the current picture is greater than a dead zone for DC coefficients for the current picture.

8. The method of claim 6 wherein dead zone regions each have a size that is a function of the quantization step size (QP), and the non-uniform dead zone is approximately 0.8*QP on each side of zero for each QP of the non-uniform quantizer.

9. One or more computer-readable media having stored thereon computer executable instructions to cause a computer to perform the method of claim 6.

10. An encoder comprising:

a frequency transformer for frequency transforming plural blocks of spatial domain information into plural blocks of transform coefficients;

a quantizer for quantizing the transform coefficients;

an entropy encoder for entropy encoding the quantized transform coefficients;

an inverse quantizer for inverse quantizing the quantized transform coefficients; and

a controller for selecting quantizers and quantization step sizes for the quantizing by choosing between plural available quantizers, each of the plural available quantizers having a different dead zone ratio, wherein the controller implements;

a first path for increasing dead zone ratio while decreasing quantization step size for a given level of rate-distortion performance;

a second path for decreasing dead zone ratio while increasing quantization step size for the given level of rate-distortion performance; and

wherein the controller follows the first path if a current picture is a B-picture type and otherwise follows the second path.

11. The encoder of claim 10 wherein the selecting comprises using picture type when choosing between the plural available quantizers and when selecting between different quantization step sizes.

12. The encoder of claim 10 further comprising:

a motion estimator for estimating motion in predicted pictures from reference pictures;

a motion compensator for applying the estimated motion; and

one or more reference picture buffers for storing the reference pictures.

13. The encoder of claim 10 wherein the plural available quantizers include a uniform quantizer and a non-uniform quantizer.

14. The encoder of claim 13 wherein the non-uniform quantizer has a larger dead zone ratio than the uniform quantizer.

15. The encoder of claim 13 wherein the non-uniform quantizer has a non-uniform dead zone for quantizing AC coefficients and a uniform dead zone for quantizing DC coefficients.

16. The encoder of claim 10 wherein selecting quantizers comprises, for a current picture, selecting a first quantizer from among the plural available quantizers if the current picture has a B-picture type and otherwise selecting a second quantizer from among the plural available quantizers.