Adaptive selection of quantization scales for video encoding

• US 6,765,962 B1
• Filed: 11/21/2000
• Issued: 07/20/2004
• Est. Priority Date: 12/02/1999
• Status: Expired due to Fees

• Current Assignee: Sarnoff Corporation (SRI International, Inc.)
• Original Assignee: Sarnoff Corporation (SRI International, Inc.)
• Inventors: Lee, Jungwoo, Binenbaum, Nurit
• Primary Examiner(s): Vo, Tung T.
• Application Number:
  US09/717,420
• Time in Patent Office: 1,337 Days
• Field of Search: G06/K.936, H04/N.724, H04/B.166, 375/240, 375/240.01, 375/240.03, 375/240.04, 375/240.07, 375/240.16, 348/384.1, 348/390.1, 348/400.1, 348/403.1, 348/405.1, 348/419.1, 382/232, 382/234, 382/236, 382/238, 382/250, 382/251
• US Class Current: 375/240.03
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/196 being specially adapted for the computation of encoding parameters, e.g. by averaging previously computed encoding parameters processing of motion vectors H04N19/513
• CPC Class Codes: H04N 19/197 including determination of the initial value of an encoding parameter H04N19/56 takes precedence
• CPC Class Codes: H04N 19/503 involving temporal prediction adaptive coding with adaptive selection between spatial and temporal predictive coding H04N19/107; adaptive coding with adaptive selection among a plurality of temporal predictive coding modes H04N19/109
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

The quantization scale selected for encoding the current frame of a video sequence is selected based on a metric generated based on a set of image data in the video sequence. For example, in MPEG encoding, the linear quantization scale is selected for use in encoding the current frame if the average quantization level used to encode the previously encoded frame is between specified high and low thresholds. Otherwise, the non-linear quantization scale is selected. As a result, medium-difficulty sequences will tend to be encoded using the linear quantization scale, while low- and high-difficulty sequences will tend to be encoded using the non-linear quantization scale. For most normal video sequences, this will result in fewer incidents of panic mode video compression processing and improved picture quality.

Claims

1. A method for encoding frames of a video sequence, comprising the steps of:

(a) generating a metric characterizing quantization levels corresponding to a set of image data in the video sequence;

(b) comparing the metric to one or more specified thresholds to select a quantization scale for a current frame in the video sequence; and

(c) encoding the current frame using the selected quantization scale, wherein;

the quantization scale is one of a linear quantization scale and a non-linear quantization scale;

the linear quantization scale represents a set of quantization levels forming a linear progression; and

the non-linear quantization scale represents a set of quantization levels forming a non-linear progression, wherein step (b) comprises the steps of;

(1) comparing the metric to a low threshold and to a high threshold; and

(2) selecting a first quantization scale, if the metric is between the low and high thresholds;

otherwise, selecting a second quantization scale.

2. The invention of claim 1, wherein step (a) comprises the step of generating the metric based on the quantization levels used to encode a previously encoded frame in the video sequence.

3. The invention of claim 2, wherein the metric is an average quantization level for the previously encoded frame.

4. The invention of claim 2, wherein the one or more specified thresholds are independent of the quantization scale used to encode the previously encoded frame.

5. The invention of claim 2, wherein the one or more specified thresholds are dependent on the quantization scale used to encode the previously encoded frame.

6. The invention of claim 5, wherein the dependence of the one or more specified thresholds on the quantization scale used to encode the previously encoded frame achieves a degree of hysteresis for the method.

7. The invention of claim 2, wherein the one or more specified thresholds are independent of whether the previously encoded frame is an I, P, or B frame.

8. The invention of claim 2, wherein the one or more specified thresholds are dependent on whether the previously encoded frame is an I, P, or B frame.

9. The invention of claim 1, wherein step (a) comprises the step of generating the metric based on quantization levels selected during a first pass of processing for the current frame.

10. The invention of claim 8, wherein the selected quantization scale is used during a second pass of processing for the current frame.

11. The invention of claim 10, wherein the quantization scale used for the first pass is the quantization scale used to encode a previously encoded frame in the video sequence.

12. The invention of claim 1, wherein the first quantization scale has a dynamic range smaller than the second quantization scale.

13. The invention of claim 12, wherein the first quantization scale is an MPEG linear quantization scale and the second quantization scale is an MPEG non-linear quantization scale.

14. The invention of claim 1, wherein step (c) comprises the step of selecting one or more quantization levels in the selected quantization scale for quantizing DCT coefficients for the current frame.

15. The invention of claim 1, wherein the one or more specified thresholds are independent of whether the current frame is an I, P, or B frame.

16. The invention of claim 1, wherein the one or more specified thresholds are dependent on whether the current frame is an I, P, or B frame.

17. The invention of claim 1, wherein:

step (a) comprises the step of generating the metric based on the quantization levels used to encode a previously encoded frame in the video sequence, wherein the metric is an average quantization level for the previously encoded frame;

the first quantization scale is an MPEG linear quantization scale and the second quantization scale is an MPEG non-linear quantization scale; and

step (b) comprises the step of selecting one or more quantization levels in the selected quantization scale for quantizing DCT coefficients for the current frame.

18. The invention of claim 1, wherein the metric is generated from the quantization levels corresponding to the set of image data in the video sequence.

19. The invention of claim 1, wherein:

the linear quantization scale comprises 31 quantization levels consisting of 2 to 62 in increments of 2; and

the non-linear quantization scale comprises 31 quantization levels consisting of 1 to 8 in increments of 1, 8 to 24 in increments of 2, 24 to 56 in increments of 4, and 56 to 112 in increments of 8.

20. An apparatus for encoding frames of a video sequence, comprising:

(a) means for generating a metric characterizing quantization levels corresponding to a set of image data in the video sequence;

(b) means for comparing the metric to one or more specified thresholds to select a quantization scale for a current frame in the video sequence; and

(c) means for encoding the current frame using the selected quantization scale, wherein;

the quantization scale is one of a linear quantization scale and a non-linear quantization scale;

the linear quantization scale represents a set of quantization levels forming a linear progression; and

the non-linear quantization scale represents a set of quantization levels forming a non-linear progression, wherein means (b) comprises;

(1) means for comparing the metric to a low threshold and to a high threshold; and

(2) means for selecting a first quantization scale, if the metric is between the low and high thresholds;

otherwise, selecting a second quantization scale.

21. A machine-readable medium, having encoded thereon program code, wherein, when the program code is executed by a machine, the machine implements a method for encoding frames of a video sequence, comprising the steps of:

(a) generating a metric characterizing quantization levels corresponding to a set of image data in the video sequence;

(b) comparing the metric to one or more specified thresholds to select a quantization scale for a current frame in the video sequence; and

(c) encoding the current frame using the selected quantization scale, wherein;

the quantization scale is one of a linear quantization scale and a non-linear quantization scale;

the linear quantization scale represents a set of quantization levels forming a linear progression; and

the non-linear quantization scale represents a set of quantization levels forming a non-linear progression, wherein step (b) comprises the steps of;

(1) comparing the metric to a low threshold and to a high threshold; and

(2) selecting a first quantization scale, if the metric is between the low and high thresholds;

otherwise, selecting a second quantization scale.

22. A method for encoding frames of a video sequence, comprising the steps of:

(a) generating a metric characterizing quantization levels corresponding to a set of image data in the video sequence;

(b) comparing the metric to one or more specified thresholds to select a quantization scale for a current frame in the video sequence; and

(c) encoding the current frame using the selected quantization scale, wherein step (b) comprises the steps of;

(1) comparing the metric to a low threshold and to a high threshold; and

(2) selecting a first quantization scale, if the metric is between the low and high thresholds;

otherwise, selecting a second quantization scale.

23. The invention of claim 22, wherein the first quantization scale has a dynamic range smaller than the second quantization scale.

24. The invention of claim 23, wherein the first quantization scale is an MPEG linear quantization scale and the second quantization scale is an MPEG non-linear quantization scale.