Distortion quantizer model for video encoding

• US 6,687,294 B2
• Filed: 04/27/2001
• Issued: 02/03/2004
• Est. Priority Date: 04/27/2001
• Status: Expired due to Fees

• Current Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Original Assignee: Koninklijke Philips Electronics N.V. (Koninklijke Philips N.V.)
• Inventors: Challapali, Kiran, Yan, Yong
• Primary Examiner(s): LEE, RICHARD J
• Application Number:
  US09/844,457
• Publication Number:
  US 20020159521A1
• Time in Patent Office: 1,012 Days
• Field of Search: H04/N.712, 329/307, 341/51, 375/240.01, 375/240-2-2, 375/240-2-2, 704/230
• 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 19/147 according to rate distortion criteria rate-distortion as a criterion for motion estimation H04N19/567
• CPC Class Codes: H04N 19/154 Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion use of rate-distortion criteria H04N19/147
• CPC Class Codes: H04N 19/60 using transform coding

Abstract

A video encoder that allows for the selection of a distortion level, comprising: a selection system for selecting a target distortion level; and a system for determining a quantization parameter q that will ensure compliance with the selected target distortion level, wherein the system includes an algorithm for calculating distortion that utilizes a Gaussian distribution having a variance that is a function of the quantization parameter q. The algorithm is based on a distortion model D(q)=N(a₁q²+a₂q+a₃, b₁q²+b₂q+b₃), wherein N is a Gaussian distribution and a₁, a₂, a₃, b₁, b₂ and b₃ are distortion model parameters.

Claims

1. A method for determining a quantization parameter q that meets a predetermined quality level for video encoding, comprising the steps of:

providing a distortion model D(q)=N(a₁q²+a₂q+a₃, b₁q²+b₂q+b₃), wherein N is a Gaussian distribution and a₁, a₂, a₃, b₁, b₂ and b₃ are distortion model parameters;

selecting a target distortion level;

calculating the quantization parameter q such that an upper bound of the distortion model D(q) is less than or equal to the target distortion level; and

using the quantization parameter q for said video encoding.

2. The method of claim 1, comprising the further step of determining a bit rate based on the calculated quantization parameter q.

3. The method of claim 1, wherein the quantization parameter q is selected from a set of sequential values 1, 2, 3 . . . Q, and wherein Q is a whole number.

4. The method of claim 3, wherein the step of calculating the quantization parameter q includes the steps of iteratively testing each of the set of sequential values from Q to 1 until the upper bound of D(g) is less than or equal to the selected target distortion level.

5. The method of claim 1, wherein the distortion model D(q) is represented as a root mean square RMS(q).

6. The method of claim 1, wherein the distortion model parameters are estimated using a maximum likelihood estimator statistical method.

7. The method of claim 1, wherein the upper bound of the distortion model D(q) is calculated as:

D(q)=a₁q²+a₂q=a₃+λ

*sqrt(b₁q²+b₂q+b₃), and wherein λ

is a confidence level constant.

8. An encoding system having quality level selection capabilities, comprising:

a system for selecting a target distortion level;

a distortion model, wherein the distortion model determines a distortion level as a function of a quantization parameter, and wherein the distortion model includes a Gaussian distribution having a variance that is a function of the quantization parameter; and

a system for calculating the quantization parameter such that the distortion level does not exceed the target distortion level.

9. The encoding system of claim 8, further comprising a system for calculating a bit rate as a function of the calculated quantization parameter.

10. The encoding system of claim 8, wherein the distortion model is defined as:

11. The encoding system of claim 10, wherein an upper bound of the distortion level is defined by the distortion model as:

12. The encoding system of claim 10, further comprising a system for estimating the distortion model parameters.

13. The encoding system of claim 8, wherein the system for calculating the quantization parameter utilizes an algorithm that iteratively tests whole number quantization parameter values.

14. A video encoder that allows for the selection of a distortion level, comprising:

a selection system for selecting a target distortion level; and

a system for determining a quantization parameter that will ensure compliance with the selected target distortion level, wherein the system for determining includes an algorithm for calculating distortion that utilizes a Gaussian distribution having a variance that is a function of the quantization parameter.

15. The video encoder of claim 14, further comprising a system for selecting a bit rate as a function of the quatization parameter.

16. The video encoder of claim 14, wherein an upper bound of distortion D_upper for a given quantization level q is calculated as follows:

17. The video encoder of claim 14, wherein distortion D for a given quantization level q is calculated as follows:

18. A program product, stored on a recordable medium, which when executed allows for the selection of a distortion level in an encoding operation, comprising:

a selection system for determining for selecting a target distortion level; and

a code for determining a quantization parameter that will ensure compliance with the selected target distortion level, wherein the code for determining includes an algorithm for calculating distortion that utilizes a Gaussian distribution having a variance that is a function of the quantization parameter.