Model based bit rate control for a macroblock encoder

• US 20050254578A1
• Filed: 05/05/2005
• Published: 11/17/2005
• Est. Priority Date: 05/13/2004
• Status: Active Grant

• Current Assignee: Ittiam Systems Pte Ltd
• Original Assignee: Ittiam Systems Pte Ltd
• Inventors: Muthukrinan, Murali Babu, Raman, Arvind, Sethuraman, Sriram, Parameswaran, Sankaranarayanan, Tamia, Dileep Kumar, Rao, Bhavani Gopalakrishna
• Granted Patent: US 7,720,145 B2
• Time in Patent Office: Days
• US Class Current: 375/240.30
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/132 Sampling, masking or truncation of coding units, e.g. adaptive resampling, frame skipping, frame interpolation or high-frequency transform coefficient masking
• CPC Class Codes: H04N 19/149 by estimating the code amount by means of a model, e.g. mathematical model or statistical model
• CPC Class Codes: H04N 19/152 by measuring the fullness of the transmission buffer
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/587 involving temporal sub-sampling or interpolation, e.g. decimation or subsequent interpolation of pictures in a video sequence
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

A method for achieving bit rate controlled encoding, e.g., constant bit rate, uses a bit rate control model based on bits allocated per macroblock and based on error in estimating bits consumed in an encoding process in a macroblock. The method computes bits consumed per macroblock as a function of the formed bit rate control model, and allocates bits for the macroblock. To this end, a quadratic (second order) equation is used for the model, which equation might default into a first order equation requiring fewer computations, in the event that no second order solutions exist. In one form, the bit rate control model computes the bit rate to satisfy video buffer verifier (VBV) compliance in accordance with MPEG requirements. The bit rate control model makes allowances for the type of macroblock, i.e., inter, intra or bidirectional, and provides for any frame-skips.

Claims

1. A method for achieving bit rate controlled encoding, comprising:

forming a bit rate control model based on bits allocated per macroblock and based on error in estimating bits consumed in an encoding process in a macroblock;

computing bits consumed per macroblock as a function of the formed bit rate control model; and

allocating bits for the macroblock.

2. The method of claim 1, wherein the step of computing the bits consumed per macroblock comprises:

computing a quantizer value as a function of the formed bit rate control model; and

computing the bits consumed based on said computed quantizer value.

3. The method of claim 2, wherein the rate control model is comprises a quadratic equation.

4. The method of claim 3, wherein the step of computing a quantizer value comprises:

if the computed quantizer value is a real number, then computing the bits consumed using said computed quantizer value; and

if the computed quantizer value is not a real number, then degenerating said quadratic equation to a linear equation, and computing the bits consumed using the computed quantizer value.

5. The method of claim 1, wherein, in forming the bit rate control model, the bit rate is computed using the equation:

bit rate/macroblock=bits per frame/number of macroblocks

6. The method of claim 1, wherein, in forming the bit rate control model, the bit rate is computed as a function of video buffer verifier (VBV) compliance to cater MPEG standard requirements.

7. The method of claim 2, wherein the step of computing the quantizer value comprises:

if the computed quantizer value of the current macroblock is not equal to the computed quantizer value of the previous macroblock, then estimating the bits used for encoding a change in quantizer values;

if estimated bits used for encoding are less than a threshold value, then using the quantizer value of a previous macroblock.

8. The method of claim 1, wherein the step of forming the bit rate control model, comprises:

forming the rate control model based on a type of macroblock chosen from a group consisting of inter, intra, and bidirectional macroblocks.

9. The method of claim 1, wherein the step of allocating the bits for the macroblock is computed using the equation:

bits allocated per macroblock=(bits in reservoir−

actual bits consumed by the previous macroblock)/remaining number of macroblocks.

10. The method of claim 8, wherein forming the bit rate control model for an inter macroblock comprises:

dividing the allocated bits per macroblock into bits for motion vector data and coefficient data.

11. The method of claim 10, wherein the bits for the motion vector data are computed by a predetermined ratio of the allocated bits per macroblock.

12. The method of claim 9, wherein the rate control model is used for computing said quantizer value based on the allocated bits per coefficient data.

13. The method of claim 10, further comprising:

scaling motion vector bits using another ratio based on a number of frame-skips.

14. The method as in claim 1, including the step of fine tuning said bit rate control model to produce fewer quantizer variations in a frame.

15. The method as in claim 1, wherein the bit rate control model provides for frame skips and dynamically changes encoding parameters based on said frame skips.

16. The method as in claim 1, wherein the step of allotting bits uses least squares estimate.

17. The method as in claim 1, wherein said step of forming the bit rate control model comprises forming a quadratic rate control model wherein, for coding intra coded frames, signal energy is assumed to be a sum of all the pixel values in said current macroblock to be encoded.

18. The method as in claim 1, including the step of computing bits available per frame based on bits per second desired by a user, a ratio of I pictures to P pictures and a factor by which bit allocation for said I pictures is higher than for P pictures.

19. The method as in claim 1, where, in case of P pictures, the method includes the step of using frame level average to determine bits consumed by motion vectors, and using a predetermined factor for allocating bits for an initial P frame.

20. A method for achieving constant bit rate encoding for motion compression, comprising the steps of:

forming a bit rate control model which estimates an encoding parameter such that a current macroblock to be encoded takes a predetermined number of bits; and

,

allotting bits for the macroblock to be coded, using said bit rate control model.

21. The method as in claim 20, wherein said encoding parameter is a quantizer scale and includes an allowed bit rate.

22. The method as in claim 20, wherein the step of allotting bits uses least squares estimate.

23. The method as in claim 20, including the step of tracking bits having video content with smallest possible delay.

24. The method as in claim 20, wherein said step of forming the bit rate control model comprises forming a quadratic rate control model wherein, for coding intra coded frames, signal energy is assumed to be a sum of all the pixel values in said current macroblock to be encoded.

25. The method as in claim 20, including the step of considering a given bit rate, a frame rate and a number of macroblocks to be coded, for allocation of bits per macroblock.

26. The method as in claim 25, where said allocation of bits per macroblock is done for a group of pictures (GOP).

27. The method as in claim 20, including the step of computing bits available per frame based on bits per second desired by a user, and a ratio of I pictures to P pictures and a factor by which bit allocation for said I pictures is higher than for P pictures.

28. The method as in claim 20, where, in case of P pictures, the method includes the step of using frame level average to determine bits consumed by motion vectors, and using a predetermined factor for allocating bits for an initial P frame.

29. The method as in claim 28, including, when a frame drop is encountered, the step of adjusting a next picture bit allocation for motion vectors to have an increased value by a scale factor.

30. The method as in claim 29, wherein the scale factor is 1.5.

31. The method as in claim 26, including the step of changing a GOP structure dynamically to suit scene changes such that a scene change is coded as an Intra frame, and a new GOP begins from said Intra frame.

32. An article comprising a storage medium having instructions thereon which when executed by a computing platform result in execution of a method for achieving bit rate controlled encoding, comprising:

forming a bit rate control model based on bits allocated per macroblock and based on error in estimating bits consumed in an encoding process in a macroblock;

computing bits consumed per macroblock as a function of the formed bit rate control model; and

allocating bits for the macroblock.

33. An article as in claim 32, wherein the bit rate control model uses a second order equation.

34. An article comprising a storage medium having instructions thereon which when executed by a computing platform result in execution of a method for achieving constant bit rate encoding for motion compression, comprising the steps of:

forming a bit rate control model which estimates an encoding parameter such that a current macroblock to be encoded takes a predetermined number of bits; and

,

allotting bits for the macroblock to be encoded, using said bit rate control model.

35. The article as in claim 34, wherein the bit rate control model uses a second order equation.