Constant bit-rate control in a video coder by way of pre-analysis of a slice of the pictures

• US 6,215,820 B1
• Filed: 10/08/1999
• Issued: 04/10/2001
• Est. Priority Date: 10/12/1998
• Status: Expired due to Term

• Current Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Original Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Inventors: De Bei, Mattia, Mian, Gian Antonio, Bagni, Daniele, Sacchi, Maria Luisa
• Primary Examiner(s): Kelley, Chris
• Application Number:
  US09/415,692
• Time in Patent Office: 550 Days
• Field of Search: 348/419, 348/405, 375/240
• US Class Current: 375/240
• CPC Class Codes: H04N 19/115 Selection of the code volume for a coding unit prior to coding
• CPC Class Codes: H04N 19/126 Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
• CPC Class Codes: H04N 19/137 Motion inside a coding unit, e.g. average field, frame or block difference
• CPC Class Codes: H04N 19/14 Coding unit complexity, e.g. amount of activity or edge presence estimation H04N19/146 takes precedence
• 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 19/152 by measuring the fullness of the transmission buffer
• CPC Class Codes: H04N 19/172 the region being a picture, frame or field
• CPC Class Codes: H04N 19/174 the region being a slice, e.g. a line of blocks or a group of blocks
• CPC Class Codes: H04N 19/176 the region being a block, e.g. a macroblock
• CPC Class Codes: H04N 19/61 in combination with predictive coding

Abstract

An algorithm based on a pre-analysis implements an efficient constant bit-rate control with a reduced requirement on a buffer memory capacity. The pre-analysis is on at least one slice (GOS) of the current whole picture, and/or on a mix of information on the pre-analysis of a slice of the preceding picture and on the actual encoding data of the preceding whole picture. The pre-analysis may be carried out by precoding the GOS with a constant reference quantizer or by entropy computation. The local control of the bit-rate is implemented by an integrative-proportional controller.

Claims

1. A method for providing a constant bit-rate (CBR) at a certain value responsive to optimization criteria for transferring a data stream of encoded picture sequences, the method comprising the steps of:

a) estimating a number of bits for encoding a current picture and a successive picture belonging to a same group of pictures (GOP) of a sequence based on a target number of coding bits for each picture of the sequence as a function of encoding data of pictures preceding the current picture;

b) determining a reference value of a quantization parameter at a local level for an n-th macroblock of data that complies with a limit established by the estimated number of bits for the n-th macroblock;

c) calculating an effective quantization parameter for each macroblock as a function of a filling state of an output buffer of the data stream of coded data and of an activity parameter of the macroblock;

d) calculating a distribution profile of target bits over a picture using a pre-analysis of at least one line of a macroblock (GOS) of less than the whole current picture, distributing the allocated bits as a function of the local complexity of the single n-th macroblock and of the number of bits effectively used in the pre-analysis, effective encoding of the GOS of the preceding picture being used for distributing the allocated bits as a function of the local complexity of the n-th single macroblock and of the number of bits effectively used during the effective coding;

e) calculating a distribution profile of target bits over the GOS given by a ratio between the estimated number of bits required for the encoding obtained using a pre-analysis of the current GOS of less than the whole current picture;

f) intermediately computing a distribution profile of the target bit among the different lines of macroblocks based upon the ratio between entropv values derived from a pre-analysis of the current GOS less than the whole current picture multiplied by the target number of a bit usage profile; and

g) the pre-analysis of the current GOS being carried out using a calculation of the entropy on histograms of the discrete cosine transform (DCT) coefficients of the whole GOS.

2. A method according to claim 1, wherein step a) further comprises estimating the number of bits based on results of the pre-analysis of step f).

3. A method according to claim 1, wherein step e) further comprises calculating the distribution profile of target bits from the effective coding of the corresponding GOS of the preceding current picture multiplied by the target number of bits for the whole picture.

4. A method according to claim 1, wherein the effective encoding of the GOS of the preceding picture in step d) is used for distributing the allocated bits as a function of the local complexity of the n-th single macroblock and of the number of bits effectively used during the effective coding.

5. A method according to claim 1, wherein the steps are implemented in accordance with the MPEG-2 standard.

6. A method according to claim 1, wherein the steps are implemented in accordance with the MPEG-1 standard.

7. A method for providing a constant bit-rate (CBR) at a certain value responsive to optimization criteria for transferring a data stream of encoded picture sequences, the method comprising the steps of:

a) estimating a number of bits necessary for encoding a current picture and a successive picture belonging to a same group of pictures (GOP) of a sequence based on a target number of coding bits for each picture of the sequence as a function of encoding data of pictures preceding the current picture;

b) determining a reference value of a quantization parameter at a local level for an n-th macroblock of data that complies with a limit established by the estimated number of bits for the n-th macroblock;

c) calculating an effective quantization parameter for each macroblock as a function of a filling state of an output buffer of the data stream of coded data and of an activity parameter of the macroblock;

d) calculating a distribution profile of target bits over a picture using a pre-analysis of at least one line of a macroblock (GOS) of less than the whole current picture, distributing the allocated bits as a function of the local complexity of the single n-th macroblock and of the number of bits effectively used in the pre-analysis;

e) calculating a distribution profile of target bits over the GOS given by a ratio between the estimated number of bits required for the encoding obtained using a pre-analysis of the current GOS of less than the whole current picture, and calculating the distribution profile of target bits from the effective coding of the corresponding GOS of the preceding picture multiplied by the target number of bits for the whole current picture;

f) intermediately computing a distribution profile of the target bit among the different lines of macroblocks based upon the ratio between entropy values derived from a pre-analysis of the current GOS of less than the whole current picture is multiplied by the target number of a bit usage profile; and

g) the pre-analysis of the current GOS being carried out using a calculation of the entropy on histograms of the DCT coefficients of the whole GOS.

8. A method according to claim 7, wherein step a) further comprises estimating the number of bits based on results of the pre-analysis of step g).

9. A method according to claim 7, wherein the effective encoding of the GOS of the preceding picture in step d) is used for distributing the allocated bits as a function of the local complexity of the n-th single macroblock and of the number of bits effectively used during the effective coding.

10. A method according to claim 7, wherein the steps are implemented in accordance with the MPEG-2 standard.

11. A method according to claim 7, wherein the steps are implemented in accordance with the MPEG-1 standard.

12. A video encoder comprising:

a controller for controlling a bit-rate of a bitstream of encoded video sequences to be transmitted over a transmission channel at a constant transmission rate, said controller comprising

an output buffer, and

an integrative-proportional controller connected to said output buffer and having an input receiving contents of said output buffer, and having an output providing a compressed video sequence of moving pictures to generate the bit-stream to be transmitted through the transmission channel, each picture comprising a plurality of macroblocks of pixels compressed with transform coding and/or temporal prediction and/or bi-dimensional motion compensated interpolation techniques to produce respectively I and/or P and/or B frames;

said controller intermediately computing a distribution profile of target bits among different lines of macroblocks, named GOS, equal to a ratio between entropy values derived from a pre-analysis of the GOS of the current picture, and multiplying the ratio by a target number of a bit usage profile, the pre-analysis being carried out using a calculation of the entropy on histograms of transform coding coefficients of the whole GOS, and calculating a target number of a bit usage profile by the effective encoding of the corresponding GOS of the previous picture multiplied by a target number of bits for the whole current picture.

13. A video encoder according to claim 12, wherein said controller calculates the target number of the bit usage profile by performing a pre-analysis of the corresponding macroblock of the previous picture.

14. A video encoder according to claim 13, wherein the intermediately computed distribution profile of target bits among different lines of macroblocks is equal to a ratio between bit values derived from the pre-analysis of the current GOS of the current picture, and multiplying the ratio by the target number of the bit usage profile.

15. A video encoder according to claim 12, wherein intermediately computing the distribution profile of target bits among different lines of macroblocks comprises determining an average between the entropy value ratio derived from the pre-analysis of the current GOS of the current picture, and the effective bits computed during coding of the previous picture multiplied by the target number of the bit usage profile.

16. A video encoder according to claim 12, further comprising a memory for storing the pre-analysis data of GOS of the current picture and the data of an effective coding.

17. A video encoder according to claim 12, wherein said controller is compatible with the MPEG-2 standard.

18. A video encoder according to claim 12, wherein said controller is compatible with the MPEG-1 standard.

19. A method for controlling a bit-rate of a bitstream of encoded video sequences to be transmitted over a channel at a constant transmission rate comprising:

compressing a video sequence of moving pictures to generate the bitstream, each picture comprising a plurality of macroblocks of pixels compressed with transform coding and/or temporal prediction and/or bi-dimensional motion compensated interpolation techniques to produce respectively I and/or P and/or B frames;

intermediately computing a distribution profile of target bits among different lines of macroblocks, named GOS, equal to a ratio between entropy values derived from a pre-analysis of the GOS of the current picture, and multiplying the ratio by a target number of a bit usage profile;

the pre-analysis being carried out using a calculation of the entropy on histograms of transform coding coefficients of the whole GOS; and

calculating a target number of a bit usage profile by the effective encoding of the corresponding GOS of the previous picture multiplied by a target number of bits for the whole current picture.

20. A method according to claim 19, wherein calculating the target number of the bit usage profile comprises performing a pre-analysis of the corresponding GOS of the previous picture.

21. A method according to claim 19, wherein intermediately computing the distribution profile of target bits among different lines of macroblocks comprises determining an average between the entropy value ratio derived from the pre-analysis of the current GOS of the current picture, and the effective bits computed during coding of the previous picture multiplied by the target number of the bit usage profile.

22. A method according to claim 19, wherein the intermediately computed distribution profile of target bits among different lines of macroblocks is equal to a ratio between bit values derived from the pre-analysis of the current GOS of the current picture, and multiplying the ratio by the target number of the bit usage profile.

23. A method according to claim 22, wherein calculating the target number of a bit usage profile comprises computing a pre-analysis of the corresponding GOS of the previous picture.