Quantization method and system for video MPEG applications and computer program product therefor

• US 7,301,999 B2
• Filed: 01/14/2004
• Issued: 11/27/2007
• Est. Priority Date: 02/05/2003
• Status: Active Grant

• Current Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Original Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Inventors: Pau, Danilo Pietro, Filippini, Gianluca, Piccinelli, Emiliano Mario
• Primary Examiner(s): DHARIA, PRABODH M
• Application Number:
  US10/757,961
• Publication Number:
  US 20040170395A1
• Time in Patent Office: 1,413 Days
• Field of Search: 345/660, 345/447, 345/505, 345/629, 345/558, 345/520, 345/547, 345/531, 345/535, 386/111, 386/112, 386/109, 386/129, 386/46, 375/240.12, 375/240.02, 375/240.03, 375/240.16, 348/446, 382/100
• US Class Current: 375/240.03
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/126 Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
• CPC Class Codes: H04N 19/40 using video transcoding, i.e. partial or full decoding of a coded input stream followed by re-encoding of the decoded output stream
• CPC Class Codes: H04N 19/90 using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals

Abstract

Digital signals are converted between a first and second format by a conversion process including generating coefficients representing the digital signals. The coefficients may be discrete cosine transform coefficient generated during encoding/transcoding of MPEG signals. The coefficients are subject to quantization by generating a dither signal that is added to the coefficients before quantization to generate a quantized signal. Preferably, each coefficient is first subject to a first quantization in the absence of any dither signal added to generate an undithered quantized coefficient. If the undithered quantized signal is equal to zero the undithered quantized coefficient is taken as the output quantized signal. If the undithered quantized coefficient is different from zero, the dither signal is added and the dithered coefficient thus obtained is subject to quantization to generate the output quantized signal.

Claims

1. A method of converting digital signals between a first and a second format, the method comprising:

generating coefficients representative of said digital signals;

subjecting said coefficients to a first quantization to generate an undithered quantized coefficient;

checking if said undithered quantized coefficient is equal to zero, such that

when said undithered quantized coefficient is equal to zero, taking said undithered quantization coefficient as a quantized signal, and

when said undithered quantized coefficient is different from zero, generating a dither signal, and adding said dither signal to said coefficients before said first quantization to generate a dithered coefficient and subjecting said dithered coefficient to a quantization step to generate a dithered quantized signal.

2. The method of claim 1, wherein subjecting said coefficients to quantization comprises subjecting said coefficients to a uniform quantization.

3. The method of claim 1, further comprising:

subjecting said quantized signal to inverse quantization; and

subtracting said dither signal from said signal subjected to inverse quantization.

4. The method of claim 1, wherein the spectrum of said dither signal is gaussian, uniform, sinusoidal or triangular.

5. The method of claim 4, wherein said dither signal is generated as a pseudo-random variable having a uniform distribution by subsequently modifying said distribution to at least one distribution of said group.

6. The method claim 1 wherein said dither signal is generated from a plurality of independent pseudo-random variables.

7. The method of claim 1 further comprising subjecting said digital signals to a discrete cosine transform to generate said coefficients to be quantized as discrete cosine transform coefficients.

8. The method of claim 1, wherein said quantization comprises a part of a transcoding process between an input stream of digital signals at a first bitrate and an output stream of digital signals at a second bitrate, said second bitrate of said output stream of digital signals being selectively controlled.

9. The method of claim 8, further comprising conducting a preanalysis process on said input stream including:

quantizing said signals with a given quantization step; and

evaluating the number of bits spent for coding said coefficients, and in that said bitrate of said output data stream is controlled as a function of said preanalysis.

10. The method of claim 9, further comprising controlling said data stream with a proportional-integrative control.

11. The method of claim 9, wherein said input stream comprises a stream of digital video signals including pictures arranged in groups of pictures, and wherein said bitrate control assigns a value of target bits for each single picture of a group of pictures.

12. The method of claim 1, wherein said quantization comprises part of a transcoding process between an input stream of digital signals at a first bitrate and an output bitrate at a second bitrate, said transcoding process including subjecting at least part of said input digital signals to low pass filtering step followed by downsampling.

13. The method of claim 12, wherein said low pass filtering is performed before conducting a preanalysis process.

14. The method of claim 12, further comprising executing a decimation phase.

15. The method of claim 1, wherein said digital signals comprise, in at least one of said first and second formats, MPEG encoded signals.

16. A system for converting digital signals between a first and second format, the system being configured for generating coefficients representative of said digital signals comprising:

a first quantizer for subjecting each said coefficient to a first quantization step to generate an undithered quantized coefficient;

a control module for checking if said undithered quantized coefficient is equal to zero;

an output element for taking said undithered quantization coefficient as a quantized signal when said undithered quantized coefficient is equal to zero;

a source of a dither signal;

an adder for adding said dither signal to said coefficient when said undithered quantized coefficient is different from zero; and

a second quantizer for subjecting said dithered coefficient to a quantization step to generate said quantized siqnal for feeding to said output element.

17. The system of claim 16 wherein said quantizer comprises a uniform quantizer.

18. The system of claim 16 further comprising:

an inverse quantizer for subjecting said quantized signal to inverse quantization; and

a subtractor for subtracting said dither signal from said signal subjected to inverse quantization.

19. The system of claim 16, wherein said source of said dither signal comprises a gaussian, uniform, sinusoidal or triangular signal source.

20. The system of claim 19, wherein said source comprises a source of a pseudo-random variable having a uniform distribution.

21. The system of any of claims 16, wherein said source of dither signal includes a plurality of sources of independent pseudo-random variables.

22. The system of claim 16 further comprising a discrete cosine transform module for subjecting said digital signals to a discrete cosine transform to generate said coefficients to be quantized as discrete cosine transform coefficients.

23. The system of claims 16, further comprising means for transcoding an input stream of said digital signals at a first bitrate into an output stream of digital signals at a second bitrate, including a bitrate control block for selectively controlling said second bitrate of said output stream of digital signals.

24. The system of claim 23, further comprising a preanalysis chain for subjecting said input stream to a preanalysis process, said chain including:

a quantizer for quantizing said signals with a given quantization step; and

a bit usage profile module for evaluating the number of bits spent for coding said coefficients,

wherein said bitrate control block is configured for controlling the bitrate of said output data stream as a function of said preanalysis.

25. The system of claim 24, wherein said bitrate control block comprises a proportional-integrative controller.

26. The system of claim 24, for use in connection with an input stream of digital video signals including pictures arranged in groups of pictures, wherein said bitrate control block is configured for assigning said value of target bits for each single picture of a group of pictures.

27. The system of claim 16, wherein said quantizer for transcoding an input stream of digital signals at a first bitrate into an output bitrate at a second bitrate, including a low pass filter and a downsampling module for subjecting at least part of said input digital signals to lowpass filtering and downsampling.

28. The method of claim 27, wherein said low pass filter is arranged upstream of a preanalysis chain.

29. The system of claim 27, further comprising a decimation module.

30. A computer program product directly loadable in the internal memory of a digital computer and including software code portions for performing, when the product is run on a computer, a method of converting digital signals between a first and a second format, the method comprising:

generating coefficients representative of said digital signals;

subjecting said coefficients to quantization;

generating a dither signal; and

checking if said quantized coefficient is equal to zero, such that when said quantized coefficient is equal to zero, taking said quantization coefficient as a quantized signal, and

when said quantized coefficient is different from zero, adding said dither signal to said coefficients before said quantization to generate a dithered coefficient and subjecting said dithered coefficient to quantization to generate a dithered quantized signal.

31. A method of converting digital signals between a first and second format, the method comprising:

subjecting said digital signals to a discrete cosine transform to generate coefficients representative of said digital signals;

generating a dither signal;

adding said dither signal to said coefficients and subjecting said coefficients to which said dither signal has been added to quantization to generate a quantized signal;

subjecting such quantized signal to inverse quantization; and

leaving said signal subjected to inverse quantization exempt from subtraction therefrom of dither signal, wherein said signal subjected to inverse quantization contains artifacts due to the non-linear characteristic of said quantization.

32. The method of claim 31, wherein said quantization step is a uniform quantization step.

33. The method of claim 31, wherein the spectrum of said dither signal is selected from the group consisting of gaussian, uniform, sinusoidal and triangular.

34. The method of claim 31, wherein said dither signal is generated as a pseudo-random variable having a uniform distribution by subsequently modifying said distribution to at least one distribution of said group.

35. The method of claim 31, wherein said dither signal is generated from a plurality of independent pseudo-random variables.

36. The method of claim 31, wherein said quantization is a part of a transcoding process between an input stream of digital signals at a first bitrate and an output stream of digital signals at a second bitrate, said second bitrate of said output stream of digital signals being selectively controlled.

37. The method of claim 31, wherein said input stream is subject to a preanalysis process including:

quantizing said signals with a given quantization step; and

evaluating the number of bits spent for coding said coefficients, and in that said bitrate of said output data stream is controlled as a function of said preanalysis.

38. The method of claim 36, wherein said control is of a proportional-integrative type.

39. The method of claim 31, wherein said input stream is stream of digital video signals including pictures arranged in groups of pictures, and in that said bitrate control assign value of target bits for each single picture of a group of pictures.