Real-time evaluation of compressed picture quality within a digital video encoder

• US 6,252,905 B1
• Filed: 02/05/1998
• Issued: 06/26/2001
• Est. Priority Date: 02/05/1998
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Stein, Charles J., Pokrinchak, Stephen P.
• Primary Examiner(s): Britton, Howard
• Assistant Examiner(s): VO, TUNG T
• Application Number:
  US09/020,904
• Time in Patent Office: 1,237 Days
• Field of Search: H04/N.736, H04/B.166, 348/400.1, 348/404.1, 348/405.1, 348/411.1, 348/412.1, 348/416.1, 375/240, 375/240.14, 382/240.03, 382/232, 382/234
• US Class Current: 375/240.14
• CPC Class Codes: G06T 9/00 Image coding bandwidth or redundancy reduction for static pictures H04N1/41; coding or decoding of static colour picture signals H04N1/64; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00

Abstract

Method, system and computer program product are provided for real time evaluation of compressed picture quality, in hardware, software or a combination thereof, during encoding of a sequence of video data. A received video signal is encoded by quantizing the signal after frequency transformation to produce a quantized video signal which is automatically evaluated in real time for data loss. Loss statistics, representative of picture quality of the quantized video signal, are generated for comparison with, for example, a predefined threshold as a dynamic control over outputting of the quantized video signal to the encode pipeline. As an enhancement, multiple quantizers are employed in parallel, each using a different quantization scale to produce multiple different quantization signals. The multiple quantization signals are evaluated in real time for picture quality and an optimum signal is selected for forwarding within the encode pipeline.

Claims

1. A method for encoding a sequence of video data, said method comprising:

compressing the sequence of video data to produce a compressed signal, said compressing comprising;

(i) compressing the sequence of video data n times, wherein n>

1, and wherein each compressing (i) of the sequence of video data said n times employs at least one different compression parameter, thereby producing n different compressed video signals from the same sequence of video data,

(ii) evaluating picture quality of said n compressed video signals in real-time, wherein said evaluating comprises comparing data loss information derived from each of said n compressed video signals, and

(iii) automatically selecting one of said n compressed video signals for output as said compressed signal, said automatically selecting being based upon said evaluating (ii) of picture quality.

2. The method of claim 1, wherein said compressing (i) comprises lossy compressing said sequence of video data, and wherein said evaluating picture quality (ii) comprises for each of said n compressed video signals evaluating data loss resulting from said lossy compressing (i) of said sequence of video data.

3. The method of claim 1, wherein said compressing of the sequence of video data comprises quantizing the sequence of video data to produce said compressed signal, and wherein said compressing (i) comprises quantizing (i) the sequence of video data n times, each quantizing employing a different quantization scale, thereby producing said n different compressed video signals.

4. The method of claim 3, wherein said evaluating (ii) of picture quality comprises for each compressed video signal of said n compressed video signals estimating in real time data loss occurring from said quantizing (i).

11. The method of claim 1, further comprising transforming the sequence of video data into a frequency domain signal prior to said compressing, and wherein said compressing comprises compressing the frequency domain signal to produce said compressed signal.

12. The method of claim 1, wherein said compressing (i) comprises employing n quantizers in parallel to compress the sequence of video data n times.

13. The method of claim 1, wherein said sequence of video data comprises a macroblock of a video frame, and wherein said method comprises repeating said compressing of the sequence of video data for each macroblock of said video frame.

5. A method for encoding a sequence of video data, said method comprising:

compressing the sequence of video data to produce a compressed signal, said compressing comprising;

(i) compressing the sequence of video data n times, wherein n≧

1, and wherein each compressing (i) of said n times employs at least one different compression parameter, thereby producing n different compressed video signals,

(ii) evaluating picture quality of said n compressed video signals in real-time,

(iii) automatically selecting one of said n compressed video signals for output as said compressed signal, said automatically selecting being based upon said evaluating (ii) of picture quality;

wherein said evaluating (ii) of picture quality comprises for each compressed video signal of said n compressed video signals estimating in real-time data loss occurring from said quantizing (i); and

wherein each of said n compressed video signals comprises a block of coefficients, and wherein said estimating of data loss comprises for each compressed video signal producing a block of difference coefficients representative of said data loss, and wherein said evaluating (ii) further comprises scaling said block of difference coefficients produced for each compressed video signal and producing therefrom quantization loss statistics representative of data loss of said compressed video signal.

6. The method of claim 5, wherein said producing of quantization loss statistics comprises for each compressed video signal summing the scaled difference coefficients in said block of scaled difference coefficients to produce said quantization loss statistics for said block.

7. The method of claim 5, wherein said scaling further comprises for each compressed video signal weighting each difference coefficient of said block of difference coefficients, wherein for each difference coefficient said weighting is based on location of said difference coefficient within said block of difference coefficients if said block comprises an intra block.

8. The method of claim 5, wherein said automatically selecting (iii) further comprises automatically selecting a compressed video signal of said n compressed video signals which has quantization loss statistics within a predefined range of acceptable quantization loss statistics.

9. The method of claim 5, wherein said automatically selecting (iii) further comprises automatically comparing the quantization loss statistics of each compressed video signal to a predefined threshold and selecting a compressed video signal obtained using a largest quantization scale in said quantizing (i) and having quantization loss statistics below said predefined threshold.

10. A method for encoding a sequence of video data, said method comprising:

compressing the sequence of video data to produce a compressed signal, said compressing comprising;

(i) compressing the sequence of video data n times, wherein n≧

1, and wherein each compressing (i) of said n times employs at least one different compression parameter, thereby producing n different compressed video signals,

(ii) evaluating picture quality of said n compressed video signals in real-time,

(iii) automatically selecting one of said n compressed video signals for output as said compressed signal, said automatically selecting being based upon said evaluating (ii) of picture quality;

wherein said evaluating (ii) of picture quality comprises for each compressed video signal of said n compressed video signals estimating in real-time data loss occurring from said quantizing (i); and

wherein estimating said data loss occurring from said quantizing (i) comprises for each coefficient of each compressed video signal determining a minimum of a remainder, produced by quantizing a corresponding coefficient of the sequence of video data with a corresponding quantization divisor used to produce said compressed video signal, or the corresponding quantization divisor minus the remainder.

14. A method for encoding a video signal, said method comprising:

encoding the video signal, said encoding comprising;

(i) compressing the video signal to produce a compressed video signal; and

(ii) automatically evaluating in real time data loss resulting from said compressing (i) of said video signal and producing therefrom loss statistics representative of picture quality of said compressed video signal, wherein said automatically evaluating occurs during encoding of said video signal.

15. The method of claim 14, wherein said compressing (i) comprises quantizing the video signal using a quantization scale to produce said compressed video signal, and wherein said encoding further comprises providing a new quantization scale if said loss statistics exceed a predefined threshold, and repeating compressing of said video signal in real time using said new quantization scale.

16. The method of claim 15, further comprising employing said compressed signal in further encoding if said loss statistics are less than said predefined threshold.

17. A method for encoding a video signal, said method comprising:

encoding the video signal, said encoding comprising;

(i) compressing the video signal to produce a compressed video signal;

(ii) automatically evaluating in real-time data loss resulting from said compressing (i) of said video signal and producing therefrom loss statistics representative of picture quality of said compressed video signal, wherein said automatically evaluating occurs during encoding of said video signal; and

wherein said compressing (i) comprises quantizing the video signal to produce said compressed video signal, and wherein said automatically evaluating in real-time (ii) comprises inverse quantizing the compressed video signal to produce a lossy video signal, and comparing a delayed version of the video signal with the lossy video signal to derive said loss statistics representative of picture quality of said compressed video signal.

18. The method of claim 17, wherein said comparing produces a block of difference coefficients, and wherein said automatically evaluating in real time (ii) further comprises scaling said block of difference coefficients and producing therefrom said loss statistics representative of data loss of said compressed video signal.

19. A digital video encoder comprising:

a quantization subsystem for compressing a sequence of video data to produce a compressed signal, said quantization subsystem comprising;

means for compressing the sequence of video data n times, wherein n>

1, and wherein each compressing of the sequence of video data n times employs a different compression parameter, thereby producing n different compressed video signals from the same sequence of video data,

means for evaluating picture quality of said n compressed video signals in real-time, wherein said means for evaluating comprises means for comparing data loss information derived from each of said n compressed video signals, and

means for automatically selecting one of said n compressed video signals for output as said compressed signal, wherein said means for automatically selecting comprises means for basing said selection upon said evaluating of picture quality.

20. The digital video encoder of claim 19, wherein said means for compressing comprises at least one quantizer, and wherein said at least one quantizer employs a different quantization scale for each quantizing of the sequence of video data n times.

21. The digital video encoder of claim 20, wherein said means for evaluating picture quality comprises means for estimating data loss in real time for each compressed video signal produced by said at least one quantizer.

25. The digital video encoder of claim 20, wherein said at least one quantizer comprises n quantizers connected in parallel within said quantization subsystem, each of said n quantizers employing a different quantization scale to produce said n different compressed video signals.

22. A digital video encoder comprising:

a quantization subsystem for compressing a sequence of video data to produce a compressed signal, said quantization subsystem comprising;

means for compressing the sequence of video data n times, wherein n≧

1, and wherein each compressing of the sequence of video data n times employs a different compression parameter, thereby producing n different compressed video signals,

means for evaluating picture quality of said n compressed video signals in real-time,

means for automatically selecting one of said n compressed video signals for output as said compressed signal, wherein said means for automatically selecting comprises means for basing said selection upon said evaluating of picture quality;

wherein said means for compressing comprises at least one quantizer, and wherein said at least one quantizer employs a different quantization scale for each quantizing of the sequence of video data n times;

wherein said means for evaluating picture quality comprises means for estimating data loss in real-time for each compressed video signal produced by said at least one quantizer; and

wherein each of said n compressed video signals comprises a block of coefficients, and wherein said means for estimating data loss comprises for each compressed video signal means for producing a block of difference coefficients representative of said data loss, and wherein said means for evaluating picture quality further comprises means for scaling each block of difference coefficients obtained for each compressed video signal and producing therefrom quantization loss statistics representative of data loss of said compressed video signal.

23. The digital video encoder of claim 22, wherein said means for scaling further comprises means for weighting each difference coefficient of said block of difference coefficients produced for each compressed video signal, wherein for each difference coefficient said weighting is based on location of said difference coefficients if said block of coefficients comprising the compressed video signal comprises an intra block.

24. The digital video encoder of claim 23, wherein said means for automatically selecting further comprises means for automatically comparing the quantization loss statistics of each compressed video signal to a predefined threshold and for selecting a compressed video signal obtained using a largest quantization scale within said at least one quantizer and having quantization loss statistics below said predefined threshold.

26. A digital video encoder comprising:

a quantizer for producing a quantized video signal from a received video signal; and

logic for automatically evaluating in real time data loss resulting from said quantizing of said video signal and producing therefrom loss statistics representative of picture quality of said quantized video signal, wherein said logic for automatically evaluating is implemented within said digital video encoder for evaluating said quantized video signal during encoding of the received video signal.

27. The digital video encoder of claim 26, wherein said quantizer produces said quantized video signal using a quantization scale, and wherein said digital video encoder further comprises means for adjusting said quantization scale if said loss statistics representative of picture quality are outside a predefined acceptable range.

28. A digital video encoder comprising:

a quantizer for producing a quantized video signal from a received video signal;

logic for automatically evaluating in real-time data loss resulting from said quantizing of said video signal and producing therefrom loss statistics representative of picture quality of said quantized video signal, wherein said logic for automatically evaluating is implemented within said digital video encoder for evaluating said quantized video signal during encoding of the received video signal; and

wherein said logic for automatically evaluating comprises an inverse quantizer for inverse quantizing the quantized video signal to produce a lossy video signal, and logic for comparing a delayed version of the received video signal with the lossy video signal to derive said loss statistics representative of picture quality of said quantized video signal.

29. The digital video encoder of claim 28, wherein said logic for comparing signals comprises logic for producing a block of difference coefficients, and for scaling said block of difference coefficients and for producing therefrom said loss statistics representative of data loss of said compressed video signal.

30. A digital video encoder comprising:

a quantizer for producing a quantized video signal from a received video signal;

logic for automatically evaluating in real-time data loss resulting from said quantizing of said video signal and producing therefrom loss statistics representative of picture quality of said quantized video signal, wherein said logic for automatically evaluating is implemented within said digital video encoder for evaluating said quantized video signal during encoding of the received video signal; and

wherein said quantizer comprises one quantizer of multiple quantizers connected in parallel within said digital video encoder to produce multiple quantization video signals from said received video signal, each quantizer of said multiple quantizers employing a different quantization scale, and wherein said logic for automatically evaluating comprises logic for automatically evaluating in real-time data loss of each of said quantized video signals, and for selecting an optimum quantization video signal for further encoding within said digital video encoder.

31. A computer program product comprising a computer usable medium having computer readable program code means therein for use in encoding a sequence of video data, said computer readable program code means in said computer program product comprising:

computer readable program code means for causing a computer to affect compressing the sequence of video data to produce a compressed signal, said compressing comprising;

(i) compressing the sequence of video data n times, wherein n>

1, and wherein each compressing (i) of the sequence of video data said n times employs at least one different compression parameter, thereby producing n different compressed video signals from the same sequence of video data,

(ii) evaluating picture quality of said n compressed video signals in real-time, wherein said evaluating comprises comparing data loss information derived from each of said n compressed video signals, and

(iii) automatically selecting one of said n compressed video signals for output as said compressed signal, said automatically selecting being based upon said evaluating (ii) of picture quality.

32. The computer readable program code means of claim 31, wherein said computer readable program code means for causing a computer to affect compressing of the sequence of video data comprises computer readable program code means for causing a computer to affect quantizing the sequence of video data to produce said compressed signal, and wherein said compressing (i) comprises quantizing (i) the sequence of video data n times, each quantizing employing a different quantization scale, thereby producing said n different compressed video signals.

33. The computer readable program code means of claim 32, wherein said evaluating (ii) of picture quality comprises for each compressed video signal of said n compressed video signals estimating in real time data loss occurring from said quantizing (i).

34. The computer readable program code means of claim 33, wherein said quantizing (i) comprises quantizing the sequence of video data n times in parallel.

35. A computer program product comprising a computer usable medium having computer readable program code means therein for use in encoding a video signal, said computer readable program code means in said computer program product comprising:

computer readable program code means for causing a computer to affect encoding of the video signal, said encoding comprising;

(i) compressing the video signal to produce a compressed video signal; and

(ii) automatically evaluating in real time data loss resulting from said compressing (i) of said video signal and producing therefrom loss statistics representative of picture quality of said compressed video signal, wherein said automatically evaluating occurs during encoding of said video signal.

36. The computer readable program code means of claim 35, wherein said compressing (i) comprises quantizing the video signal using a quantization scale to produce said compressed video signal, and wherein said computer readable program code means for causing a computer to affect encoding of the video signal further comprises computer readable program code means for causing a computer to affect providing a new quantization scale if said loss statistics exceed a predefined threshold, and repeating compressing of said video signal in real time using said new quantization scale.

37. The computer readable program code means of claim 35, wherein said compressing (i) comprises quantizing the video signal to produce said compressed video signal, and wherein said automatically evaluating in real time (ii) comprises inverse quantizing the compressed video signal to produce a lossy video signal, and comparing a delayed version of the video signal with the lossy video signal to derive said loss statistics representative of picture quality of said compressed video signal.

38. The computer readable program code means of claim 37, wherein said comparing produces a block of difference coefficients, and wherein said automatically evaluating in real time (ii) further comprises scaling said block of difference coefficients and producing therefrom said loss statistics representative of data loss of said compressed video signal.