Range reduction

• US 7,502,415 B2
• Filed: 07/16/2004
• Issued: 03/10/2009
• Est. Priority Date: 07/18/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Srinivasan, Sridhar, Hsu, Pohsiang, Holcomb, Thomas W., Lin, Bruce Chih-Lung
• Primary Examiner(s): SENFI, BEHROOZ M
• Application Number:
  US10/893,801
• Publication Number:
  US 20050013373A1
• Time in Patent Office: 1,698 Days
• Field of Search: 348/426, 348/261.2, 348/411, 348/412, 348/413, 348/414, 348/398, 348/402, 348/420.1, 348/719, 348/699, 348/397, 348/404, 348/405, 348/407, 348/416, 348/427, 348/443, 348/444, 382/248, 382/238, 382/250, 382/246, 382/260, 382/298, 382/299, 382/300, 382/233, 382/234, 382/236, 382/239, 382/251, 375/240.16, 375/240.12, 375/240.21, 375/240.17, 375/240.18, 375/240.19, 375/240.2, 375/240.25, 375/240.15
• US Class Current: 375/240.18
• CPC Class Codes: H04N 19/126 Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
• CPC Class Codes: H04N 19/136 Incoming video signal characteristics or properties
• CPC Class Codes: H04N 19/182 the unit being a pixel
• CPC Class Codes: H04N 19/44 Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
• CPC Class Codes: H04N 19/46 Embedding additional information in the video signal during the compression process H04N19/517, H04N19/68, H04N19/70 take precedence
• CPC Class Codes: H04N 19/51 Motion estimation or motion compensation
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/70 characterised by syntax aspects related to video coding, e.g. related to compression standards

Abstract

Techniques and tools are described for decoding video data having samples that have been scaled in the spatial domain. For example, a decoder receives a bit stream that includes coded video data for a current frame. The decoder processes at least one syntax element (e.g., sequence layer flag, frame layer flag) that indicates whether the current frame should be scaled up in value in a spatial domain. If so, then the samples for the current frame are scaled up in value in the spatial domain. As another example, for a reference frame used in motion compensation for a current frame, a decoder scales samples of the reference frame so the range of the reference frame matches the range of the current frame.

Claims

1. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

receiving a bit stream at the video decoder, the bit stream comprising coded video data for a current frame in a transform domain;

with the video decoder, processing at least one syntax element that indicates whether samples of the current frame should be scaled up in value in a spatial domain;

with the video decoder, decoding the video data for the current frame to produce a reconstructed current frame in the spatial domain; and

if the at least one syntax element indicates that the samples of the current frame should be scaled up in value in the spatial domain, then with the video decoder scaling up the samples of the reconstructed current frame, including mapping samples of at least one plane of the reconstructed current frame from a smaller range to a larger range, wherein the mapping comprises scaling original values in the smaller range to produce new values in the larger range as follows;

Y_p[n]=CLIP(((Y[n]−

128)<

<

1)+128);

U_p[n]=CLIP(((U[n]−

128)<

<

1)+128); and

V_p[n]=CLIP(((V[n]−

128)<

<

1)+128).

2. The method of claim 1, wherein the at least one syntax element comprises a sequence flag that indicates whether samples of any of a plurality of frames in a sequence should be scaled up in value in the spatial domain.

3. The method of claim 2, wherein the sequence flag is part of an initial sequence header for the bit stream.

4. The method of claim 2, wherein, if the sequence flag indicates that the samples of at least one of the plurality of frames should be scaled up in value in the spatial domain, then the at least one syntax element further comprises a current frame flag that indicates whether the samples of the current frame should be scaled up in value in the spatial domain.

5. The method of claim 4, wherein the current frame flag is part of a frame header for the current frame.

6. The method of claim 1, wherein a midpoint of the larger range is substantially the same as a midpoint of the smaller range.

7. The method of claim 1, wherein:

the larger range is from 0 to 255; and

the smaller range is from 64 to 191.

8. The method of claim 1, wherein the method further comprises, during the decoding, performing motion compensation for the current frame using video data for a reference frame, including:

matching a range of the current frame with a range of the reference frame; and

reconstructing the current frame based at least in part on the reference frame.

9. The method of claim 8, wherein matching the range of the current frame with the range of the reference frame comprises scaling samples of the reference frame.

10. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

receiving coded video data in a bit stream at the video decoder, the video data comprising coded video data for a current frame;

with the video decoder, determining whether the current frame has been range-reduced in a pixel domain and whether a reference frame for the current frame has been range-reduced in the pixel domain;

with the video decoder, selectively performing motion compensation for the current frame using a range-reduced reference frame in the motion compensation, including, if the current frame but not the reference frame has been range-reduced, then scaling samples of the reference frame to reduce range of the reference frame such that the range of the reference frame matches range of the current frame and performing the motion compensation using the range-reduced reference frame after the scaling; and

with the video decoder, reconstructing the current frame based at least in part on the motion compensation from the reference frame.

11. The method of claim 10, wherein the determining comprises processing a sequence flag that indicates whether any frames of the sequence have been range-reduced in the pixel domain.

12. The method of claim 11, wherein the sequence flag is part of an initial sequence header for the bit stream.

13. The method of claim 10, wherein the determining comprises processing a current frame flag that indicates whether the current frame has been range-reduced in the pixel domain and processing a reference frame flag that indicates whether the reference frame has been range-reduced in the pixel domain.

14. The method of claim 13, wherein the reference frame flag is part of a frame header for the reference frame and the current frame flag is part of a frame header for the current frame.

15. The method of claim 10, wherein the scaling the samples of the reference frame comprises scaling down as follows:

Y_p[n]=((Y[n]−

128)<

<

1)+128;

U_p[n]=((U[n]−

128)<

<

1)+128; and

V_p[n]=((V[n]−

128)<

<

1)+128.

16. The method of claim 10, wherein the scaling the samples of the reference frame comprises scaling up as follows:

Y_p[n]=CLIP(((Y[n]−

128)<

<

1)+128);

U_p[n]=CLIP(((U[n]−

128)<

<

1)+128); and

V_p[n]=CLIP(((V[n]−

128)<

<

1)+128).

17. The method of claim 10, wherein reconstructing the current frame further comprises transforming the video data for the current frame from a transform domain to the pixel domain.

22. The method of claim 10 wherein the selectively performing motion compensation further includes:

if both of the current frame and the reference frame have been range-reduced, performing the motion compensation using the range-reduced reference frame.

18. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

with the video decoder, parsing a sequence layer syntax element that indicates whether range reduction is possible for a sequence of frames;

if the sequence layer syntax element indicates that range reduction is possible for the sequence of frames, then with the video decoder performing additional processing that allows the use of range reduction for the sequence of frames, wherein range reduction includes scaling samples from a larger range of sample values to a smaller range of sample values, and wherein the additional processing includes;

determining whether range reduction has been used for a current frame and whether range reduction has been used for a reference frame of the current frame;

if range reduction has been used for the current frame but not the reference frame, scaling samples of the reference frame down to reduce range prior to using the range-reduced reference frame in motion compensation for the range-reduced current frame; and

if the sequence layer syntax element indicates that range reduction is not possible for the sequence of frames, then with the video decoder skipping the additional processing that allows the use of range reduction for the sequence of frames.

19. The method of claim 18, further comprising with the video decoder parsing a picture level syntax element for the current frame if the sequence layer syntax element indicates that range reduction is possible for the sequence of frames, the picture level syntax element for the current frame indicating whether range reduction has been used for the current frame.

20. The method of claim 19, further comprising with the video decoder scaling up samples of the current frame after it has been decoded and prior to display only if the picture-level syntax element indicates that range reduction has been used for the current frame.

21. The method of claim 19, further comprising with the video decoder parsing a picture level syntax element for the reference frame, the picture level syntax element for the reference frame indicating whether range reduction has been used for the reference frame.

23. The method of claim 18 wherein the additional processing further includes:

if range reduction has been used for the reference frame but not the current frame, scaling samples of the range-reduced reference frame up prior to using the scaled reference frame in motion compensation for the current frame;

if range reduction has been used for both the reference frame and the current frame, using the range-reduced reference frame in the motion compensation for the range-reduced current frame; and

if range reduction has not been used for the reference frame or the current frame, using the reference frame in the motion compensation for the current frame.

24. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

receiving coded video data in a bit stream at the video decoder, the video data comprising coded video data for a current frame;

with the video decoder, determining whether the current frame has been range-reduced in a pixel domain and whether a reference frame for the current frame has been range-reduced in the pixel domain;

if one but not both of the current frame and reference frame have been range-reduced, then with the video decoder scaling samples of the reference frame such that a range of the reference frame matches a range of the current frame, wherein the scaling the samples of the reference frame comprises scaling down as follows;

Y_p[n]=((Y[n]−

128)<

<

1)+128;

U_p[n]=((U[n]−

128)<

<

1)+128; and

V_p[n]=((V[n]−

128)<

<

1)+128; and

with the video decoder, reconstructing the current frame based at least in part on motion compensation from the reference frame.

25. A computer-implemented method of using a video decoder to transform video from a compressed form to a reconstructed form, the method comprising:

receiving coded video data in a bit stream at the video decoder, the video data comprising coded video data for a current frame;

with the video decoder, determining whether the current frame has been range-reduced in a pixel domain and whether a reference frame for the current frame has been range-reduced in the pixel domain;

if one but not both of the current frame and reference frame have been range-reduced, then with the video decoder scaling samples of the reference frame such that a range of the reference frame matches a range of the current frame, wherein the scaling the samples of the reference frame comprises scaling up as follows;

Y_p[n]=CLIP(((Y[n]−

128)<

<

1)+128);

U_p[n]=CLIP(((U[n]−

128)<

<

1)+128); and

V_p[n]=CLIP(((V[n]−

128)<

<

1)+128); and

with the video decoder, reconstructing the current frame based at least in part on motion compensation from the reference frame.