Efficient architecture for layered VDR coding

1. In a layered video encoder, a method of coding video signals to reduce color format conversion operations, the method comprising:

receiving a first video signal having a first dynamic range in a first color space and a first chrominance sampling format of a first color format;

receiving a second video signal corresponding to the same content as the first video signal, wherein the second video signal has a second dynamic range in a second color space and a second chrominance sampling format of a second color format, wherein the second dynamic range is smaller than the first dynamic range, wherein the second chrominance sampling format of the second color format is different from the first chrominance sampling format of the first color format, and wherein the second color space of the second color format is different from the first color space of the first color format;

generating a third video signal based on the first video signal by converting the first video signal in the first color space of the first color format into the third video signal in the second color space and the second chrominance sampling format of the second color format while preserving the first dynamic range;

determining a predicted video signal of the third video signal in the second color space of the second color format based at least on the second video signal, the predicted video signal having the first dynamic range in the second color space and the second chrominance sampling format of the second color format;

coding the first video signal using a coded base layer stream and a coded enhancement layer stream of a coded bitstream, wherein the coded base layer stream comprises a coded representation of the second video signal, and the coded enhancement layer stream comprises a coded representation of a residual between the third video signal and the predicted video signal;

causing a recipient device of the coded bitstream to generate a decoded image from the coded bitstream for rendering.

2. The method of claim 1 , wherein the first video signal comprises a visual dynamic range (VDR) signal and the second video signal comprises a standard dynamic range (SDR) signal.

3. The method of claim 2 , wherein the VDR signal comprises an RGB 4:4:4 color format and the SDR signal comprises a YUV or YCbCr 4:2:0 color format.

4. The method of claim 1 , wherein determining the predicted video signal comprises using a multiple-color channel multiple-regression (MMR) predictor.

5. The method of claim 1 , wherein generating the coded enhancement layer stream further comprises:

quantizing the residual with a non-linear quantizer to generate a quantized residual signal with the second dynamic range; and

coding the quantized residual signal with a video encoder to generate the coded enhancement layer stream.

6. The method of claim 1 , wherein in a decoder, a method of decoding the coded first video signal comprises:

decoding the coded base layer stream to generate a decoded base layer signal in the second color format;

decoding the coded enhancement layer stream to generate a decoded enhancement layer signal in the second color format;

determining, using a predictor, a decoder predicted signal in the second color format in response to the decoded base layer signal;

generating an intermediate output signal based on the decoded enhancement layer signal and the decoder predicted signal, the intermediate output signal having the first dynamic range and the second color format; and

determining a decoded representation of the first signal by transforming the intermediate output signal from the second color format into a final output color format, the final output color format being different from the second color format.

7. The method of claim 6 , wherein the final output color format comprises the first color format.

8. A method of decoding a visual dynamic range (VDR) video signal in a video codec comprising a base layer and an enhancement layer, the method comprising:

receiving a coded VDR stream comprising a base layer stream and an enhancement layer stream, wherein both the base layer stream and the enhancement layer stream are to be used to generate a reconstructed version of a first video signal having a first dynamic range in a first color space and a first chrominance sampling format of a first color format;

wherein the base layer stream comprises a coded representation of a second video signal having a second dynamic range in a second color space and a second chrominance sampling format of a second color format;

wherein the enhancement layer stream comprises a coded representation of a residual between a third video signal and a predicted video signal of the third video signal;

wherein the third video signal was generated by an upstream encoder based on the first video signal by converting the first video signal in the first color space of the first color format into the third video signal in the second color space and the second chrominance sampling format of the second color format while preserving the first dynamic range;

wherein the predicted video signal of the third video signal was determined by the upstream encoder based at least on the second video signal, the predicted video signal having the first dynamic range in the second color space and the second chrominance sampling format of the second color format;

wherein the second chrominance sampling format of the second color format is different from the first chrominance sampling format of the first color format and wherein the second color space of the second color format is different from the first color space of the first color format;

decoding the base layer stream to generate the second video signal having the second dynamic range in the second color space and the second chrominance sampling format of the second color format;

decoding the enhancement layer stream to generate the residual between the third video signal and the predicted video signal of the third video signal;

generating, using a predictor, the predicted video signal of the third video signal in response to the decoded second video signal;

generating a VDR intermediate output signal in the second color space and the second chrominance sampling format of the second color format based on the residual between the third video signal and the predicted video signal of the third video signal and the predicted video signal of the third video signal;

transforming the VDR intermediate signal from the second color space and the second chrominance sampling format of the second color format to the first color space and the first chrominance sampling format of the first color format to generate the reconstructed version of the first video signal having the first dynamic range in the first color space and the first chrominance sampling format of the first color format;

causing a reconstructed image in the reconstructed version of the first video signal to be rendered.

9. The method of claim 8 , wherein the second color format comprises a YUV or YCbCr 4:2:0 color format.

10. The method of claim 8 , wherein the first color format comprises an RGB 4:4:4 color format.

11. The method of claim 8 , wherein generating the decoded VDR enhancement layer signal further comprises:

decoding the enhancement layer stream with a video decoder to generate an SDR enhancement layer signal;

dequantizing the decoded SDR enhancement layer signal with a non-linear dequantizer to generate the decoded VDR enhancement layer signal.

12. An apparatus comprising one or more processors and one or more non-volatile computer-readable storage media storing instructions, which when executed by the one or more processors, cause the one or more processors to perform a method as recited in claim 1 .

13. An apparatus comprising one or more processors and one or more non-volatile computer-readable storage media having stored instructions, which when executed by the one or more processors, cause the one or more processors to perform a method as recited in claim 8 .

14. One or more non-volatile computer-readable storage media having stored instructions, which when executed by one or more processors, cause the one or more processors to perform a method as recited in claim 1 .

15. One or more non-volatile computer-readable storage media having stored instructions, which when executed by one or more processors, cause the one or more processors to perform a method as recited in claim 8 .