Distribution of multi-format high dynamic range video using layered coding

Patent metadata

Field	Value
Publication number	US-2016014422-A1
Application number	US-201414772051-A
Country	US
Kind code	A1
Filing date	Feb 14, 2014
Priority date	Mar 11, 2013
Publication date	Jan 14, 2016
Grant date	—

Abstract

An encoder receives an input enhanced dynamic range (EDR) image to be stored or transmitted using multiple coding formats in a layered representation. A layer decomposer generates a lower dynamic range (LDR) image from the EDR image. One or more base layer (BL) encoders encode the LDR image to generate a main coded BL stream and one or more secondary coded BL streams, where each secondary BL stream is coded in a different coding format than the main coded BL stream. A single enhancement layer (EL) coded stream and related metadata are generated using the main coded BL stream, the LDR image, and the input EDR image. An output coded stream includes the coded EL stream, the metadata, and either the main coded BL stream or one of the secondary coded BL streams. Computation-scalable decoding and display management processes for EDR images are also described.

First claim

What is claimed is:

1 - 29 . (canceled)

30 . A method for encoding an enhanced dynamic range (EDR) image using multiple coding formats, the method comprising:
receiving an input EDR image with a first spatial resolution; decomposing the input EDR image with a layer decomposer to generate a lower dynamic range (LDR) image; coding the LDR image with a first base layer (BL) encoder to generate a first coded BL stream;
characterized by
coding the LDR image with a second BL encoder to generate a second coded BL stream, wherein the second coded BL stream comprises at least one coding format parameter different than the coding format parameter of the first coded BL stream, the coding format parameter comprising a coding algorithm, a coding bit rate, or a spatial resolution of the coded bit stream;
applying a prediction function with a predictor to the first coded BL stream to generate a predicted EDR image;
computing a residual image representing differences between pixel values in the predicted EDR image and the input EDR image;
applying a clip mask to the residual image to generate a clip-mask image, wherein applying a clip mask comprises setting pixels values of the residual image to a predetermined fixed value if the corresponding pixel value of the LDR image is below a predetermined enhancement layer (EL) threshold;
applying an EL quantizer to the clip-mask image to generate an EL image;
coding the EL image with an EL encoder to generate a coded EL stream; and
encoding and storing data representing the first coded BL stream, the second coded BL stream and the single coded EL stream, and a single set of metadata representative of the prediction function, wherein both the first coded BL stream and the second coded BL stream share the single coded EL stream;
transmitting said data representing the coded EL stream, the prediction function, and, depending on the required coding format parameter, either the first coded BL stream or the second coded BL stream.

31 . Method according to claim 30 , wherein the first and second coded BL streams are non-backward compatible and wherein the first and second coded BL streams together with the single coded El stream are optimized for reconstructing high quality EDR images for viewing on EDR displays.

32 . Method as claimed in claim 30 , wherein the image data in a coded BL stream is optimized to contain an optimal amount of image data in a lower bit depth image container for the purpose of minimizing an overall bit requirement for the coded EDR image and to improve the overall quality of a final decoded EDR image.

33 . The method of claim 30 , further comprising down-sampling the clip-mask image to a second spatial resolution lower than the first spatial resolution to generate a down-sampled EL image, wherein the down-sampling may be executed before or after applying the EL quantizer, and the coded EDR stream is based on the down-sampled EL image.

34 . The method of claim 30 , wherein the second coded BL stream is coded in a different bit rate than the first coded BL stream.

35 . The method claim 30 , wherein the second coded BL stream is coded at a different spatial resolution than the first coded BL stream.

36 . The method of claim 30 , wherein the first BL encoder complies to a first coding algorithm and the second BL encoder complies to a second coding algorithm which is different than the first coding algorithm.

37 . The method claim 30 , wherein the first BL encoder comprises an HEVC-compliant encoder and the EL encoder comprises an H.264-compliant encoder;

38 . The method claim 30 , wherein the first coded BL stream and the coded EL stream are coded using a bit-depth of 8 bits, the predetermined EL threshold is 250, and the predetermined fixed value is 0.

39 . A method to reduce the computational load during the decoding of high-resolution enhanced dynamic range (EDR) signals encoded according to the method of any of the preceding claims, the method comprising:
receiving a coded stream comprising a coded base layer (BL) stream in a first spatial resolution and a lower dynamic range (LDR) than a target enhanced dynamic range (EDR), a coded enhancement layer stream in a second spatial resolution and a lower dynamic range (LDR) than the target EDR, and parameters for a prediction function, wherein the first spatial resolution is higher than the second spatial resolution; decoding the coded BL stream with a BL decoder to generate a decoded BL image; decoding the coded EL stream with an EL decoder to generate a decoded EL image; down-sampling the decoded BL image to generate a second BL image with the second spatial resolution; generating a predicted EDR image by applying the prediction function to the second BL image; compositing a first EDR image using the predicted EDR image and the decoded EL image; and up-sampling the first EDR image to generate an output EDR image with the first spatial resolution.

40 . A method for encoding an enhanced dynamic range (EDR) image using multiple coding formats, the method comprising:
receiving an input EDR image with a first spatial resolution; decomposing the input EDR image with a layer decomposer to generate a lower dynamic range (LDR) image; coding the LDR image with a first base layer (BL) encoder to generate a coded BL stream;
characterized by
applying a prediction function with a predictor to the coded BL stream to generate a predicted EDR image;
computing a residual image representing differences between pixel values in the predicted EDR image and the input EDR image;
applying a clip mask to the residual image to generate a clip-mask image, wherein applying a clip mask comprises setting pixels values of the residual image to a predetermined fixed value if the corresponding pixel value of the LDR image is below a predetermined enhancement layer (EL) threshold;
applying an EL quantizer to the clip-mask image to generate an EL image;
coding the EL image with a first EL encoder to generate a first coded EL stream;
coding the EL image with a second EL encoder to generate a second coded EL stream, wherein the second coded EL stream comprises at least one coding format parameter different than the coding format parameter of the first coded EL stream, the coding format parameter comprising a coding algorithm, a coding bit rate, or a spatial resolution of the coded bit stream; and
encoding and storing data representing the single coded BL stream, the prediction function, the first coded EL stream and the second coded EL stream;
transmitting said data representing the coded BL stream, the prediction function, and, depending on the required coding-format, either the first coded EL stream or the second coded EL stream.

41 . The method of claim 40 , further comprising down-sampling the EL image to generate a down-sampled EL image and generating the second coded EL stream based on the down-sampled EL image.

42 . The method of claim 40 , wherein the coding format parameter comprises a coding algorithm, a coding bit rate, or a spatial resolution of the coded bi stream.

43 . The method of claim 40 , wherein the second coded EL stream is coded in a different bit rate than the first coded BL stream.

44 . The method of claim 40 , wherein the second coded EL stream is coded at a different spatial resolution than the first coded EL stream.

45 . The method of claim 40 , wherein the first EL encoder complies to a first coding algorithm and the second EL encoder complies to a second coding algorithm which is different than the first coding al

Assignees

• Dolby Lab Licensing Corp

Inventors

• Su Guan-Ming
• Hulyalkar Samir N
• Chen Tao
• Qu Sheng

Classifications

• G06T3/40

  Scaling of whole images or parts thereof, e.g. expanding or contracting · CPC title

• H04N19/157

  Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter · CPC title

• H04N19/44

  Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder · CPC title

• H04N19/593

  involving spatial prediction techniques · CPC title

• H04N19/124

  Quantisation · CPC title