HDR Image Tone-Mapping Encoding for Legacy Display Compatibility

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Legacy digital images with 24-bit formats are limited in dynamic range, failing to accurately represent real-world scenes, while modern HDR images exceed the capabilities of legacy displays, necessitating a solution for efficient encoding and rendering across both modern and legacy display technologies.

Innovation Solution

The proposed solution involves encoding HDR images using a tone-mapping process that generates a tone-mapped base image and additional metadata, allowing both modern HDR and legacy LDR displays to render images effectively. This includes histogramming log-luminances, multi-scale filtering, and quantization, with the use of JPEG-HDR format to support wide color gamut and multiple color spaces, enabling backwards compatibility and enhanced dynamic range rendering.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If HDR images are encoded with full dynamic range, then image fidelity and detail representation are improved, but compatibility with legacy LDR displays deteriorates

Engineering Contradiction:
Improveimage fidelityVSAvoiddisplay compatibility
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The encoding process segments HDR image data into multiple components: a base image at reduced dynamic range and separate metadata containing the difference information. This segmentation allows legacy displays to render the base image while HDR-capable displays can recover the full dynamic range using the metadata, thus resolving the contradiction between fidelity and compatibility.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a tone-mapped base image as an intermediary representation that bridges HDR and LDR domains. This base image serves as a compatible intermediate form that can be displayed on legacy devices, while preserving the ability to reconstruct full HDR information on modern displays through the metadata.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If legacy 24-bit formats are used, then compatibility with existing displays is maintained, but dynamic range representation capability deteriorates

Engineering Contradiction:
Improvedisplay compatibilityVSAvoiddynamic range representation
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent extends the traditional 24-bit color space by adding a temporal or layered dimension through metadata. The base image maintains 24-bit compatibility while the appended metadata layer provides additional dynamic range information, effectively moving from a single-plane to a multi-plane data structure that preserves both compatibility and enhanced capability.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Adaptability or versatility

If tone-mapping is applied to reduce dynamic range, then compatibility with legacy displays is improved, but image detail and contrast information are lost

Engineering Contradiction:
Improvelegacy display compatibilityVSAvoidimage detail
Core Design Contradiction:
Adaptability or versatilityVSLoss of information

Solution Approach 1:

The patent discards full dynamic range information only temporarily for legacy display compatibility, but recovers it by storing the difference between original HDR and tone-mapped versions in metadata. This allows information to be discarded for compatibility purposes while being preserved for later recovery on HDR-capable devices.

Inventive Principle:
Principle #34Discarding and recovering

Data Source

PatentEP3168809B1Image processing for HDR images
Publication Date: 2023.08.30 DOLBY LABORATORIES LICENSING CORP
  • EP3168809B1 patent drawingFigure 1A
  • EP3168809B1 patent drawingFigure 1B
  • EP3168809B1 patent drawingFigure 2

AI summary

Image encoding is described. Log-luminances in an HDR input image are histogrammed to generate a tone-map, along with which a log global tone-mapped luminance image is computed. The log global tone-mapped luminance image is downscaled. The log-luminances and the log global tone-mapped luminance image generate a log ratio image. Multi-scale resolution filtering the log ratio image generates a log multi-scale ratio image. The log multi-scale ratio image and the log luminances generate a second log tone-mapped image, which is normalized to output a tone-mapped image based on the downscaled log global tone-mapped luminance image and the normalized image. The HDR input image and the output tone-mapped image generate a second ratio image, which is quantized.