SoC Floating-Point Image Processing for HDR Dynamic Range
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mobile devices and systems on chip struggle to produce high-quality images due to limitations in dynamic range, which restricts the realization of realistic image quality.
Innovation Solution
A mobile device and system on chip configuration that includes a processor for converting raw image data into floating-point format, a memory for storing this data, and a display processing unit capable of performing high dynamic range (HDR) processing on the floating-point format image data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If LDR or SDR processing is used in mobile devices, then device complexity is reduced and processing speed is improved, but image quality and dynamic range are limited
Solution Approach 1:
The processor performs preliminary conversion of raw image data into floating-point format image data before it is stored in memory. This preliminary action ensures that when the display processing unit retrieves the data, it is already in the appropriate format for HDR processing, eliminating the need for format conversion at display time and reducing overall system complexity
Solution Approach 2:
The system changes the data format parameter from traditional fixed-point (LDR/SDR) to floating-point format, enabling HDR processing. This parameter change allows the image data to represent a wider dynamic range (up to 10,000 nits) while maintaining compatibility with existing processing pipelines through the use of standard floating-point arithmetic operations
2Manufacturing precision
If floating-point format image data is stored in memory for HDR processing, then image quality and dynamic range are improved, but memory usage and data size increase
Solution Approach 1:
The system uses floating-point format (e.g., 32-bit float) to represent image data, which allows for a much wider dynamic range compared to traditional 8-bit or 16-bit fixed-point formats. Although floating-point numbers use more bits per pixel, the trade-off is justified by the ability to represent both very bright and very dark areas simultaneously, achieving up to 10,000 nits dynamic range as perceived by the human eye
3Manufacturing precision
If HDR processing is performed on floating-point format image data, then dynamic range and image realism are extended, but processing time and computational load increase
Solution Approach 1:
The processor converts raw image data to floating-point format in advance, before the data is stored in memory. This preliminary conversion ensures that when the display processing unit needs to perform HDR processing, the data is already in the correct format, eliminating the need for time-consuming format conversion during display operations
Solution Approach 2:
The patent merges the raw image data processing and floating-point conversion functions into the processor, and integrates the display processing unit with memory and processor into a unified system on chip (SoC). This consolidation allows for optimized data flow and reduces the time required for data transfer and processing between separate components
Data Source
AI summary
A system on chip and a mobile device are provided, the mobile device including a processor that receives raw image data, processes the raw image data into floating-point format image data, and output the floating-point format image data; a memory that stores therein the floating-point format image data; and a display processing unit that receives the floating-point format image data stored in the memory therefrom, and performs high dynamic range (HDR) processing on the floating-point format image data.