GPU Stereoscopic View Processing via Primitive Generator
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current graphics processing technologies face challenges in rendering stereoscopic 3D graphics without modifying existing applications or graphics processing units (GPUs), as they require conversion of mono view images to stereoscopic views in real-time, while managing attribute information for multiple views efficiently.
Innovation Solution
A graphics processing unit (GPU) generates attribute information for corresponding primitives in different views by combining or keeping attribute information separate, depending on storage and processing thresholds, and uses a primitive generator to output this information to a vertex shader program, allowing for efficient rendering of stereoscopic views without modifying the application or GPU.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If attribute information for multiple views is processed separately, then processing accuracy is maintained, but processing time and power consumption increase
Solution Approach 1:
The patent segments the processing of attribute information by creating a primitive generator that selectively combines or separates attribute information based on view count thresholds. For a first number of views, attribute information is combined into a single set, while for a second number of views, separate attribute information is generated for each view. This segmentation allows the system to optimize between processing efficiency and accuracy based on the specific rendering scenario.
Solution Approach 2:
The system dynamically adjusts its processing mode based on the number of views being rendered. The primitive generator automatically switches between combining attribute information (for fewer views) and maintaining separate attribute information (for more views), allowing the processing approach to adapt to changing rendering requirements and optimize both time and accuracy.
2Productivity
If attribute information is combined for multiple views, then processing efficiency improves, but storage requirements increase
Solution Approach 1:
The patent segments storage requirements by conditionally combining or separating attribute information based on view count. For a first number of views, attribute information is combined to reduce storage usage, while for a second number of views, separate attribute information is maintained. This segmentation allows the system to balance storage efficiency with the processing benefits of having separate view data when needed.
Solution Approach 2:
The system changes the storage parameter dynamically based on the number of views. When the view count exceeds a threshold, the system transitions from combining attribute information (lower storage) to maintaining separate attribute information (higher storage), allowing optimization of storage usage based on the specific rendering scenario.
3Measurement precision
If vertex shader program is executed multiple times for different views, then view-specific accuracy is improved, but power consumption increases
Solution Approach 1:
The patent segments the vertex shader execution strategy by using a primitive generator that combines attribute information for a first number of views, allowing a single vertex shader execution to handle multiple views. For a second number of views, separate attribute information enables targeted shader executions. This segmentation reduces the total number of shader executions and associated power consumption while maintaining view-specific accuracy when necessary.
Solution Approach 2:
The system uses copying by generating a single set of combined attribute information that can be reused across multiple views, avoiding the need to process each view separately through the vertex shader. This copying approach significantly reduces power consumption by eliminating redundant shader executions while still producing accurate view-specific results through the combined attribute data.
4Adaptability or versatility
If existing applications and GPUs are modified to support stereoscopic rendering, then rendering capability is improved, but device complexity increases
Solution Approach 1:
The patent introduces a primitive generator as an intermediary component between the existing graphics pipeline and the stereoscopic rendering requirements. This intermediary selectively combines or separates attribute information based on view count thresholds, enabling stereoscopic rendering capability without modifying existing applications or GPUs. The intermediary handles the complexity of multi-view attribute management while leaving the core rendering system unchanged.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
Techniques are described for stereoscopic view generation. A graphics processing unit (GPU) may combine attribute information for two or more corresponding vertices of corresponding primitives in different views. The GPU may process the combined attributed information to generate graphics data for the stereoscopic view.