Viewport State Data Objects for Video Compression

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video compression methods are inefficient in reducing data requirements for rendering realistic video frames, as they often require re-rendering each frame from scratch, which is computationally intensive and bandwidth-heavy, especially when rendering from different spatial locations and view transforms.

Innovation Solution

The use of Viewport State Data Objects (VSDOs) and corresponding render instructions allows for the efficient re-use of computing resources and bandwidth by enabling a video render client to render multiple frames from a single set of VSDOs, using techniques like layered cube maps and fragment shaders to trace rays and update frames based on spatial locations and view transform parameters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If traditional video compression methods re-render each frame from scratch, then realistic video frames can be generated, but computational load and bandwidth requirements increase significantly

Engineering Contradiction:
Improverendering qualityVSAvoidcomputational load
Core Design Contradiction:
Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

The patent performs preliminary rendering by pre-computing view-dependent information (depth buffers, normal maps, texture coordinates) and storing them as VSDO assets before video playback. This allows the rendering system to skip frame-by-frame re-rendering and instead reconstruct frames by combining pre-computed assets with minimal view transform calculations, dramatically reducing real-time computational load while maintaining rendering quality

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates reusable VSDO asset copies that can be shared across multiple frames. Instead of rendering each frame independently, the system copies and reuses pre-rendered view state data objects, applying only necessary view transforms to generate new frames. This copying approach eliminates redundant rendering computations while preserving visual fidelity

Inventive Principle:
Principle #26Copying

2Productivity

If traditional video compression methods re-render each frame from scratch, then complete video sequences can be generated, but bandwidth requirements increase significantly

Engineering Contradiction:
Improvevideo frame generationVSAvoiddata transmission
Core Design Contradiction:
ProductivityVSQuantity of substance

Solution Approach 1:

The system performs preliminary rendering offline to generate VSDO assets containing view-dependent geometry, materials, and lighting information. These pre-computed assets are transmitted once to the client, enabling local frame reconstruction without requiring continuous transmission of full video frames. This dramatically reduces bandwidth consumption while maintaining complete video sequence generation capability

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent transmits compact VSDO asset copies instead of large video frame data. The client reconstructs video frames by combining these small asset copies with local view transform parameters, reducing data transmission requirements by orders of magnitude compared to traditional methods that transmit or re-render complete frames for each timestamp

Inventive Principle:
Principle #26Copying

3Manufacturing precision

If ray tracing algorithms are used to generate realistic images, then image quality improves, but computational complexity increases

Engineering Contradiction:
Improveimage realismVSAvoidcomputational complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent performs computationally intensive ray tracing operations in advance during VSDO asset generation, storing the results in optimized data structures. This preliminary ray tracing captures complex lighting, shadows, and material interactions once, allowing subsequent frame generation to reuse these pre-computed effects with minimal additional computation, maintaining image realism while reducing real-time complexity

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system copies pre-rendered ray-traced view state data instead of performing new ray tracing for each frame. The VSDO assets contain cached ray tracing results that can be efficiently combined with view transforms to generate new frames, eliminating redundant ray tracing computations while preserving the high image quality that ray tracing provides

Inventive Principle:
Principle #26Copying

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach significantly reduces the computational load and bandwidth needed for rendering multiple frames, allowing for efficient video compression and transmission, while maintaining high-quality rendering of realistic video sequences.

Implementation Method 1

Ray tracing involves determining the path taken by rays of light starting at a light source, which then pass through a hypothetical lens to a focal point. The process of following the light rays is called ray tracing.

Methodology Applied
Scientific EffectRay tracing:

Data Source

PatentUS10504276B2Re-utilization of render assets for video compression
Publication Date: 2019.12.10 URBACH JULIAN MICHAEL
  • US10504276B2 patent drawing
  • US10504276B2 patent drawing
  • US10504276B2 patent drawing

AI summary

Methods, apparatuses and systems directed to using viewport state data objects (VSDO) to render a series of video frames according to render instructions to achieve video compression. In a particular implementation, the video compression format exposes the VSDO and render instructions to a video render client, allowing the video render client to finish rendering a sequence of video frames from different spatial locations and view transform parameters. Implementations of this video compression format further allow any arbitrary client with enough graphics processing power and bandwidth to retransmit novel, completely rendered viewports to thin clients without taxing the root remote render devices.