Video Stream Splitting for VR Bandwidth Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for video data transmission in virtual reality applications, such as near-eye displays, face bandwidth and computing power limitations, leading to noticeable lag and visual artifacts when changing viewing angles, as they require large amounts of data to maintain seamless high-resolution images across all viewing angles.

Innovation Solution

The technique involves partitioning video images into regional cross sections based on user viewport, foveal vision, and scenes of interest, encoding these sections into multiple video sub-streams at varying spatiotemporal resolutions, allowing for efficient data transmission and seamless high-resolution rendering without lag.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If video data is transmitted at high resolution for all viewing angles, then visual quality is improved, but bandwidth requirements increase enormously

Engineering Contradiction:
Improvevisual qualityVSAvoidbandwidth requirements
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent segments the video content into multiple pyramid faces, where only the base pyramid face (corresponding to the current FOV) is transmitted at high resolution, while other faces are transmitted at lower resolutions or not at all. This segmentation allows the system to provide high visual quality for the active viewing area while significantly reducing overall bandwidth requirements.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by transmitting high-resolution video data only for the base pyramid face that corresponds to the user's current field of view, while other pyramid faces receive lower resolution data. This ensures that computational and bandwidth resources are concentrated on the visually critical regions, maintaining high visual quality where needed while reducing overall data transmission requirements.

Inventive Principle:
Principle #3Local quality

2Productivity

If pyramid partitioning is used to reduce bandwidth, then data transmission efficiency is improved, but lag increases when changing FOV to non-base faces

Engineering Contradiction:
Improvedata transmission efficiencyVSAvoidlag
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-computing and preparing video data for multiple pyramid faces, including potential future FOV directions. When the user changes their field of view, the system can quickly switch to the pre-prepared base pyramid face data, minimizing lag. This preliminary preparation ensures that high-resolution data is ready for rapid transmission when needed.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements dynamics by making the base pyramid face selection adaptive and dynamic based on the user's current and predicted field of view. The system continuously monitors FOV changes and dynamically adjusts which pyramid face becomes the base, ensuring that high-resolution data is always available for the active viewing area while maintaining efficient bandwidth utilization.

Inventive Principle:
Principle #15Dynamics

3Productivity

If pyramid partitioning is used to reduce bandwidth, then transmission efficiency is improved, but FOV may intersect two pyramid faces causing resolution loss

Engineering Contradiction:
Improvetransmission efficiencyVSAvoidresolution
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent applies partial or excessive action by transmitting slightly more data than strictly necessary for the current FOV. When the FOV intersects multiple pyramid faces, the system transmits high-resolution data for the entire base pyramid face area, which may extend beyond the strict boundaries of the current FOV. This excessive transmission ensures complete coverage and maintains resolution continuity, preventing the visual artifacts that would result from precise but incomplete face boundary adherence.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3793205B1Content based stream splitting of video data
Publication Date: 2023.09.13 DOLBY LABORATORIES LICENSING CORP
  • EP3793205B1 patent drawingFigure 1
  • EP3793205B1 patent drawingFigure 2A
  • EP3793205B1 patent drawingFigure 2B

AI summary

Scenes in video images are identified based on image content of the video images. Regional cross sections of the video images are determined based on the scenes in the video images. Image portions of the video images in the regional cross sections are encoded into multiple video sub-streams at multiple different spatiotemporal resolutions. An overall video stream that includes the multiple video sub-streams is transmitted to a streaming client device.