Adaptive 360 Video Streaming via GOP Boundary Switching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

High-quality 360-degree video content transmission is hindered by large file sizes and bandwidth restrictions, leading to lower-than-desired resolution and significant latency when switching viewpoints, which disrupts the immersive experience in virtual reality applications.

Innovation Solution

The implementation of a video mapping system that utilizes a cube format to divide video content into multiple tiles, each with varying resolutions and group of pictures (GOP) durations, allowing for adaptive streaming and reduced latency by switching between videos at their respective GOP boundaries, thereby minimizing the delay in changing viewpoints.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If high-resolution video content is provided for all viewpoints, then video quality is improved, but file size and bandwidth requirements increase significantly

Engineering Contradiction:
Improvevideo qualityVSAvoidfile size
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The video content is segmented into multiple viewports, each representing a different viewing direction. Instead of providing high-resolution content for all directions simultaneously, the system divides the 360-degree video into discrete segments (viewports) and transmits only the relevant segment at high resolution based on the user's current viewing direction, while other segments are transmitted at lower resolution or not at all.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system applies different quality levels to different regions of the video content based on their importance. The viewport corresponding to the user's current viewing direction is transmitted at high resolution, while other viewports are transmitted at lower resolution. This creates a local quality differentiation where only the relevant portion receives maximum quality treatment, optimizing bandwidth usage while maintaining perceived video quality.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If video switching is performed when viewpoint changes, then high-quality image in viewing direction is provided, but latency is caused by switching delay

Engineering Contradiction:
Improvevideo qualityVSAvoidswitching latency
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

Multiple viewport videos are pre-encoded and prepared in advance for different viewing directions. When the user changes their viewpoint, the system can immediately switch to the pre-prepared viewport that corresponds to the new viewing direction, eliminating the need for real-time encoding or processing delays. The preliminary preparation of multiple viewports enables instant switching without quality loss or latency.

Inventive Principle:
Principle #10Preliminary action

3Quantity of substance

If lower-resolution content is provided to reduce file size, then bandwidth requirements are reduced, but display quality degrades below display capabilities

Engineering Contradiction:
Improvebandwidth usageVSAvoiddisplay quality
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The system transmits video content at different quality levels for different viewports based on their relevance to the user's current viewing direction. The active viewport (corresponding to the user's viewing direction) is transmitted at high resolution matching display capabilities, while inactive viewports are transmitted at lower resolution. This local quality differentiation ensures that the user always receives high-quality content for their current view while minimizing overall bandwidth consumption.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS11394944B2Image processing device and system for outputting higher-resolution videos
Publication Date: 2022.07.19 SONY INTERACTIVE ENTERTAINMENT LLC
  • US11394944B2 patent drawing
  • US11394944B2 patent drawing
  • US11394944B2 patent drawing

AI summary

A video providing device for providing video content having a plurality of viewpoints, the device including a video storage unit operable to store a plurality of versions of a video for each of a plurality of viewpoints in the video content, each video corresponding to the same viewpoint being encoded with a different resolution and a different group of pictures duration, a viewpoint determining unit operable to determine a requested viewpoint in the video content, a video selection unit operable to select a video from amongst the plurality of versions corresponding to the requested viewpoint, the selected video having a next-occurring group of pictures boundary, a video transmitting unit operable to transmit the selected video to a video receiving device, and a video switching unit operable to, if the selected video is not the highest-resolution video corresponding to that viewpoint, perform successive switches to higher-resolution videos at the next-occurring group of pictures boundary in a higher-resolution video until a switch is performed to the highest-resolution video for the determined viewpoint, where the video transmitting unit is operable to transmit a switched video instead of the selected video.