Omnidirectional Video Tiling for Accurate ABR Bitrate Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive bitrate (ABR) algorithms for omnidirectional video streaming struggle to accurately estimate a single, all-inclusive bitrate that accounts for both viewport and non-viewport tiles, leading to suboptimal quality adaptation in varying network conditions.
Innovation Solution
A method and apparatus for determining an estimated picture bitrate based on tile stream bitrates and quality levels, considering viewport and non-viewport tiles, to facilitate adaptive bitrate control in omnidirectional video streaming.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If adaptive bitrate algorithms use tile stream bitrates separately for viewport and non-viewport tiles, then quality adaptation can be performed for different regions, but accurate estimation of a single all-inclusive bitrate becomes difficult
Solution Approach 1:
The patent merges the separate bitrate information of viewport tiles and non-viewport tiles into a single all-inclusive picture bitrate estimation. This is achieved by summing the bitrates of all tiles at different quality levels to create a comprehensive bitrate value that represents the entire omnidirectional picture, enabling accurate ABR algorithm decision-making.
Solution Approach 2:
The patent creates a universal bitrate estimation method that serves multiple functions: it provides the total picture bitrate for ABR algorithms, maintains separate tile bitrate information for regional quality control, and enables both overall quality adaptation and localized quality management in omnidirectional video streaming.
2Manufacturing precision
If separate bitrate control is applied to viewport and non-viewport tiles, then regional quality optimization is achieved, but the complexity of bitrate management increases
Solution Approach 1:
The patent segments the omnidirectional picture into viewport tiles and non-viewport tiles, allowing independent bitrate control for each region. This segmentation enables precise quality management where viewport tiles receive higher quality encoding while non-viewport tiles use lower quality, optimizing overall streaming efficiency.
Solution Approach 2:
The patent applies local quality control by assigning different quality levels and bitrates to different tile regions based on their importance. Viewport tiles, which are currently visible to the user, receive higher quality treatment, while non-viewport tiles receive lower quality, creating a localized quality optimization strategy.
3Measurement precision
If a single all-inclusive bitrate is estimated, then ABR algorithms can make accurate quality selections, but the ability to optimize individual tile qualities is reduced
Solution Approach 1:
The patent creates a universal bitrate estimation method that serves multiple functions: it provides the total picture bitrate for ABR algorithms, maintains separate tile bitrate information for regional quality control, and enables both overall quality adaptation and localized quality management in omnidirectional video streaming.
Solution Approach 2:
The patent merges the separate bitrate information of viewport tiles and non-viewport tiles into a single all-inclusive picture bitrate estimation while preserving the individual tile bitrate data, enabling both aggregate and granular quality control simultaneously.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There are disclosed various methods, apparatuses and computer program products for adaptive bitrate control for tiled omnidirectional video. In some embodiments information of a size of a viewport to view omnidirectional video data is received and an orientation of a viewport within the omnidirectional video data is determined. A first set of tiles belonging to the viewport and a second set of tiles outside the viewport are determined. A bitrate is estimated for a full picture using a first quality for the first set, and a second quality for the second set. The determination is repeated for a plurality of viewports. A representative viewport orientation is selected using the estimation results. A plurality of bitrates are estimated for a full picture for the third set of tiles and for the fourth set of tiles; and estimated bitrates and the representative viewport orientation are provided to an adaptive bitrate algorithm.