Bandwidth Estimation for Video Encoding Using Running Averages
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies for video streaming face challenges in adapting to unpredictable network conditions, leading to inaccurate bandwidth estimation and resulting in reactive changes in video quality, which can cause unnecessary loss of quality, delay, or jitter due to short-term network fluctuations.
Innovation Solution
A system that employs a bandwidth estimation method using a combination of network monitors and machine learning techniques to calculate running averages of network conditions over different time periods, adjusting the estimated bandwidth based on these averages to stabilize video encoding quality and avoid drastic changes in response to transient network conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If bandwidth estimation responds quickly to network condition changes, then video quality adapts faster to network capacity, but short-term fluctuations cause unnecessary quality changes and jitter
Solution Approach 1:
The patent changes the time parameter of bandwidth estimation by calculating running averages over different time periods (short-term and long-term averages). This parameter transformation filters out short-term network fluctuations while preserving long-term bandwidth trends, resolving the contradiction between fast response and stability.
Solution Approach 2:
The patent introduces running average calculations as an intermediary between raw network conditions and video encoding decisions. This intermediary layer smooths out transient fluctuations before they reach the video encoder, preventing unnecessary quality changes while maintaining adaptive response to genuine bandwidth changes.
2Device complexity
If bandwidth estimation uses simple reactive methods, then implementation is straightforward, but inaccurate estimation leads to unnecessary loss of video quality
Solution Approach 1:
The patent segments the bandwidth estimation process into multiple independent components: short-term average calculation, long-term average calculation, and comparison logic. This segmentation maintains implementation simplicity while improving accuracy by considering multiple time scales separately and combining their insights.
3Productivity
If video encoding adapts to every network fluctuation, then bandwidth utilization is maximized, but transmission delays increase due to frequent quality changes
Solution Approach 1:
The patent implements periodic bandwidth estimation using running averages calculated over fixed time periods. This periodic approach with inherent smoothing reduces the frequency of encoding parameter changes, thereby reducing transmission delays while maintaining effective bandwidth utilization through systematic monitoring.
Data Source
AI summary
Techniques are generally described for remote estimation of bandwidth. In various examples, a video stream may be received at a first bit rate over a first communication channel. A first value of a network condition of the video stream may be determined over a first time period. A determination may be made that the first value is less than a threshold value. A first bandwidth estimate of the communication channel may be determined. The first bandwidth estimate may comprise the first bit rate reduced by a first percentage. A second value of the network condition may be determined over a second time period. A determination may be made that the second value is greater than the threshold value. A second bandwidth estimate of the communication channel may be determined. The second bandwidth estimate may be less than the first bandwidth estimate.