Median Filter for Media Rate Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional adaption algorithms for streaming media in mobile environments are ineffective due to bursty packet loss and delay, leading to unnecessary reduction in media quality as they overreact to short bursts of packet loss and delay.
Innovation Solution
A system that applies a median filter to feedback data on packet events to adjust the encoding rate, reducing the impact of short bursts and maintaining media quality by focusing on more persistent packet loss, thereby smoothing the streaming experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If conventional adaption algorithms adjust encoding rate downward in response to packet loss and delay, then media display smoothness is improved, but media quality is reduced
Solution Approach 1:
A median filter is introduced as an intermediary component between the packet loss/delay detection and the encoding rate adjustment. The filter processes the feedback data about packet events, smoothing out transient bursts while preserving sustained degradation patterns, thereby mediating between the raw network conditions and the adaption algorithm's response.
Solution Approach 2:
The system changes the parameter used for detecting packet loss from raw count-based metrics to median-filtered percentage-based metrics. This parameter transformation allows the system to distinguish between transient bursts (which would cause unnecessary quality reduction) and sustained degradation (which warrants rate adjustment), resolving the contradiction between smoothness and quality.
2Manufacturing precision
If adaption algorithms respond to bursty packet loss in mobile environments, then media quality may be maintained, but the algorithms become overly sensitive to transient conditions
Solution Approach 1:
The median filter serves as a mediator that processes raw packet loss feedback before it reaches the adaption algorithm. By filtering out transient bursts through median calculation, the filter protects the algorithm from over-reacting to temporary conditions while still allowing response to genuine degradation, thus improving quality without excessive sensitivity.
Solution Approach 2:
The system performs preliminary processing of packet loss data through the median filter before the adaption algorithm makes its decision. This preliminary action smooths the feedback signal in advance, removing the sensitivity to transient conditions that would cause unnecessary quality reduction, while preserving the ability to respond to sustained problems.
Data Source
AI summary
In an exemplary embodiment of this disclosure, a computer-implemented method includes encoding a first portion of media at an encoding rate. The first portion of the media is streamed over a network from a sending device to a receiving device. Feedback is received related to packet events occurring during transmission of the first portion of the media over the network. Adjusted feedback is generated, by a computer processor, by applying a median filter to the feedback data. The encoding rate of the media is adjusted based on the adjusted feedback data. An additional portion of the media is encoded at the adjusted encoding rate.