Video Player Ad Playout Prediction for Stalling Events
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Video-on-demand streaming experiences quality degradations such as rebuffering and stalling events due to network throughput fluctuations, which negatively impact user experience and energy efficiency, as advertisement placement is not optimized for network conditions during video streaming.
Innovation Solution
A method that predicts stalling events based on network quality metrics, allowing for the controlled playout of advertisement content to replace rebuffering indicators, thereby minimizing user annoyance and energy consumption by synchronizing advertisement display with predicted or detected stalling events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If advertisement content is played out at pre-defined locations during video streaming, then advertisement placement is optimized for profit maximization, but stalling events may occur sequentially with advertisement playout reducing Quality of Experience
Solution Approach 1:
The system dynamically adjusts advertisement playout timing based on real-time network conditions and predicted stalling events. Instead of fixed pre-defined locations, advertisement playout is adapted to occur during predicted stalling periods, transforming a static scheduling approach into a dynamic one that responds to changing network conditions.
Solution Approach 2:
The system converts harmful stalling events into beneficial advertisement delivery opportunities. By predicting stalling events and scheduling advertisement playout during these periods, the system transforms what would otherwise be pure negative user experience into a dual-purpose solution that maintains advertisement delivery while masking network issues.
2Reliability
If video player adapts bitrate to match network throughput to prevent rebuffering, then stalling events are reduced, but network throughput fluctuations still cause playback buffer underflow
Solution Approach 1:
The system performs preliminary prediction of stalling events using network quality metrics before they actually occur. By analyzing current network conditions and predicting future stalling periods, the system proactively prepares advertisement content for playout during these predicted periods, rather than reactively responding after stalling begins.
Solution Approach 2:
The system implements a feedback loop where network quality metrics are continuously monitored, stalling events are predicted based on this data, and advertisement playout timing is adjusted accordingly. This closed-loop control allows the system to adapt to changing network conditions and improve its predictions over time.
3Productivity
If advertisement content is downloaded and injected at pre-defined locations, then advertisement placement is controlled, but sequential advertisement and stalling events waste energy and extend session duration
Solution Approach 1:
The system dynamically optimizes advertisement playout timing to coincide with predicted stalling events, thereby reducing the need for sequential advertisement delivery. This dynamic scheduling consolidates advertisement playout into periods when the video player would otherwise be idle due to buffering, reducing overall session duration and energy consumption.
Data Source
AI summary
A method (100) of controlling playout of advertisement content during video-on-demand video streaming on an end-user client terminal comprising steps of: receiving (110) advertisement content from an advertisement server; receiving (112) video-on-demand, VoD, content from a content delivery network; obtaining (114) network quality metrics between the end-user terminal and the content delivery network; predicting (116) whether a stalling event will occur during playout of VoD content within a prediction time window based on the network quality metrics; and playing out (118) received VoD content and received advertisement content, wherein playout of advertisement content within the prediction time window is dependent on a result of the prediction.