Chunk-Based ABR Logic for Low-Latency Streaming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive bitrate (ABR) schemes for HTTP chunked transfer encoding in CMAF-based low-latency streaming suffer from inaccurate throughput estimation, leading to sub-optimal bitrate selection and unsatisfactory quality of experience (QoE) due to the bursty nature of shorter chunks and the reliance on segment-based measurements.
Innovation Solution
A novel ABR scheme using a sliding window moving average throughput estimation algorithm and an online linear adaptive filter to estimate future bandwidth, allowing for chunk-based bitrate decisions that minimize stalls and maintain target latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If segment-based throughput estimation is used in existing ABR schemes, then the measurement process is simple, but the throughput estimation becomes inaccurate due to the bursty nature of shorter chunks
Solution Approach 1:
The patent segments the throughput estimation process into chunk-level measurements rather than segment-level measurements. By measuring throughput at the chunk level (smaller units), the system captures the bursty nature of data transmission more accurately. The sliding window mechanism further segments the measurement history into recent chunks, allowing the system to adapt to changing network conditions without being skewed by outdated segment-level averages.
Solution Approach 2:
The patent introduces a dynamic sliding window approach where the measurement window moves forward in time, continuously incorporating new chunk-level throughput measurements and excluding older ones. This dynamic mechanism allows the throughput estimation to adapt to changing network conditions in real-time, providing accurate measurements even when network bandwidth fluctuates during the streaming session.
2Reliability
If chunk-based bitrate decisions are implemented, then stalls are reduced and latency is maintained, but the complexity of bitrate selection increases
Solution Approach 1:
The patent implements a feedback mechanism where the adaptation logic continuously monitors actual chunk download throughput and uses this information to adjust bitrate decisions. The sliding window moving average provides a feedback loop that compares expected vs. actual throughput, allowing the system to detect stalls and network conditions in real-time and respond by adjusting the selected bitrate to maintain reliable playback.
Solution Approach 2:
The patent uses preliminary throughput measurements from recent chunks to predict future network conditions and make proactive bitrate selections. By analyzing the sliding window of past chunk throughput values, the system can anticipate potential stalls or bandwidth changes and adjust the bitrate before actual playback issues occur, maintaining reliability without waiting for problems to manifest.
3Measurement precision
If sliding window moving average throughput estimation is used, then throughput estimation accuracy improves, but computational complexity increases
Solution Approach 1:
The patent applies partial action by using a sliding window that processes only the most recent N chunks rather than analyzing the entire historical record. This partial approach focuses computational resources on the most relevant recent data while ignoring outdated information, achieving accurate throughput estimation with reduced computational complexity compared to processing all historical measurements.
Solution Approach 2:
The patent changes the parameter of measurement granularity from segment-level to chunk-level, and introduces a sliding window size parameter N that controls the trade-off between accuracy and computational complexity. By adjusting N, the system can optimize the balance between estimation accuracy and processing resources based on the specific application requirements and available computational power.
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
A multimedia player downloads chunks (parts of the segment file) during the download of a segment of a stream of segments of a low-latency stream. The first chunks of a segment may be downloaded while the segment is still being written to the CDN server. A chunk-based prediction adaptation logic uses throughput measurements on a chunk instead of a segment and specifically looks at bursts in a sliding window. This data is used to build a prediction of future throughput by applying linear adaptive filter algorithms that may rely on recursive least squares. This adaptation logic leads to very accurate bandwidth predictions and as a consequence, better user experience, compared to existing adaptation algorithms.