Per-scene encoding, also known as Content-Aware Encoding (CAE), is a revolutionary technique that adjusts encoder settings dynamically for each segment of a video. This ensures optimal compression while maintaining high visual quality. By analyzing content in real-time, this method tailors encoding parameters to suit the complexity and motion of each scene. The result? File size reductions of up to 70% without compromising quality.
VisualOn’s Optimizer: Redefining Transcoding with Per-Scene Encoding
VisualOn’s Optimizer leads the way in per-scene transcoding, providing an efficient solution that analyzes content to determine the best transcoder settings for achieving target quality. This cutting-edge solution integrates seamlessly with existing delivery systems, eliminating the need for changes to transcoders or workflows. Learn more about VisualOn Optimizer here.
Benchmarking with Axinom
In a recent study conducted by Axinom, the open-source short film “Tears of Steel” was used to demonstrate the power of per-scene encoding. This 12-minute video, containing diverse scenes with varying levels of complexity and motion, served as an ideal test case.
Axinom compared the video encoded without optimization to videos encoded using per-scene encoding at different target Video Multimethod Assessment Fusion (VMAF) scores. VMAF, developed by Netflix, is a reliable metric for predicting video quality. Read Axinom’s full article here.
Results
The test results were compelling:
Encoding Variant
Target VMAF
File Size (MB)
Percentage Saving
Original
–
570.0
–
Per-Scene
99
455.0
20%
Per-Scene
98
296.5
48%
Per-Scene
95
187.6
67%
Encoding with a target VMAF of 95 resulted in an impressive 67% reduction in file size while maintaining high visual quality.
Ensuring Quality
The optimized videos consistently delivered high VMAF scores, demonstrating that file size reduction didn’t come at the cost of quality. In some cases, the optimized videos even outperformed the original unoptimized video in terms of visual fidelity.
Conclusion
Per-scene encoding, as implemented by VisualOn’s Optimizer, is a game-changer for the video streaming industry. By applying customized encoding settings to individual scenes, content providers can achieve significant bandwidth savings, enhanced streaming efficiency, and an exceptional viewing experience for audiences worldwide.
To explore the technical details behind this methodology, visit Axinom’s original article: Per-Scene Encoding: A Comparison.
