Rethinking the Video Transcoding Pipeline: How Pre-Filtering Holds the Key to Efficiency and Quality
summary
Proposes a production-ready input pre-filter that combines pixel- and transform-domain methods with a perceptual thresholding module to preserve texture and creative intent while reducing bitrate. Reports objective BD-rate gains of 6–15%, a subjective expert-panel improvement of 6–12%, real-time deployment at about 18% of an x264 encoder's compute, and examples of ~1 Mbps per-stream bandwidth saving. Claims the filter is codec-agnostic, integrates with AWS Elemental workflows, and avoids the blocking/blur artifacts of prior pixel-domain filters.
AI-written summary · about
chapters (6)
- 0:00Video compression and AWS Elemental
- 2:37Pre-filtering improves compression
- 5:28Designing input filters
- 11:13Perceptual thresholding module
- 16:16Pre-filtering results and impact
- 18:25Conclusion
related talks
10mGoing Beyond Bitrate: Adaptation Based on Video Context
20mStreaming at 250 kbps: raise your expectations
26mOptimizing Video Performance for Web Developers
25mOptimal Video Compression Using Pixel Shift Tracking
26mNew video coding techniques under consideration for VP10 - the successor to VP9
26mBandwidth Ballet: Optimizing Video Performance for Web Maestros
22mProviding better video experiences for the next billion users.
11mFrom Artifacts to Art: A Data-Driven Guide to Android Video Encoding