New video coding techniques under consideration for VP10 - the successor to VP9
summary
Project targets a 30–40% bit‑rate drop over VP9 but current baseline additions already yield modest gains: ~6% reported against VP9 on a large set, with experimental results showing ~8% for 8‑bit VP10, ~12% for 12‑bit, ~12–14% at medium resolutions and ~20% on screen content. Work builds high‑bit‑depth buffers, 128×128 base blocks, T‑shaped and wedge partitioning, five‑to‑six reference frames, palette coding, non‑transform block modes, super‑transforms and global motion; prototypes exist and early tests show measurable but incomplete progress toward the 30–40% goal.
AI-written summary · about
chapters (8)
- 0:00Intro
- 0:38What is VP10
- 7:07New features
- 10:35Improved inter prediction
- 14:30Global motion prediction
- 18:59Filtering
- 22:47Nontransform
- 24:40Results
related talks
10mAdventures with VP9
29mHow to build a high quality 'live' HD VP9 Encoder
31mAV1 Update
17mDelivering better video encodes for legacy devices
19mRethinking the Video Transcoding Pipeline: How Pre-Filtering Holds the Key to Efficiency and Quality
20mBoosting Video Codec With RISC-V Vector Extension
11mFrom Artifacts to Art: A Data-Driven Guide to Android Video Encoding
30mA paradigm shift in codec standards: MPEG-5 Part 2 LCEVC, higher quality with any...