The AV1 Abstract
The abstract breaks down the AOMedia Video 1 (AV1) codec, the open, royalty-free standard designed for the next generation of video streaming. This application breaks down the key technologies that give AV1 its superior compression efficiency.
Core Architecture & Concepts
AV1 builds upon a hybrid block-based coding framework, but enhances it with more flexible structures and parallelism. This section introduces the fundamental building blocks of the codec.
AV1 Encoding Pipeline
Input Frame
The process starts with a single video frame.
Partitioning
The frame is divided into large Superblocks (up to 128x128), which are recursively partitioned.
Prediction
Each block is predicted using information from the same frame (Intra) or other frames (Inter).
Transform & Quantize
The prediction error (residual) is transformed to frequency domain and its precision is reduced.
Entropy Coding
The quantized values and metadata are losslessly compressed into the final bitstream.
Advanced Prediction Techniques
Prediction is key to reducing redundancy. AV1 introduces a vast toolkit of new prediction methods to handle diverse content types, from natural video to screen captures. Select a category below to explore the specific tools.
Directional Modes
AV1 expands from VP9's 8 directional modes to 56 modes, including finer angles. This allows for much more accurate prediction of edges and textures, reducing the residual that needs to be coded.
Chroma from Luma (CfL)
A powerful chroma-only predictor. It leverages the high correlation between luma (brightness) and chroma (color) by modeling chroma pixels as a linear function of the reconstructed luma pixels in the same block.
Intra Block Copy (IBC)
Extremely effective for screen content (text, graphics). It allows a block to be predicted by copying a previously decoded block from within the same frame, essentially performing motion compensation locally.
Palette Mode
Designed for content with few colors, like animations or screen sharing. It identifies a small palette of unique colors within a block and then encodes each pixel by its palette index, achieving massive compression for such content.
Smooth & Paeth Predictors
AV1 introduces new smooth prediction modes for handling gradients and replaces VP9's TrueMotion with the Paeth predictor, which excels at predicting corners and sharp edges by selecting the best of three neighboring pixel values.
Recursive Filtering
The FILTER_INTRA modes treat a block as a 2D Markov process, using pre-designed 7-tap filters on small patches to generate a highly accurate prediction, especially for textured areas.
Efficient Coding & Quality Enhancement
After prediction, AV1 uses advanced transform, quantization, and filtering techniques to compress the residual data and clean up the final image, maximizing both efficiency and visual quality.
Transform & Quantization
- Flexible Transforms: Supports both square and rectangular transform shapes (e.g., 64x16, 8x32) up to 64x64, adapting better to object shapes.
- Multiple Kernels: Uses a rich set of transform kernels (DCT, ADST, Flip-ADST, Identity) to better compact energy for different signal types.
- Quantization Matrices (QMS): Applies different quantization levels to different frequencies based on human perception, preserving important details while saving bits.
In-Loop Filtering Pipeline
1. Deblocking Filter (DBF)
Smooths sharp artifacts across transform block boundaries.
2. Constrained Directional Enhancement (CDEF)
A powerful deringing filter that removes "mosquito noise" around sharp edges by identifying edge directions and filtering along them.
3. Loop Restoration Filter (LRF)
The final step, using Wiener and Self-Guided filters to restore detail and reverse some of the damage caused by compression.
Performance & Codec Comparison
The combination of these advanced techniques results in significant compression gains over previous generations. The chart below shows typical bitrate savings for equivalent visual quality, according to industry benchmarks.
AV1 Bitrate Savings vs. Predecessors
Feature Showdown: AV1 vs. HEVC vs. H.264
| Feature | H.264 (AVC) | H.265 (HEVC) | AV1 |
|---|---|---|---|
| Max Block Size | 16x16 | 64x64 | 128x128 |
| Intra Prediction Modes | 9 | 35 | 56+ |
| Motion Compensation | Translational | Advanced | Warped, Affine, OBMC |
| In-Loop Filters | Deblocking | Deblocking, SAO | Deblocking, CDEF, LRF |
| Screen Content Tools | None | Limited | Intra Block Copy, Palette |
| Royalty Status | Complex Licensing | Complex Licensing | Royalty-Free |