SVT-AV1-PSY

SVT-AV1-PSY is the Scalable Video Technology for AV1 (SVT-AV1 Encoder and Decoder) with perceptual enhancements for psychovisually optimal AV1 encoding. The goal is to create the best encoding implementation for perceptual quality with AV1.

Feature Additions

• --fgs-table Argument for providing a film grain table for synthetic film grain (similar to aomenc's --film-grain-table= argument).
• --variance-boost-strength Provides control over our augmented AQ mode 2 which can utilize variance information in each frame for more consistent quality under high/low contrast scenes. Five curve options are provided, and the default is curve 2. 1: mild, 2: gentle, 3: medium, 4: aggressive
• --new-variance-octile Enables a new 8x8-based variance algorithm and picks an 8x8 variance value per superblock to use as a boost. Lower values enable detecting more false negatives, at the expense of false positives (bitrate increase). There are four options. 0: disabled, use 64x64 variance algorithm instead 1: enabled, 1st octile 4: enabled, median 8: enabled, maximum. The default is 6.
• Preset -2 A terrifically slow encoding mode for research purposes.
• Tune 3 A new tune based on Tune 2 (SSIM) called SSIM with Subjective Quality Tuning. Generally harms metric performance in exchange for better visual fidelity.
• --sharpness A parameter for modifying loopfilter deblock sharpness and rate distortion to improve visual fidelity. The default is 0 (no sharpness).

Modified Defaults

SVT-AV1-PSY has different defaults than mainline SVT-AV1 in order to provide better visual fidelity out of the box. They include:

• Default 10-bit color depth. Might still produce 8-bit video when given an 8-bit input.
• Disable film grain denoising by default, as it often harms visual fidelity.
• Default to Tune 2 instead of Tune 1, as it reliably outperforms Tune 1 on most metrics.
• Enable quantization matrices by default.
• Set minimum QM level to 0 by default.

We are not in any way affiliated with the Alliance for Open Media or any upstream SVT-AV1 project contributors who have not also contributed here.

Scalable Video Technology for AV1 (SVT-AV1 Encoder and Decoder)

The Scalable Video Technology for AV1 (SVT-AV1 Encoder and Decoder) is an AV1-compliant software encoder/decoder library. The work on the SVT-AV1 encoder targets the development of a production-quality AV1-encoder with performance levels applicable to a wide range of applications, from premium VOD to real-time and live encoding/transcoding. The SVT-AV1 decoder implementation targets future codec research activities.

The SVT-AV1 project was initially founded by Intel in partnership with Netflix, and was then adopted by the Alliance of Open Media (AOM) Software Implementation Working Group (SIWG), in August 2020, to carry on the group's mission.

The canonical URL for this project is at https://gitlab.com/AOMediaCodec/SVT-AV1

License

Up to v0.8.7, SVT-AV1 is licensed under the BSD-2-clause license and the Alliance for Open Media Patent License 1.0. See LICENSE and PATENTS for details. Starting from v0.9, SVT-AV1 is licensed under the BSD-3-clause clear license and the Alliance for Open Media Patent License 1.0. See LICENSE and PATENTS for details.

Documentation

Guides

• System Requirements
• How to run SVT-AV1 within ffmpeg
• Standalone Encoder Usage
• Decoder Usage
• List of All Parameters
• Build Guide
• ARM Build Guide

Common Questions/Issues

• What presets do
• Scene change detection
• GOP size selection
• Threading and efficiency
• Practical advice about grain synthesis
• Improving decoding performance
• Tuning for animation
• 8 vs. 10-bit encoding
• HDR and SDR video
• Options that give the best encoding bang-for-buck
• Multi-pass encoding
• CBR, VBR, and CRF modes

Presentations

• Big Apple Video 2019
• Video @ Scale 2021

Papers and Blogs

• Netflix Blog 2020
• SPIE 2020
• SPIE 2021
• SVT-AV1 - Tech Blog 2022
• SPIE 2022
• Adaptive Steaming Common Test Conditions
• ICIP 2023

Design Documents

• Encoder Design
• Decoder Design

Technical Appendices

• Adaptive Prediction Structure Appendix
• Altref and Overlay Pictures Appendix
• CDEF Appendix
• CfL Appendix
• Compliant Subpel Interpolation Filter Search Appendix
• Compound Mode Prediction Appendix
• Deblocking Loop Filter (LF) Appendix
• Film Grain Synthesis
• Global Motion Appendix
• Intra Block Copy Appendix
• IPP Pass Appendix
• Local Warped Motion appendix
• Mode Decision Appendix
• Motion Estimation Appendix
• Overlapped Block Motion Compensation Appendix
• Palette Prediction Appendix
• Rate Control Appendix
• Recursive Intra Appendix
• Restoration Filter Appendix
• SQ Weight Appendix
• Super-resolution Appendix
• Temporal Dependency Model
• Transform Search Appendix
• Reference Scaling Appendix

How Can I Contribute?

• SVT-AV1 Contribution Guide