pat-e/av1-encoding-scripts
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Replace with newest version from Github 2/2

Browse Source
This commit is contained in:
pat-e
2026-04-13 13:06:40 +02:00
parent c15aed1b87
commit 1d7b8b22fd
2 changed files with 893 additions and 873 deletions
Download Patch File Download Diff File Expand all files Collapse all files

216
README.md
View File

@@ -1,100 +1,116 @@
# AV1 Encoding Scripts # AV1 Encoding Scripts
## Overview ## Overview
This repository contains Python scripts for batch-processing MKV files to encode video to AV1 and audio to Opus. The scripts prioritize high-quality encoding, handle complex audio track downmixing, VFR (Variable Frame Rate) conversions, and offer features like automatic cropping and resumable encoding. This repository contains Python scripts for batch-processing MKV files to encode video to AV1 and audio to Opus. The scripts prioritize high-quality encoding, handle complex audio track downmixing, VFR (Variable Frame Rate) conversions, and offer features like automatic cropping and resumable encoding.
## Scripts Overview ## Scripts Overview
* **`aom_opus_encoder.py`**: Uses the `aom` encoder (specifically designed for the `aom-psy101` fork) via `av1an`. It is tuned for high perceptual quality with specific psychovisual parameters and film grain synthesis. * **`aom_opus_encoder.py`**: Uses the `aom` encoder (specifically designed for the `aom-psy101` fork) via `av1an`. It is tuned for high perceptual quality with specific psychovisual parameters and film grain synthesis.
* **`svt_opus_encoder.py`**: Uses the `svt-av1` encoder (specifically designed for the `SVT-AV1-Essential` fork) via `av1an`. It provides a good balance between encoding speed and quality, and allows customization of speed, quality, and film-grain presets from the command line. * **`svt_opus_encoder.py`**: Uses the `svt-av1` encoder (specifically designed for the `SVT-AV1-Essential` fork) via `av1an`. It provides a good balance between encoding speed and quality, and allows customization of preset, crf, and film-grain from the command line.
* **`hdr_svt_opus_encoder.py`**: A specialized script for 4K HDR movies using the `SVT-AV1-Essential` encoder. It is designed for pre-processed CFR inputs, preserves original surround sound audio without downmixing, and retains HDR metadata.
## Encoding Parameters Documentation
## Encoding Parameters Documentation
For detailed information on the specific FFmpeg arguments, audio downmixing logic, VFR-to-CFR conversion processes, and the special SVT-AV1/AomEnc parameters used by these scripts, please refer to the [`parameters.md`](parameters.md) file.
For detailed information on the specific FFmpeg arguments, audio downmixing logic, VFR-to-CFR conversion processes, and the special SVT-AV1/AomEnc parameters used by the standard scripts, please refer to the [`parameters.md`](parameters.md) file.
## Prerequisites
For the HDR-specific encoder parameters and settings used in `hdr_svt_opus_encoder.py`, please see [`parameters_hdr.md`](parameters_hdr.md).
Both scripts require several external tools to be installed and available in your system's `PATH`:
## Prerequisites
* **ffmpeg** & **ffprobe**: For video/audio extraction, filtering (cropdetect), and loudnorm analysis.
* **mkvtoolnix** (`mkvmerge`, `mkvpropedit`): For remuxing the final MKV file. Both scripts require several external tools to be installed and available in your system's `PATH`:
* **opusenc** (opus-tools): For encoding audio tracks to the Opus codec.
* **mediainfo**: For extracting detailed media information (especially frame rate details). * **ffmpeg** & **ffprobe**: For video/audio extraction, filtering (cropdetect), and loudnorm analysis.
* **av1an**: The core chunking encoder used to run multiple encode workers in parallel. * **mkvtoolnix** (`mkvmerge`, `mkvpropedit`): For remuxing the final MKV file.
* **HandBrakeCLI**: Used as a fallback/pre-processor to convert VFR (Variable Frame Rate) video to CFR (Constant Frame Rate) before the main encode. * **opusenc** (opus-tools): For encoding audio tracks to the Opus codec.
* **ffmsindex** (ffms2): For indexing the intermediate UTVideo file for Vapoursynth. * **mediainfo**: For extracting detailed media information (especially frame rate details).
* **Vapoursynth**: Required by `av1an` as the frame server via the generated `.vpy` scripts. * **av1an**: The core chunking encoder used to run multiple encode workers in parallel.
* *(Specific to `aom_opus_encoder.py`)*: **aom-psy101** encoder. You must download the correct version from [Damian101's aom-psy101 GitLab](https://gitlab.com/damian101/aom-psy101). * **HandBrakeCLI**: Used as a fallback/pre-processor to convert VFR (Variable Frame Rate) video to CFR (Constant Frame Rate) before the main encode.
* *(Specific to `svt_opus_encoder.py`)*: **SVT-AV1-Essential** encoder. You must download the correct version from [nekotrix's SVT-AV1-Essential GitHub](https://github.com/nekotrix/SVT-AV1-Essential/). * **ffmsindex** (ffms2): For indexing the intermediate UTVideo file for Vapoursynth.
* **Vapoursynth**: Required by `av1an` as the frame server via the generated `.vpy` scripts.
## Features * *(Specific to `aom_opus_encoder.py`)*: **aom-psy101** encoder. You must download the correct version from [Damian101's aom-psy101 GitLab](https://gitlab.com/damian101/aom-psy101).
* *(Specific to `svt_opus_encoder.py`)*: **SVT-AV1-Essential** encoder. You must download the correct version from [nekotrix's SVT-AV1-Essential GitHub](https://github.com/nekotrix/SVT-AV1-Essential/).
* **Automated Batch Processing**: Simply place your `.mkv` files in the same directory as the script. The script will process them one by one.
* **Resumable Encoding**: Because it uses `av1an`, if an encode is interrupted, you can restart the script, and it will resume from where it left off. ## Features
* **Audio Normalization and Downmixing**:
* Extracts audio tracks to FLAC. * **Automated Batch Processing**: Simply place your `.mkv` files in the same directory as the script. The script will process them one by one.
* Applies a 2-pass `loudnorm` normalization (Target: -23 LUFS, True Peak: -1 dB). * **Resumable Encoding**: Because it uses `av1an`, if an encode is interrupted, you can restart the script, and it will resume from where it left off.
* Downmixes 5.1/7.1 surround sound to stereo (unless `--no-downmix` is specified). * **Audio Normalization and Downmixing**:
* Encodes to Opus with bitrates automatically chosen based on the channel count (e.g., 128k for Stereo, 256k for 5.1). * Extracts audio tracks to FLAC.
* Directly remuxes existing `aac` or `opus` tracks without re-encoding. * Applies a 2-pass `loudnorm` normalization (Target: -23 LUFS, True Peak: -1 dB).
* Preserves track languages, titles, and delays. * Downmixes 5.1/7.1 surround sound to stereo (unless `--no-downmix` is specified).
* **VFR to CFR Conversion**: Detects Variable Frame Rate (VFR) media and automatically converts it to Constant Frame Rate (CFR) using HandBrakeCLI (virtually lossless `x264_10bit` CRF 0 intermediate) to prevent audio desync issues. * Encodes to Opus with bitrates automatically chosen based on the channel count (e.g., 128k for Stereo, 256k for 5.1).
* **Automatic Cropping**: Optional `--autocrop` flag detects black bars and determines the optimal cropping parameters before encoding. * Directly remuxes existing `aac` or `opus` tracks without re-encoding.
* **Organized Output**: * Preserves track languages, titles, and delays.
* Completed files are moved to a `completed/` directory. * **VFR to CFR Conversion**: Detects Variable Frame Rate (VFR) media and automatically converts it to Constant Frame Rate (CFR) using HandBrakeCLI (virtually lossless `x264_10bit` CRF 0 intermediate) to prevent audio desync issues.
* Original files are moved to an `original/` directory. * **Automatic Cropping**: Optional `--autocrop` flag detects black bars and determines the optimal cropping parameters before encoding.
* Per-file processing logs are saved in a `conv_logs/` directory. * **Organized Output**:
* Temporary files are automatically cleaned up upon success. * Completed files are moved to a `completed/` directory.
* Original files are moved to an `original/` directory.
## Usage * Per-file processing logs are saved in a `conv_logs/` directory.
* Temporary files are automatically cleaned up upon success.
Place the script in the directory containing your `.mkv` files and execute it.
## Usage
### `aom_opus_encoder.py`
It is highly recommended to place these scripts (or symbolic links to them) in a directory that is included in your system's `PATH` variable (e.g., `~/bin` on Linux/macOS, or a custom Scripts folder on Windows). This allows you to run the commands directly from any directory.
```bash
python aom_opus_encoder.py [options] To use the scripts, open your terminal (bash, PowerShell, etc.), navigate to the folder containing your `.mkv` files, and simply type the name of the script.
```
### `aom_opus_encoder.py`
**Options:**
* `--no-downmix`: Preserve original audio channel layout (do not downmix 5.1/7.1 to stereo). ```bash
* `--autocrop`: Automatically detect and crop black bars from the video. aom_opus_encoder.py [options]
* `--grain <int>`: Set the `photon-noise` value for grain synthesis (default: 8). ```
* `--crf <int>`: Set the constant quality level (`cq-level`) for video encoding (default: 28).
**Options:**
### `svt_opus_encoder.py` * `--no-downmix`: Preserve original audio channel layout (do not downmix 5.1/7.1 to stereo).
* `--autocrop`: Automatically detect and crop black bars from the video.
```bash * `--grain <int>`: Set the `photon-noise` value for grain synthesis (default: 8).
python svt_opus_encoder.py [options] * `--crf <int>`: Set the constant quality level (`cq-level`) for video encoding (default: 25).
```
### `svt_opus_encoder.py`
**Options:**
* `--no-downmix`: Preserve original audio channel layout (do not downmix 5.1/7.1 to stereo). ```bash
* `--autocrop`: Automatically detect and crop black bars from the video. svt_opus_encoder.py [options]
* `--speed <str>`: Set the SVT-AV1 encoding speed preset (e.g., `slower`, `slow`, `medium`, `fast`, `faster`). Defaults to `slower`. ```
* `--quality <str>`: Set the SVT-AV1 encoding quality preset (e.g., `lowest`, `low`, `medium`, `high`, `higher`). Defaults to `medium`.
* `--grain <int>`: Set the `film-grain` value. Adjusts the film grain synthesis level. Defaults to 6. **Options:**
* `--no-downmix`: Preserve original audio channel layout (do not downmix 5.1/7.1 to stereo).
## Process Workflow * `--autocrop`: Automatically detect and crop black bars from the video.
* `--preset <int>`: Set the SVT-AV1 encoding speed preset (e.g., 0-13). Lower is slower and yields better compression. Defaults to 1.
1. **Preparation**: Scans for `.mkv` files and checks for required tools. * `--crf <int>`: Set the SVT-AV1 Constant Rate Factor (CRF) for video quality (e.g., 0-63). Lower is better quality. Defaults to 30.
2. **Analysis**: Examines video and audio tracks using `ffprobe` and `mediainfo`. * `--grain <int>`: Set the `film-grain` value. Adjusts the film grain synthesis level. Defaults to 6.
3. **Video Processing**:
* Runs crop detection (if `--autocrop` is enabled). ### `hdr_svt_opus_encoder.py`
* Converts VFR to CFR (if VFR is detected).
* Extracts an intermediate lossless video (`utvideo`). ```bash
* Encodes the video using `av1an`. hdr_svt_opus_encoder.py [options]
4. **Audio Processing**: ```
* Remuxes AAC/Opus.
* Normalizes, downmixes (if applicable), and encodes other formats to Opus. **Options:**
5. **Muxing**: Combines the newly encoded video and audio tracks using `mkvmerge`, preserving synchronization delays, metadata, and languages. * `--preset <int>`: Set the SVT-AV1 encoding speed preset (e.g., 0-13). Lower is slower and yields better compression. Defaults to 1.
6. **Cleanup**: Moves files to respective folders (`completed/`, `original/`) and deletes temporary working files. * `--crf <int>`: Set the SVT-AV1 Constant Rate Factor (CRF) for video quality (e.g., 0-63). Lower is better quality. Defaults to 30.
* `--grain <int>`: Set the `film-grain` value. Adjusts the film grain synthesis level. Defaults to 12.
## Notes
## Process Workflow
- Encoding AV1 takes a significant amount of time and CPU resources.
- Ensure you have sufficient disk space, as the scripts generate intermediate lossless `utvideo` files which can be very large depending on the length and resolution of the source media. 1. **Preparation**: Scans for `.mkv` files and checks for required tools.
2. **Analysis**: Examines video and audio tracks using `ffprobe` and `mediainfo`.
## License 3. **Video Processing**:
* Runs crop detection (if `--autocrop` is enabled).
This project is licensed under the MIT License - see the [`LICENSE.md`](LICENSE.md) file for details. * Converts VFR to CFR (if VFR is detected).
* Extracts an intermediate lossless video (`utvideo`).
* Encodes the video using `av1an`.
4. **Audio Processing**:
* Remuxes AAC/Opus.
* Normalizes, downmixes (if applicable), and encodes other formats to Opus.
5. **Muxing**: Combines the newly encoded video and audio tracks using `mkvmerge`, preserving synchronization delays, metadata, and languages.
6. **Cleanup**: Moves files to respective folders (`completed/`, `original/`) and deletes temporary working files.
## Notes
- Encoding AV1 takes a significant amount of time and CPU resources.
- Ensure you have sufficient disk space, as the scripts generate intermediate lossless `utvideo` files which can be very large depending on the length and resolution of the source media.
## License
This project is licensed under the MIT License - see the [`LICENSE.md`](LICENSE.md) file for details.

1550
svt_opus_encoder.py
View File

File diff suppressed because it is too large Load Diff