https://github.com/7gxycn08/pyanime4k-gui
Anime4K batch video upscaler using predefined settings for AMD GPU's.
https://github.com/7gxycn08/pyanime4k-gui
amd anime-upscaling anime4k convolutional-neural-networks denoising-algorithms esrgan ffmpeg python real-esrgan super-resolution upsampling upscaling video-processing
Last synced: 2 months ago
JSON representation
Anime4K batch video upscaler using predefined settings for AMD GPU's.
- Host: GitHub
- URL: https://github.com/7gxycn08/pyanime4k-gui
- Owner: 7gxycn08
- License: mit
- Created: 2024-12-07T17:18:43.000Z (11 months ago)
- Default Branch: main
- Last Pushed: 2025-02-10T19:13:04.000Z (9 months ago)
- Last Synced: 2025-03-27T15:55:54.921Z (7 months ago)
- Topics: amd, anime-upscaling, anime4k, convolutional-neural-networks, denoising-algorithms, esrgan, ffmpeg, python, real-esrgan, super-resolution, upsampling, upscaling, video-processing
- Language: GLSL
- Homepage:
- Size: 2.19 MB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
## PyAnime4K-GUI
PyAnime4K-GUI is a graphical user interface for upscaling video files using FFmpeg and [Anime4K](https://github.com/bloc97/Anime4K) custom shaders. It provides an easy way to manage video processing tasks with real-time progress updates and log outputs.
# Features
1. Batch Video Upscaling: Select multiple video files for upscaling which will be saved to disk.
2. Configurable Settings: Customize output `dimensions`, `bitrate`, `codec`, and `shaders` via a config file using a text editor.
3. Real-Time Progress: Monitor upscaling progress with a visual progress bar.
4. Log Viewer: View live progress in the GUI and ffmpeg stderr in output.txt file.
5. Cancel Operations: Cancel ongoing upscaling tasks at any time.
6. Output Folder Access: Quickly navigate to the output folder.
7. Multiple Subtitle Stream Copy: Includes all subtitle streams from input file.
8. Compare Two Videos Side-by-Side: Video compare function that display quality changes in real-time.
# Requirements
- Python 3.9+
- PySide6
- Opencv-Python
- FFmpeg with Vulkan support
- Additional Python libraries:
* tqdm
* ffmpeg-progress-yield
# Installation
1. Clone this repository:
```
git clone https://github.com/your-repo/PyAnime4K-GUI.git
cd PyAnime4K-GUI
```
2. Install dependencies:
```
pip install -r requirements.txt
```
3. Place the required ffmpeg.exe and ffprobe.exe in the `ffmpeg/` folder.
4. Ensure the `Resources/Config.ini` file contains your desired settings. You can open `Config.ini` directly from the GUI.:
```
[Settings]
width = 3840
height = 2160
bit_rate = 10M
preset = slow
codec = hevc_amf # for Nvidia GPU's use hevc_nvenc
shader = Anime4k-ModeA-UL.glsl
```
# Usage
1. Run the application:
```
python PyAnime4K.py
```
2. Use the GUI to:
* Edit Configuration: Open the Config.ini file for adjustments.
* Select Video Files: Choose videos to upscale.
* Start Upscaling: Begin processing with the "Upscale" button.
* Cancel Upscaling: Stop the current task with "Cancel."
* Access Output Folder: Open the output directory to view results.
3. stderr Logs are saved in `output.txt` and progress updates will appear in the application window.
# Custom Shaders
[Click Here for Shader Details](https://github.com/bloc97/Anime4K/blob/master/md/GLSL_Instructions_Advanced.md#modes)
Shaders for upscaling are located in the `shaders/` directory. Modify or add your shaders as needed and reference It in `Resources/Config.ini` file.
You can use your own custom shader combinations example:
`Anime4K-R+DTD+U+R+DTD+U.glsl` This combination takes 10 minutes to 2-pass upscale an episode to 4K on a `RX 7900xtx` gpu.
```
# Included shaders:
Anime4K_Restore_CNN_UL.glsl
Anime4K_Upscale_DTD_x2.glsl # Pass-1
Anime4K_Upscale_CNN_x2_UL.glsl
Anime4K_AutoDownscalePre_x2.glsl
Anime4K_AutoDownscalePre_x4.glsl
Anime4K_Restore_CNN_UL.glsl
Anime4K_Upscale_DTD_x2.glsl # Pass-2
Anime4K_Upscale_CNN_x2_UL.glsl
```
# License
This project is licensed under the MIT License. See the LICENSE file for details.