A collection of VapourSynth wrappers and utility functions focused on advanced denoising, masking, and edge fixing.
Follow the docstring for more information about the parameters.
You can install vsdirty via pip:
pip install vsdirtyIf you want to enable GPU acceleration for neural networks, install the package with the appropriate extra based on your hardware:
- NVIDIA (TensorRT):
pip install vsdirty[nvidia]
- AMD (DirectML / ORT):
pip install vsdirty[amd]
- OpenCL (Generic):
pip install vsdirty[cl]
Or build from source:
git clone https://github.com/PingWer/vs-dirty
cd vs-dirty
pip install .[nvidia] # or .[amd], .[cl]This package relies on several external VapourSynth plugins. Most of them are automatically downloaded via pip, but ensure they are available in your VapourSynth environment:
| Plugin | Used For |
|---|---|
| vapoursynth-edgemasks | High-performance convolution masks (Kroon, Sobel, Kirsch, Prewitt) |
| vapoursynth-cas | Contrast Adaptive Sharpening |
| vapoursynth-mvtools | Motion interpolation and analysis |
| vapoursynth-vszipcu | Primary NLM and BM3D implementation (CUDA) |
| vapoursynth-vszipcl | Primary NLM and BM3D implementation (OpenCL) |
| vapoursynth-bm3d | BM3D implementation fallback (CUDA / HIP) |
| vapoursynth-bm3dcpu | BM3D implementation fallback (CPU) |
| vapoursynth-nlm-ispc | NLM implementation fallback (CPU) |
| vapoursynth-akarin | Fast mathematical expression evaluation (Expr) |
| libbore | Fix frame edges (dirtyfixer.bore) |
Note: Currently, libbore is not available via pip. You will need to download and install it manually from its GitHub repository.
MIT License