oidn-cmd

• Is aiming to be a simple command line utility for denosiing .exr sequences using Intel Open Image Denoise.
• For Linux appimage is provided, which should work out of the box.
• For Windows and MacOS building should via cargo should work, but I did not test it yet.

Limitations

• The denoised .exr file contains only the denoised layer, other layers from the original file are stripped.

Arguments

  -b, --beauty <beauty>            a beauty .exr file or sequence using the foo.####.exr pattern
  -a, --albedo <albedo>            a albedo .exr file or sequence using the foo.####.exr pattern
  -n, --normal <normal>            a normal .exr file or sequence using the foo.####.exr pattern
  -e, --layerbeauty <layerbeauty>  OPTIONAL - name of layer to denoise inside the beauty exr, defaults to main_layer
  -l, --layeralbedo <layeralbedo>  OPTIONAL - name of layer to denoise inside the albedo exr, defaults to main_layer
  -o, --layernormal <layernormal>  OPTIONAL - name of layer to denoise inside the normal exr, defaults to main_layer
  -x, --license                    Print license information
  -h, --help                       Print help
  -V, --version                    Print version

Running the .appimage

• download and make it executable
• run via commandline:

./oidn-cmd-0.1.0-x86_64.AppImage -b tests/beauty.####.exr -a tests/denoising_albedo.####.exr -n tests/denoising_normal.####.exr

Running & building via cargo

• clone
• download and extract OIDN into include/oidn-2.2.2.x86_64.linux/
• build / run:

cargo run --release -- -b tests/beauty.####.exr -a tests/denoising_albedo.####.exr -n tests/denoising_normal.####.exr

Building the Appimage (Linux only)

• same steps as for cargo
• download the appimage-builder and make it exectuable, I also tend to remove the .AppImage in the file name.
• you can put it to e.g. /home/your-user/.local/bin/ to have it available everywhere.
• cd into oidn-cmd directory and run appimage-builder

Acknowledgements

• Will Usher - creator of oidn-rs crate
• Johannes Vollmer - creator of exrs crate

License

This repository is licensed under MIT license.

The .appimage release contains unaltered files of Intel® Open Image Denoise release, which is licensed under Apache License Version 2.0,

see https://github.com/RenderKit/oidn/blob/master/LICENSE.txt for more information.