From dc9fff182a9a436f7a83b95a3d8caa70b3da1780 Mon Sep 17 00:00:00 2001 From: Simon Stucki Date: Sat, 24 Aug 2024 10:40:47 +0200 Subject: [PATCH] Extend denoising documentation --- .../advanced/{denoising.md => denoising.mdx} | 21 +++++++++++++++++++ 1 file changed, 21 insertions(+) rename website/docs/advanced/{denoising.md => denoising.mdx} (69%) diff --git a/website/docs/advanced/denoising.md b/website/docs/advanced/denoising.mdx similarity index 69% rename from website/docs/advanced/denoising.md rename to website/docs/advanced/denoising.mdx index f2e4ce2..c06156e 100644 --- a/website/docs/advanced/denoising.md +++ b/website/docs/advanced/denoising.mdx @@ -1,3 +1,5 @@ +import TracerExperiment from "@site/src/components/TracerExperiment"; + # Denoising Denoising can be used as a final step to reduce undesired artifacts in the rendering. @@ -29,6 +31,25 @@ runPathTracer(target, model, { }); ``` +- `sigma` is the standard deviation of the Gaussian kernel. +- `kSigma` is the kernel size multiplier. +- `threshold` is the edge sharpening threshold for the denoising. + + + +:::note +The example does few samples to emphasize the denoising effect. +::: + ### OIDN [Open Image Denoise](https://github.com/RenderKit/oidn) is a denoise library which uses machine learning techniques to reduce noise. To configure it, you must provide a URL to a weight file. One option is: [oidn-weights on GitHub.com](https://github.com/RenderKit/oidn-weights/blob/master/rt_hdr_alb_nrm.tza).