Denoise Isotropix Clarisse renders with Altus

Overview

Altus-Clarisse plugin integrates the denoiser directly into Clarisse as a whole image filter. This enables you to denoise automatically and see the result in the image viewer. The plugin uses render information to generate a high quality denoised image. This guide will explain how to setup and use the Altus-Clarisse plugin.

Generally, Altus uses AOVs to enhance denoise quality and retain extra details. AOVs provide a way to render any arbitrary shaded component (AOV) into different images. Typically renderers will only produce a final color for each pixel of the image, but you can break out renders into their component parts such as indirect lighting, diffuse color, reflections, shadows, mattes, etc. and save them as AOVs individually. The denoiser quality will improve with more AOVs given as inputs to Altus.

Install the Plugin

Altus-Clarisse plugin comes in a packaged zip with two libraries.

Copy the libraries to the ‘module’ folder inside the installation directory of Clarisse. For example: C:/Program Files/Isotropix/Clarisse iFX 3.6 SP8b/Clarisse/module/

InputAOVs

Recommended List of AOVs to use with Altus:

AOV type Altus Input Name Clarisse AOV Name
World Position pos coordinates
Bump Normals nrm world normal
Visiblity vis diffuse front
Albedo alb albedo
Reflection extra reflection
Caustics cau (Optional) Caustics

We recommend adding AOVs that contain important information. Ie, if you’re denoising skin you might add the sss AOV.

Using the Plugin

Now that the plugin can be found and loaded, lets add it to a project. Here is a scene with one camera, pathfinder, and one rendered image.

../../_images/step1.png

Altus Requires two renders of the same camera each with a different sample seed. This will give Altus two images with differeing noise patterns that it can use to preserve actual detail. We recommend rendering these two images at half the samples as Altus will internally combine all the information into the denoiser.

There are many ways to get a second image to render. Here is a quick setup guide:

Right click on the rendered image (“image1”) and click “Copy with Dependencies”

../../_images/step2.png

This will copy the image, along with the camera and path tracer. Now when you paste the files it will create a second image + dependencies.

../../_images/step3.png

Navigate to the pathtracer being used by the second image and change the sampling seed to a different number. You can find the pathtracer reference inside the layer of the second image.

../../_images/step4.png

This is also where you could lower the samples to half of the original, for both pathtracers, if you desired.

Now you can add Altus to the original rendered image. Go to the layer of the original image and add “Altus Denoiser” under the filters section.

../../_images/step5.png

Finally, we have to tell Altus where the second image is. Go to the Altus reference and click the button to browse the project to select the second image.

../../_images/step6.png

Now when you preview the image, Altus should run the denoiser automatically.

../../_images/result.png

Plugin Settings

Altus Inputs:
  • Second Image: The reference to the second rendered image.
  • AOV inputs: A list of AOVs to assist denoise quality. We recommend at least 4: pos, normals, albedo, shadows.
  • Use Preview Quality: Toggles preview/production quality. Preview is great for iterative workflows, production gives the best results.
Altus Filter Settings:
  • FilterWindow: The overall constraint to the filter. Larger numbers will increase quality at a cost of speed. The default of 10 is a good balance.
  • kc1,kc2,kc3: Color sensitivity at various frequences. (kc1 = high freq details, kc2 = mid details, kc3 = large details) Low numbers give more detail, higher numbers will aggressively denoise.
  • Firefly: Enables the firefly suppressor which reduces high intensity pixels.
Altus Performance:
  • Tiling: Toggles between denoising the entire image at once (buffer mode), or denoise each bucket as it finishes rendering (bucket mode).
  • Tile-Size: Sets the maximum tile size. Smaller is slower but uses less RAM/VRAM.