question

Diego Trazzi avatar image
Diego Trazzi asked ·

AMD GPUs - Denoising workflow for VFX studios with Arnold and Nuke

My goal is to create an optimized denoising workflow for AMD Graphics using Arnold (for Maya) and Nuke (possibly).

Premise:
* Arnold has 2 denoising apps: CPU and GPU (requires Optix - NVIDIA)
* Nuke is an industry standard, so 99% of studios would have Maya and Nuke.
* Nuke Denoiser can run on AMD GPUs (OpenCL)
* Running a denoiser on GPU is faster than CPU

Question:
I would like to understand if the denoiser in Nuke (which has OpenCL capabilities) could be utilized to efficiently replace the non-existing option provided with Arnold to run a denoising algorithm on AMD Graphics.
In other words, I would like to understand if the 'noice' app provided by Arnold equally works on 2D and 3D images, or it has some specific portion of the algorithms that utilizes 3D information derived by the Arnold renders, and if Nuke similarly denoises the renders with any notion of 3D or additional AOs that can be written as AOVs from the lighting renders.

Case study:
The Nuke denoiser could be part of the denoising workflow and lift this additional step from the lighting TDs to the Compers which could be running as part of their renders utilizing the GPU power.

The Nuke Denoiser takes a sample area of noise which can be fed as separate image so I think I could output a sample render frame from Arnold on a white card, similar to what we do on set when using a color checker card [I would create a 2D plane aligned with the camera with a white shader and using the same render setting generate an image containing the noise sample which I can then feed to Nuke].

How sound is this idea/approach ? Has anybody else used a similar approach ? What are your thoughts about this ? How much performance improvement could I get by running it on a GPU and would it make sense to run it on the final image, or per AOV?

gpunoisedenoisingnukemacos
10 |600 characters needed characters left characters exceeded

Up to 5 attachments (including images) can be used with a maximum of 2.0 MiB each and 9.8 MiB total.

1 Answer

· Write an Answer
Heribert Raab avatar image
Heribert Raab answered ·

the Nuke Denoiser is the most a terrible solution for renderings in my experience. the Intel Open Image Denoiser is better and much faster with CPU then Nuke Denoiser with a large GPU.

Share
10 |600 characters needed characters left characters exceeded

Up to 5 attachments (including images) can be used with a maximum of 2.0 MiB each and 9.8 MiB total.

Write an Answer

Hint: Notify or tag a user in this post by typing @username.

Up to 5 attachments (including images) can be used with a maximum of 2.0 MiB each and 9.8 MiB total.

Welcome to the Arnold Answers community.

This is the place for Arnold renderer users everywhere to ask and answer rendering questions, and share knowledge about using Arnold, Arnold plugins, workflows and developing tools with Arnold.

If you are a new user to Arnold Answers, please first check out our FAQ and User Guide for more information.

When posting questions, please be sure to select the appropriate Space for your Arnold plugin and include the plugin version you are using.

Please include images, scene and log files whenever possible as this helps the community answer your questions.

Instructions for generating full verbosity log files are available for MtoA, MaxtoA, C4DtoA, HtoA, KtoA, and Kick.

If you are looking for Arnold Documentation and Support please visit the Arnold Support site.

To try Arnold please visit the Arnold Trial page.

Bottom No panel present for this section.