You are viewing an old version of this page. View the current version.

Compare with Current View Page History

Version 1 Next »


The Arnold Denoiser is a stand-alone program post-process denoiser executable. It works on EXR files with multiple layers (merged) and outputs an EXR file with the denoised layers. It takes into account multiple frames and multiple light AOVs. Like most denoisers, it considers a neighborhood around the current pixel and looks for similar neighborhoods inside a given search area to blend together. To use the Arnold Denoiser, you will need to render images via the Arnold EXR driver.

  • Noice can work with the gaussian_filter or box_filter and sets up any additional filters it may need.
  • A denoising tutorial can be found here.


For the denoiser to work you first need to setup AOVs by using a combination of ArnoldOutputChannelDefine and RenderOutputDefine node pairs, where you set up the AOV in the former, creating a channel, and then the latter uses that channel to set up an actual render output.

Once they are defined add an ArnoldDenoise node where you can specify which outputs are to be denoised.  It creates the basic needed AOVs for you automatically (note: not render outputs, just "AOVs"), but you have to create additional outputs if you want them to be denoised or if you want them to be used as a denoiser: albedo, etc. 



The path to the image or image sequence to be denoised.


Set the denoised output file path. The output file will include the denoised layers with the same names as the original layers. For each file, a new one is saved, postfixed by '_denoised' (the denoised images).

Note that the file output size differs from the input files because only the main RGBA layers are saved (without the auxiliary AOVs).


The strength of the filter is determined by the variance parameter, the higher the variance the more forceful the denoising will be. For variance maybe 0.25 / 0.5 (the default) / 0.75 are good low/min/max values.
How aggressive noice is in removing noise can be controlled by setting a variance threshold with the command line argument -variance (-v). The default value is 0.5, higher values will make the denoising more aggressive by considering similar neighborhoods that have bigger color disparities.

Pixel Search Radius

This is the area over which similar neighborhoods are found. The higher the better, but it will increase the cost of denoising. For every pixel noice will search a square area with a radius set with the command line argument -searchradius (-sr). The bigger this area the bigger the denoising stability and the higher the chance that similar neighborhoods to be considered will be found. The default value is 9, which gives a 19x19 square neighborhood. Setting it to 21 (a search window of 42 x 42) will look over a pixel area equivalent to loading 5 frames. 

Pixel Patch Radius

This increases the softness of the denoising (while always preserving the features). The default is 3 (on the big side), but a low value would be maybe 0 or 1, middle 3 and high 5.

For every pixel noice will consider its neighborhood patch and look for other pixels with similar neighborhood patches. The radius of this neighborhood can be controlled with the -patchradius (or -pr) command line argument. The default value is set to 3, which gives a 7x7 square neighborhood.

Temporal Padding

The number of extra images to be used for temporal stability. Multiple additional frames can be specified to increase temporal stability (for instance render001.exr, render002.exr, render003.exr).

Light Group AOVs

Additional light AOVs to be co-denoised can be specified.


Define the number of threads to be used. The default is to leave one thread free, so it should not 'freeze' your computer while denoising.


  • No labels
Privacy settings / Do not sell my personal information / Privacy/Cookies