Altus rendering denoising system Tweet

[Leonides02]

How come nobody has gotten a straight answer on which passes to use?

[MildMustard]

In my humble opinion, the denoise function should be from Otoy, just as Chaos Group has created one for Vray. It makes much better marketing sense this way too, as it would be a feature within Octane (regardless that is doesn't lie within core) and no need for another purchase to get quicker cleaner renders. Even though it isn't perfect, Vray does a supurb job! I use both Octane and Vray for Modo and it would be a cherry on the top if we had this natively in the program.

[mark0spasic]

How come nobody has gotten a straight answer on which passes to use?

Use this passes:
beauty - beauty
normals - geo.norm
position - position
albedo - diffuse filter
visibility - opacity
caustics - refraction filter

[Leonides02]

Weird, that gave me a black screen. :-\

[Goldorak]

In my humble opinion, the denoise function should be from Otoy, just as Chaos Group has created one for Vray. It makes much better marketing sense this way too, as it would be a feature within Octane (regardless that is doesn't lie within core) and no need for another purchase to get quicker cleaner renders. Even though it isn't perfect, Vray does a supurb job! I use both Octane and Vray for Modo and it would be a cherry on the top if we had this natively in the program.

I agree. See my other post on the 3.1 roadmap. Adaptive Sampling is coming, at the latest in 3.1, which will be non-destructive (and therefore higher quality then brute force denoising). We are also looking for feedback on Altus for users who have been using it with 3.0 - what works and doesn't. This would be helpful before we proceed with our own version (we have some novel ideas for denoising we are testing in any case).

[ff7darkcloud]

Got a black screen too. How can we test it when we don't know the passes to use?

[Goldorak]

Got a black screen too. How can we test it when we don't know the passes to use?

We had a separate team validate and test this using Altus' instructions and it worked as shown in GTC roadmap slides. I can ping them and see if they can help in the forums if you are still stuck.

[mark0spasic]

I don`t know why you have problems with Altus from my test
you can see I used following passes:

Render two different images with same max samples in this test I set max samples to 1000, turn off static noise also set coherent ratio to 0 save image to exr.
Keep in mind Altus uses GPU memory for processing, quote from their site:
"Can I process 4k images with the Altus?
Yes 4k image filtering is possible with Altus. However, there is a significant increase in the memory usage/requirement. Since all the images are uncompresed in memory, you will need at least 16GB of RAM available at the time of filtering for a single image assuming you are using every input. Currently this is possible on CPU and is not currently possible on a single GPU. It is theoretically possible on GPU using a high end SLI solution such as a Tesla cluster."

[Leonides02]

Ok, so it works for me only if I don't include the OPACITY.EXR

The required three (beauty, normals, position) works and so does albedo and caustics.

Opacity gives me a black image.

Can anyone else confirm?

[Leonides02]

Also, there's no way we can use this for animation, correct? I mean, I don't know of a way to export render passes inside the animation tab...