OCIO config based noise threshold calculation

Status: pending

I’ve requested this years ago but I was told that this is nonsene. So, now that I have used it in Karma for a while (and yes, it works) I propose (again) to have noise variance be determined by the output ocio transform. This saves rendertimes in highlights and adds samples in shadow areas where the tonemapping can make a significant difference and the gamma based variance is not good enough. Attached is a screenshot of Karma’s implementation.


Submitted on AHA Ideas (VRAYMAX-I-52).

I’ve myself never claimed that this is nonsense; of course it can be done - right now the adaptive sampler is assuming sRGB transform of the output - but it can use any other color transformation. Let’s see how many votes this gets and we will look into it.


Submitted on AHA Ideas by Vladimir Koylazov.

Wasn’t implying it was you. Someone else from Chaos wasn’t exactly .. supportive of the idea.


Submitted on AHA Ideas.

the task is still locked btw, so nobody can vote on it.


Submitted on AHA Ideas.

Open for voting.


Submitted on AHA Ideas by Peter Matanov.

AHA Ideas status: Pending.
Set during AHA Ideas to Discourse migration.