OTOY Forums

Hmmm.
Can you tell which is done with AS on?
Noise threshold .08
Group pix 2x2 (what difference does this make?)
Expected exposure 1.0 (what difference does this make?)
Min adaptive 20
Both cooked for 6 minutes at 3500x2625

Am I missing something here?

adaptive - gives you the speed but it is only needed to cut a clean render and not waste time cores graphics card to render - not to seek in it the function of Denoiser !!!!! adaptive rendering is not quite working well with the light he needed to focus on the noise on the highlights is personally my thoughts

...............
!!!!!!!!!!!!!
the second number of passages from 4000 and more !!!!

sorry but I can't. TEST ==== fixed hang !!!! down 4.30

you have given us to play with a broken toy

this is a simple example why the need for adaptiv

Hi Rik,
AS works by concentrating render cleaning where the noise is.
In potentially high noise scenes, AS reduce calculation times.
Regards
Paride

Goldisart wrote:sorry but I can't. TEST ==== fixed hang !!!! down 4.30

2017-03-29_14-16-15.jpg

you have given us to play with a broken toy

Goldisart,
what do you mean?
Regards
Paride

When you say Phoenix FD 3 support, does that mean Foam and Splashes now get rendered as well?? (please say yes! )

Thanks Paride.

I honestly appreciated the efforts you're putting in but...

I'm currently making multiple adjustments to the 4 variables and having to produce many, many tests to try and check that the grey parts are turning green quickly so Octane can work on the bits I know are going to still be noisy.

Its a real balancing act/guessing game though. For instance, if I go too high with the noise threshold it does not spend long enough on areas which are still noisy and will turn them green when in fact they're still noisy.
Too low with the noise threshold and it spends too long on bits with no noise.

Guess my point is that if its going to take 2 hours to get the 4 variables right for each image, then what's the point of it?

I'm running the tests at 3500x2625 so it's taking around 3 minutes each time to see what effect the tiny adjustments I'm making are having.

Please tell me I'm doing it wrong and there's an easier way?

Would be better if we have a "NOISE LEVEL" on render STATS, right? (even without using AS), so that way we can compare even better the optimization, improvements, setups. What do you think?

Hi guys,
this is my way to test AS to fine-tune the render:

Pixel grouping: 2x2 (unless you have very detailed texture materials like satin finishes, micro bumb etc, keep it 2x2
Noise threshold: 0,1
Min. adaptive samples: 1024
Expected exposure: leave it at 1 for now
if the render has no noise, then reduce max. samples until it is possible.
if the render has a little noise then increase max. samples.
if the image is too noisy then:

Noise threshold: 0,05
Min. adaptive samples: 512
if the render has no noise, then reduce max. samples until it is possible.
if the render has a little noise then increase max. samples.
if the image is too noisy then:

Noise threshold: 0,025
Min. adaptive samples: 256
if the render has no noise, then reduce max. samples until it is possible.
if the render has a little noise then increase max. samples.
if the image is too noisy then:

Noise threshold: 0,0125
Min. adaptive samples: 128
if the render has no noise, then reduce max. samples until it is possible.

if it is necessary, you can use intermediate values of Noise threshold and Min. adaptive samples to better define the right setting.
Regards
Paride