Chaos Corona Forum

Chaos Corona for 3ds Max => [Max] I need help! => Topic started by: alessandro deana on 2017-03-06, 17:09:08

Title: Indirect lighting issue
Post by: alessandro deana on 2017-03-06, 17:09:08

Hi Everybody,
I would like to know your opinion about this issue.
I have a lamp 3d model and I want to test its photometry while I move the light source.

The light source in not visible directly because it is inserted in its support but the light source (corona light) has all its visibility parameters turned on.
(https://dl.dropboxusercontent.com/u/36670899/sampei_model.jpg)

The light should reflect on the upper chrome dome and than should bounce on the ground illuminating the scene.
(https://dl.dropboxusercontent.com/u/36670899/sampei_light%20source.jpg)

I tested it in Maxwell render and it works fine, because it is totally unbiased. This is the result I want to achieve:
(https://dl.dropboxusercontent.com/u/36670899/sampei_maxwel.jpg)

Strangely I can't reproduce this behaviour in Corona, unless I set the light direction on both sides or I flip it toward the ground. This is the dark render:
(https://dl.dropboxusercontent.com/u/36670899/sampei_model2.jpg)

I am testing it both with a disk light and a sphere. Obviously the sphere light source produce some lights but it is still hidden by the support:
(https://dl.dropboxusercontent.com/u/36670899/sampei_model3.jpg)

I tried to turn on the caustic on the chrome material of the reflecting dome and I tried to set the MSI to 0, but it doesn't help.
I know I can flip the light toward the ground and adjust the directionality parameter in order to achieve the result but is is not physically accurate:
(https://dl.dropboxusercontent.com/u/36670899/sampei_corona.jpg)

Can somebody help me fixing this issue?

Many thanks

Title: Re: Indirect lighting issue
Post by: cecofuli on 2017-03-06, 18:29:39

You can, but only with Bidir/VCM.
Corona Team know this problem veeery well!
Corona, unfortunately, has a big problems with caustics

(http://www.francescolegrenzi.com/Temp/Corona/0139_Spot.jpg)

Title: Re: Indirect lighting issue
Post by: alessandro deana on 2017-03-06, 19:03:39

Yes, it works fine with BIDIR/VCT!
Great to know..

Many thanks.

Title: Re: Indirect lighting issue
Post by: cecofuli on 2017-03-07, 00:25:29

Non c'è problema ;-) Alla prossima

and, it's better than Maxwell =)

Title: Re: Indirect lighting issue
Post by: Romas Noreika on 2017-03-08, 16:08:29

What is BIDIR/VCT Where can I find it myself?

Title: Re: Indirect lighting issue
Post by: FrostKiwi on 2017-03-08, 16:30:15

Quote from: RomasNoreika on 2017-03-08, 16:08:29

What is BIDIR/VCT Where can I find it myself?

"Bidir" is the experimental Bi-directional pathtracer.
It is not meant for any use outside of experimental and otherwise impossible setups like these.

To enable it you have to enable the hidden "Development/Experimental Stuff" tab by checking the Checkbox in Render Setup->System->System Settings->enable Devel/Debug mode
Then under Performance-> "Development/Experimental Stuff" -> Render Engine -> Switch to Bidir/VCM.
The entry enables a list down below with 6 more render engines, where Bidir render engine is selected by default.

Title: Re: Indirect lighting issue
Post by: Romas Noreika on 2017-03-09, 10:02:53

thank you

Title: Re: Indirect lighting issue
Post by: maru on 2017-03-15, 17:36:50

Currently the solution would be to use separate object for the visible part of the lamp with light emission off (e.g. light bulb), and separate object for the light source that illuminates the scene (e.g. coronalight or max light with ies profile).
Fun fact is that you can bake IES light yourself. You can even use BIDIR to produce realistic result, and then save it as IES. Here is how:
https://forum.corona-renderer.com/index.php/topic,11630.msg75086.html#msg75086