Isotropix Forums

Questions about the incomprehensibility of glass material

General Discussion about Isotropix and CG related topics

Questions about the incomprehensibility of glass material

Unread postby flansrjf » Sat Sep 11, 2021 8:10 am

Here are two pictures, one without glass and the other with glass. I found that even though the transmission is already the highest, it still cuts the bright light, and the indoor picture becomes darker.
What is the reason for this?
Attachments
Snipaste_2021-09-11_15-04-23.jpg
Glass material settings
Snipaste_2021-09-11_15-05-09.jpg
With glass
Snipaste_2021-09-11_15-01-01.jpg
No glass
flansrjf
 
Posts: 59
Joined: Sat Dec 22, 2018 1:56 am

Re: Questions about the incomprehensibility of glass materia

Unread postby sam » Sat Sep 11, 2021 8:46 am

Sam Assadian
Isotropix
CEO/Founder
User avatar
sam
 
Posts: 1711
Joined: Fri Jan 25, 2013 11:33 pm

Re: Questions about the incomprehensibility of glass materia

Unread postby flansrjf » Sat Sep 11, 2021 4:16 pm

Thanks for reply
I watched this video when it was released. The problem now is that the scene seen through the glass is obviously darkened. This seems to have nothing to do with the three modes introduced in the video. I have tried each of them.
flansrjf
 
Posts: 59
Joined: Sat Dec 22, 2018 1:56 am

Re: Questions about the incomprehensibility of glass materia

Unread postby Watchyato » Wed Sep 22, 2021 4:58 pm

We are also experiencing this exact issue, seem to get the same issue with autodesk shader, standard and dielectric. Changing path depths glossy/specular doesnt seem to have much effect (other than value being less than 2 which is understandable!). We did find that increasing the Clamping threshold helps to aleviate the issue but i dont see why this should be the work, afaik clapming threshold isnt linked to transmission?

Is this the intended effect of transmissive materials, or possibly a bug?

Thanks

Olly
Watchyato
 
Posts: 9
Joined: Wed May 05, 2021 11:40 am

Re: Questions about the incomprehensibility of glass materia

Unread postby aaute » Thu Sep 23, 2021 8:40 am

Hi,

Do you a project, reproducing what you experiencing ?
I try on my side and didn't see the difference you are observing.

Cheers
Aymeric
Isotropix
Technical Artist - Clarisse Specialist
User avatar
aaute
 
Posts: 90
Joined: Mon Nov 02, 2020 11:59 am

Re: Questions about the incomprehensibility of glass materia

Unread postby Watchyato » Thu Sep 23, 2021 11:28 am

Heres a simple scene setup with images as well, if you flip between the two images you can see a difference in the white values with glass (darker) without glass (lighter) in this example im using the autodesk shader, but we get the same results with dielectric and standard, even disney. I would note, as I realise I havnt already, we are using Clarrise 5 SP1, we are currently adding sp4 to see if this resolves the issue.
Attachments
glass_issue_with_glass.png
with glass
glass_issue_with_glass.png (139.2 KiB) Viewed 2101 times
glass_issue_without_glass.png
without_glass
glass_issue_without_glass.png (140.25 KiB) Viewed 2101 times
glass_issue.zip
(274.9 KiB) Downloaded 10 times
Watchyato
 
Posts: 9
Joined: Wed May 05, 2021 11:40 am

Re: Questions about the incomprehensibility of glass materia

Unread postby flansrjf » Thu Sep 23, 2021 4:54 pm

I may have found the reason: the default clamping value is 10, and after I changed him to 100, the brightness inside the glass was significantly improved. I also saw that the help documentation did mention this
flansrjf
 
Posts: 59
Joined: Sat Dec 22, 2018 1:56 am

Re: Questions about the incomprehensibility of glass materia

Unread postby flansrjf » Thu Sep 23, 2021 5:00 pm

I may have found the reason: the default clamping value is 10, and after I changed him to 100, the brightness inside the glass was significantly improved. I also saw that the help documentation did mention this
Attachments
Snipaste_2021-09-23_23-56-38.png
flansrjf
 
Posts: 59
Joined: Sat Dec 22, 2018 1:56 am

Re: Questions about the incomprehensibility of glass materia

Unread postby aaute » Thu Sep 23, 2021 5:08 pm

Hi,

I investigated your project.
It's a normal behavior because you have an IOR of 1.52 and rays are deviated on the glass as well as because of energy conservation some rays should be reflected (hence a slight loss of energy).
If you put an IOR of 1 you get the same result.

There is also the fact you need 2 more bounces to go through the glass. If you increase all the path depth with higher number you should have closer result but still with 2 bounces difference between you render with and without glass.

Removing the Clamping can be also a solution.

Cheers
Aymeric
Isotropix
Technical Artist - Clarisse Specialist
User avatar
aaute
 
Posts: 90
Joined: Mon Nov 02, 2020 11:59 am

Re: Questions about the incomprehensibility of glass materia

Unread postby Watchyato » Thu Sep 23, 2021 5:28 pm

aaute wrote:Hi,

I open an investigate your project.
It's a normal behavior because you have an IOR of 1.52 and energy conservation let's deviate some of rays on the glass.
If you put an IOR of 1 you get the same result.

There is also the fact you need 2 more bounces to go through the glass. If you increase all the path depth with higher number you should have closer result but still with 2 bounces difference between you render with and without glass.

Removing the Clamping can be also a solution.

Cheers


Hi thanks for your speedy response!

I understand that the IOR can deviate some of the light rays but certainly not to the extent we are seeing in our scene, we are also using 8 glossy/spec bounces in our production scene and increasing this value has no effect. When we turn our IOR down to 1 the is a marked improvement but there is still darkening. Also if we are trying to produce glass for it to be realistic and IOR of 1 is not suitble, it should be 1.5-1.6. This is all very strange as comparing similar set ups in maya/arnold we dont get the same results.


flansrjf wrote:I may have found the reason: the default clamping value is 10, and after I changed him to 100, the brightness inside the glass was significantly improved. I also saw that the help documentation did mention this


I understand that clamping can be a solution, however to get our scene back to the same look as without glass, increasing the clamping value, creates far too many hotspots in areas that are not occluded by glass. I also struggle to see how a global light clamp threshold should have an impact on whether or not the bright area is seen through transmission. What it looks like is when viewing an object through transmission, the transmission doesn't have the bit depth to display bright values, compared to the rest of the scene

From my tests I have found that the Autodesk shader set to the glass preset produces the best results, however the darkening is still very obvious.
Watchyato
 
Posts: 9
Joined: Wed May 05, 2021 11:40 am

Next

Return to General Discussion