Author Topic: General question about very glossy refractions vs. speed  (Read 2788 times)

2015-07-02, 09:59:52

iLEZ

  • Active Users
  • **
  • Posts: 107
    • View Profile
    • andersryttar.se
Hello! I am working on a shot with glowing bluetooth speakers. They are lit up from the inside with an intense spherical corona light.
Does the intensity of the light and the low refraction (0.1) of the material contribute to the "noisy" results?
After 600 passes on 1250px resolution there are still noisy parts from the refracted light and on the surface of the speakers themselves. Can I construct this material some other way with similar results but less noise?
Thanks. =)

2015-07-02, 11:26:31
Reply #1

Frood

  • Active Users
  • **
  • Posts: 1309
    • View Profile
    • Rakete GmbH
Do you need refractions for some reason? I´d use translucency only if not.

Good luck!

Never underestimate the power of a well placed level one spell.

2015-07-02, 12:47:23
Reply #2

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 9058
  • Marcin
    • View Profile
Frood is probably right.
Another idea: the noise could be actually (or partially) caused by reflection. Is it still so noisy if you disable reflectivity?

2015-07-02, 13:36:01
Reply #3

iLEZ

  • Active Users
  • **
  • Posts: 107
    • View Profile
    • andersryttar.se
Do you need refractions for some reason? I´d use translucency only if not.

Good luck!

Thanks for the quick reply!
Well, I'll certainly try it with a translucent material, and we'll see if it is as realistic as the glossy transparent one. I'll post when I've tried it. =)

2015-07-03, 08:02:30
Reply #4

iLEZ

  • Active Users
  • **
  • Posts: 107
    • View Profile
    • andersryttar.se
Do you need refractions for some reason? I´d use translucency only if not.

Good luck!

I tried it, and I think it is a bit quicker. Although I get some undesired self-shadow casting inside the object. I let it run for 140+ passes. I assume that I can't use translucency on "thick" objects, so I disabled my shell modifier. I think I mix up Vray and Corona a bit, because you CAN use translucency on "thick" objects in Corona, right? I'll try that too.
I also rendered the scene with glossy refraction again but with different values and got a pretty good result with some camera raw editing.



« Last Edit: 2015-07-03, 08:11:10 by iLEZ »

2015-07-03, 09:06:03
Reply #5

iLEZ

  • Active Users
  • **
  • Posts: 107
    • View Profile
    • andersryttar.se
Frood is probably right.
Another idea: the noise could be actually (or partially) caused by reflection. Is it still so noisy if you disable reflectivity?

Thanks, I just tried to disable ALL reflections, but no luck, still a bit noisy. Good idea though. Perhaps I should just accept that the scene is a bit hard to render because the mail source of light is refracted/transluced(?) through a material. I made another attempt with a translucent material on a "thick" plastic. (2mm). This also has white absorbtion set to 10mm, and white scattering. I appreciate that the surface reflections show up clearly when using translucency. On the left side of the green speaker on top for example. I think this is perhaps the most realistic way to do it. 279 passes this time, at 1k rez.

(I'm *really* hoping for a native region render in the Corona VFB in future releases! :D)
« Last Edit: 2015-07-03, 10:30:44 by iLEZ »

2015-07-03, 09:20:16
Reply #6

Ondra

  • Administrator
  • Active Users
  • *****
  • Posts: 8896
  • Turning coffee to features since 2009
    • View Profile
(I'm *really* hoping for a native region render in the Corona VFB in future releases! :D)

me too ;)
Rendering is magic.
Private scene uploader | How to get minidumps for crashed/frozen 3ds Max | Sorry for short replies, brief responses = more time to develop Corona ;)

2015-07-03, 10:46:59
Reply #7

Frood

  • Active Users
  • **
  • Posts: 1309
    • View Profile
    • Rakete GmbH
you CAN use translucency on "thick" objects in Corona, right? I'll try that too.

Yes, right. And it looks more realistic/interesting so I wouldn´t deactivate the shell mod. What´s that background in your scene? Maybe something other is causing the noise (background/floor material) - I would tweak/test in a neutral environment.

The attached scene (for further testing if you or someone else wants to) is pretty clean after 150 passes / 1k. I even activated reflection on the floor and on the cubes material.

Good Luck!
Never underestimate the power of a well placed level one spell.

2015-07-03, 11:05:32
Reply #8

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 9058
  • Marcin
    • View Profile

2015-07-03, 11:15:03
Reply #9

iLEZ

  • Active Users
  • **
  • Posts: 107
    • View Profile
    • andersryttar.se
you CAN use translucency on "thick" objects in Corona, right? I'll try that too.

Yes, right. And it looks more realistic/interesting so I wouldn´t deactivate the shell mod. What´s that background in your scene? Maybe something other is causing the noise (background/floor material) - I would tweak/test in a neutral environment.

The attached scene (for further testing if you or someone else wants to) is pretty clean after 150 passes / 1k. I even activated reflection on the floor and on the cubes material.

Good Luck!

Excellent! Thanks for all the help! I'll try it in a neutral environment, but right now in my interative window the environment is *really* clean after 2500 passes, and the cubes and the immediate light from the cubes on the wood floor is still very noisy, so I suspect the material. =) I'll try a complete reset of the scene, and a neutral environment, maybe I've screwed up in some unexpected way.

2015-07-03, 11:40:32
Reply #10

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 9058
  • Marcin
    • View Profile
Showing render settings could help too as I think 250 passes might be too much for this scene, I would suggest increasing GIvsAA (to 32 maybe?) and/or LSM, the result of this would be:
-better GI, glossy reflections/refractions in the same render time
-less passes in the same render time = worse AA quality, but it should not have much impact here

Also these might be helpful:
https://coronarenderer.freshdesk.com/support/solutions/articles/5000516731
https://coronarenderer.freshdesk.com/support/solutions/articles/5000526159
https://coronarenderer.freshdesk.com/support/solutions/articles/5000515640

2015-07-03, 14:34:23
Reply #11

iLEZ

  • Active Users
  • **
  • Posts: 107
    • View Profile
    • andersryttar.se
Just replying before leaving work, with a comparison between the different GI vs. AA and LSM settings:

2015-07-03, 15:50:10
Reply #12

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 5801
  • Let's move this topic, shall we?
    • View Profile
    • My Models
Your comparison is not valid because you're tested with fixed passes when you should test with fixed time.
I'm not Corona Team member. Everything i say, is my personal opinion only.

2015-07-03, 16:21:07
Reply #13

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 9058
  • Marcin
    • View Profile
Your comparison is not valid because you're tested with fixed passes when you should test with fixed time.
I don't see anything bad in checking with fixed passes, but it looks like they are not even fixed here. :)

2015-07-03, 23:34:56
Reply #14

iLEZ

  • Active Users
  • **
  • Posts: 107
    • View Profile
    • andersryttar.se
I dumb. :D

I'll do a proper like-for-like test on monday, I can't even begin to explain what went wrong with this test! :)