Author Topic: romullus wips  (Read 148665 times)

2015-12-27, 22:40:20
Reply #225

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 6665
  • Let's move this topic, shall we?
    • View Profile
    • My Models
Another material session, this time felt.
I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2015-12-28, 00:20:57
Reply #226

Juraj

  • Active Users
  • **
  • Posts: 4114
    • View Profile
    • studio website
that's brutal :- ) Care you show more of it ?

I do a lot of felt fabrics of furniture maker but with approach that does not look best at close-up. This looks marvellous.
talcikdemovicova.com  Website and blog
be.net/jurajtalcik   Our studio Behance portfolio
Instagram   Our studio Instagram, managed by Veronika

2015-12-28, 10:33:59
Reply #227

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 10158
  • Marcin
    • View Profile
My guess would be "simply" hair geometry scattered over the surface but I am also curious about any tricks used.

2015-12-28, 10:47:54
Reply #228

Nekrobul

  • Primary Certified Instructor
  • Active Users
  • ***
  • Posts: 1025
    • View Profile
Looks mindblowing. How you achived it? X_X
---------------------------------------------------------------
https://www.blackbellstudio.com/
https://www.behance.net/blackbell3d
CEO at "Blackbell Studio"

2015-12-28, 11:13:35
Reply #229

johan belmans

  • Primary Certified Instructor
  • Active Users
  • ***
  • Posts: 727
    • View Profile
    • belly.be

2015-12-28, 11:20:25
Reply #230

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 6665
  • Let's move this topic, shall we?
    • View Profile
    • My Models
Thanks guys, not so much tricks here :] First i searched for a felt texture, found a really good one at wildtextures. Then fed this texture to Knald and made normal map from it. Also made height and AO textures as well. CrazyBump, NDO, B2M or other similar program can be used here. Even Photoshop. Material itself is very simple and it looks pretty good, especially at front facing angles. At glancing angles it's very flat and CGish though. Displacement can help a great deal here, but still not very convincing, especially in close ups. So i swapped displacement with simple spline modeled fibres, scattered all over model with CoronaScatter and view instantly became much more realistic. Fibres material is extremely simple - CoronaMultiMap in instances mode in diffuse map and transluceny from 0,3 to 0,8 depending on what look you want to achieve. Less translucency shows fibres colour better, more translucency helps to better blend in fibres with felt material and looks very nice in contrary light. As an added benefit, fibres can help to mask shadow terminator problem, from which this material suffers quite badly, due to fine and strong normal map. Additionally i added falloff map to scatter's density slot, to steer more fibres in areas where they needed most. Unfortunatelly CoronaScatter refuses to respect color map curve adjustments, so it's hard to fine tune distributin. AO doesn't work here too, in case someone wants to try it :]
« Last Edit: 2015-12-28, 11:25:42 by romullus »
I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2015-12-28, 11:30:07
Reply #231

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 10158
  • Marcin
    • View Profile
I wonder if flat "sprites" with opacity-mapped fiber textures would work here. :)

2015-12-28, 11:38:51
Reply #232

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 6665
  • Let's move this topic, shall we?
    • View Profile
    • My Models
I can't see why it wouldn't. If you don't affraid opacity OFC :] Hair&Fur could do a trick as well.
I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2016-02-28, 16:46:29
Reply #233

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 6665
  • Let's move this topic, shall we?
    • View Profile
    • My Models
Rerendered my ants anmation with adaptivity and denoise. Not only the picture became much cleaner, but aslo it saved  about 30% of render time. Although i'm fairly confident that i could try to reduce render time even further if needed. Must say, i'm impressed. Even more than i thought i would be :]

I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2016-02-28, 17:46:03
Reply #234

Nekrobul

  • Primary Certified Instructor
  • Active Users
  • ***
  • Posts: 1025
    • View Profile
Wow.
---------------------------------------------------------------
https://www.blackbellstudio.com/
https://www.behance.net/blackbell3d
CEO at "Blackbell Studio"

2016-02-29, 08:22:33
Reply #235

rozpustelnik

  • Active Users
  • **
  • Posts: 120
    • View Profile
@romullus
Can You provide some more tests on denoising/adaptivity? Some "real-world" scenarios ie. exterior shot with foliage and dof, interior shot with complicated GI and reflective materials...

2016-02-29, 10:48:57
Reply #236

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 6665
  • Let's move this topic, shall we?
    • View Profile
    • My Models
I will do... probably. But i think build with adaptivity should go public very soon, so everyone can do tests by themselves.

Meanwhile i did some more comparison shots of my little fellows, because i find that particular scene benefits from adaptivity very much. Without DOF and MB this scene samples very fast, but as soon as those effects kicks in, some areas requires very high amount of samples, for noise to be at acceptable level. And adaptivity fights that problem very efficiently. This is by no mean a scientific test, i just fired a render, waited for noisiest part of an image to became acceptably clean and stopped rendering. Settings were such: GIvsAA - 4, LSM - 2, adaptivity - 0, denoising - 1.
I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2016-02-29, 10:58:07
Reply #237

Nekrobul

  • Primary Certified Instructor
  • Active Users
  • ***
  • Posts: 1025
    • View Profile
I will do... probably. But i think build with adaptivity should go public very soon, so everyone can do tests by themselves.

Meanwhile i did some more comparison shots of my little fellows, because i find that particular scene benefits from adaptivity very much. Without DOF and MB this scene samples very fast, but as soon as those effects kicks in, some areas requires very high amount of samples, for noise to be at acceptable level. And adaptivity fights that problem very efficiently. This is by no mean a scientific test, i just fired a render, waited for noisiest part of an image to became acceptably clean and stopped rendering. Settings were such: GIvsAA - 4, LSM - 2, adaptivity - 0, denoising - 1.

There are couple of strange things going on. Maybe it is because the number of pses is very different? Could you make one more test with same amount of pases lets say 20-40?
---------------------------------------------------------------
https://www.blackbellstudio.com/
https://www.behance.net/blackbell3d
CEO at "Blackbell Studio"

2016-02-29, 11:11:23
Reply #238

romullus

  • Global Moderator
  • Active Users
  • ****
  • Posts: 6665
  • Let's move this topic, shall we?
    • View Profile
    • My Models
It wouldn't be very fair comparison, because with adaptivity you cannot to directly compare passes with passes anymore.
I'm not Corona Team member. Everything i say, is my personal opinion only.
My Models | My Videos | My Pictures

2016-02-29, 11:26:14
Reply #239

Nekrobul

  • Primary Certified Instructor
  • Active Users
  • ***
  • Posts: 1025
    • View Profile
Then may be same time for let's say a minute?
---------------------------------------------------------------
https://www.blackbellstudio.com/
https://www.behance.net/blackbell3d
CEO at "Blackbell Studio"