Page 3 of 10

Re: h

Posted: Sat Nov 16, 2019 5:54 am
by burnin
Another good one!
I'm also eager to see how you'll achieve Anti-reflective coating, last time I tried to make it, it was impossible...

Re: h

Posted: Sun Nov 24, 2019 11:43 am
by yonosoy
Hi burnin.

In principle it does not have much mystery. By reading a bit, it is obtained that a quarter layer antireflective coating is composed of an IOR of the square root of the lens IOR and the thickness of a quarter of the wavelength that is to be blocked (usually 650 nm). I can´t find the spectral data of the absorption.

The problem is that it seems that the coating material does not work well with a specular material like substrate. In the Flint_Ar.pigm example, open the preview material window and navigate between the coating and the specular. The coating cancels the dispersion of the substrate. Also is not practic to use a coating material with a system of thirtenn lenses :).
Usually they order the reflexions, but don´t prevent it.
(618 Bytes) Downloaded 28 times

So I dedicate myself to try to materialize the concept and resort to a modeled coating. The thickness of the quarter layer is 1.375e-7m.

Beautiful glitch.
Is it possible to adjust what was formerly called origin_ray_nudge_distance below this dimension so that indigo does not make this mistake? Indigo applies an empty/vacuum medium to other mediums...
Glitch over 1.375e-7.
It works with a coating thicknes of 1.375e-5. There are ordered reflections but less marked than with the coating material. Some hope ("") for the real one if possible.

Re: h

Posted: Wed Nov 27, 2019 8:53 am
by yonosoy
I offer to play.

24h. 2900 spp.

A Sky.
(174.58 KiB) Downloaded 30 times
(5.99 KiB) Downloaded 29 times
It is a hard game, modulate your CPU.

The code for the clouds is a bastard of CTZn work; not better but possible for an Earth cloud layer (two).

Re: h

Posted: Sat Nov 30, 2019 4:47 am
by yonosoy
Got it. Over the wall, all working.

17000 spp; 18:30h.

The sign on the wall.

Re: h

Posted: Mon Dec 02, 2019 5:11 am
by yonosoy
An orthograpfic camera is not a sensor...

You must define a film (diffuse pure white, as fused explained years ago). Then, the ortho cam face the film.

7 h. 3000 spp.

Roots 44 color.

Re: h

Posted: Tue Dec 03, 2019 1:33 am
by yonosoy
Cathedral mockup (E 1:100).

Re: h

Posted: Sun Dec 08, 2019 8:28 am
by yonosoy
18 h.

(1013.48 KiB) Downloaded 34 times

Re: h

Posted: Fri Dec 13, 2019 2:55 am
by yonosoy
An attempt for Archviz.
Archinteriors Vol 50 Scene 001.

34h, 8050 spp.
Clay render + sgg-planilux.

Re: h

Posted: Fri Dec 13, 2019 3:20 am
by pixie
This the stuff that dreams are made of! :O

Amazing stuff

Re: h

Posted: Fri Dec 13, 2019 10:53 pm
by kscore
I have tried your shared files, now I understand your long rendertimes. You would need thousands of cores to render a short animation with these complex scenes :D

Amazing and very interesting these stuff!

Re: h

Posted: Sat Dec 14, 2019 11:42 am
by Originalplan®
@yonosoy This is incredible work!

Re: h

Posted: Sun Dec 15, 2019 3:11 am
by yonosoy
Yes pixie, is a materialized dream for me too. But this is the origin (roots).

Hey ksccore, thanks for give it a try and understand... There are several minor problems for this workflow (ten perfect spheres, no polygons). This is one reason for release it (A sky).
In the other hand, we have indeed thousands of cores in the forum. I want to push for a cooperation for made short animations. Ten on board, a snowball...

So many thanks Originalplan. Again, this are the roots, we will see what they become...

Materials... First approach, as is, except for a desaturation in the wood. Ugly!
The post production make it better and darker.

Archinteriors Vol 50 Scene 001. Take two.
12 h. I LOVE the grain that optics give. The spherical lost of focus (called bokeh in Analog plugin of Nik Collection) is cool too. Spherical aberration not compensed. I will let render 12 h more and reupload the image...
17:30 h, 2000 spp. Enough by the moment.
Another take. 33:30 h, 3900 spp. Some problems solved :).

Re: h

Posted: Sun Dec 15, 2019 7:07 am
by pixie
How helpful would be if apart of perfect spheres one could access also other perfect solids and boolean operators?

Re: h

Posted: Sun Dec 15, 2019 9:12 am
by yonosoy
I see it for conceptual work.
In truth I will be able to model the Troncotron with primitives and precedence (a special kind of boolean operators in this moments :)).
Could be fun!

Re: h

Posted: Sat Dec 21, 2019 11:13 am
by yonosoy
Take 3. 22 h.

I want to take it easy. It is necessary to remodel the glass part of the table and I want to specify the rest of the materials.
They will be about 100 h. rendering for a 3:2 3 K format.

An HDR + a modeled halogen (strangely well modeled, no possibility of use for VRay).
I like it very much.
Take3.2 20h.
This one fix some things.
I learned that it´s not the same Photoshop than AfterEffects to develope the EXR.
I will proceed with it in next days.
I really like it.
ISSUE 06_2015_00000.jpg