All posts related to V2
User avatar
By m-Que
#313684
PA3K wrote:Tom, there is no refraction in pool...
Exactly, Master Tom, there is no refraction.

The point is that it works with any liquid (simple plane or full-volume) that cast caustics without it's additional editing.
And although the caustics are not real, they're very close to those real ones (if they would be seen through liquid). And it's one pass & no post, too.
It's just that there's something more... :wink:
User avatar
By m-Que
#313686
Well... if I would use them as textures, then they wouldn't change if I would move a sun or emitter position.
But they DO change, so they're sort of unique to every 'light source' position, that's why it was possible to animate the water.
By PA3K
#313690
Tom, i deleted my scene with cup, so i cannot post wireframe, but it is simple. That single sided geometry of broken line surface must be attached to main surface, because otherway it will became black in render. Is there another way, you created that broken vase ?

Bubba, your forest is very nice :wink: . Very effective 5S. Also i noticed that there is lack of z-depth in many exterior renders not only in maxwell. From some distance there should be desaturation of colors and bluish tint. This can be done in post production with z-pass mask, but it will be nice option if it can be done automaticly with some manual control adjustment - as a part of maxwell atmosphere.

m-que, what do you mean not absolutely real caustic?
#313694
PA3K wrote:From some distance there should be desaturation of colors and bluish tint. This can be done in post production with z-pass mask, but it will be nice option if it can be done automaticly with some manual control adjustment - as a part of maxwell atmosphere.
Yes, I think this is one major limitation of Maxwell right now. Volumetric atmosphere. Dust/water particles scattering light. Fog. Not using SSS, but true volumetric effects with programmable parameters.
User avatar
By tom
#313706
PA3K wrote:That single sided geometry of broken line surface must be attached to main surface, because otherway it will became black in render. Is there another way, you created that broken vase ?
Yes, I use single sided geometry as well but it's not specific to vase or another shape. I can use same geometry to crack any object. I mean I'm not fitting or modeling a crack specifically for that object.
PA3K wrote:m-que, what do you mean not absolutely real caustic?
Yes, I wonder this, too. If they are sensitive to sun direction and it's possible to animate, then why are they still fake?
#313732
tom wrote:
PA3K wrote:m-que, what do you mean not absolutely real caustic?
Yes, I wonder this, too. If they are sensitive to sun direction and it's possible to animate, then why are they still fake?
Because the real caustics (those, that are casted by dielectric material itself) are not visible through it. So, there should be 'something' that CAN be seen through dielectric material, that would produce 'caustic-looking-effect' that would be similar to the real caustics and also sensitive to light source position.
By similar I mean that it would reproduce the similar 'pattern' of the real caustics.
I already told more than enough :)
#313754
Bubbaloo wrote:
Fernando Tella wrote:Are you masking the water material with a caustics-like pattern so it let's non-caustic light in?
Ta da! I think we have a winner, no?
Ehhh... not yet.

Maybe that would work, but:
- the water surface would probably look like marble (dark and light areas all over the surface)
- in this case you would probably see only 'caustics' on the bottom and walls of the pool - the rest would be too dark;
Overall 3 light levels should be seen: shadow area (the area not lightened by the sun), light area (the area lightened by the sun), caustics area (caustics within the light area)
- that wouldn't work with full-volume water.
- how would I know that I'm using the right caustics-pattern to simulate the real caustics?
Anyways, the water material is always left untouched.
But you're definitely thinking the right way, Fernando Tella :wink:
#313761
Fernando Tella wrote:Two water objects:

1.- Real water mxm hidden to GI
2.- Masked ags mxm (caustics pattern) hidden to camera & reflection/refractions

:?: :P

http://www.maxwellrender.com/forum/view ... 2&start=30
...
Well, basically yes. I "reinvented" your method. ( :oops: Crap! How did I missed your thread???)

There're just a few minor differences.
I use:
1.- Dielectric mxm (plane or full-volume) hidden from GI and reflections/refrections
2.- A simple white material with a clip map for caustics hidden to camera & reflection/refractions
I too started with clipped/semi-clipped combination, but I didn't like that caustics are slightly visible. So I went with a clip map only for caustics for a better contrast. Works very well, meaning you can still see the lightened area of the pool and caustics look more real.

There's one more thing - my priority was to make fake caustics recreate a real caustic-pattern, so here's what I did:
You take the water object you're going to use and set a 'hidden from camera' parameter to it.
Afterwards, apply a default material to pool bottom and walls. Then set sun position to the top, in navigator go to 'top view' and render the image - this way you get a realistic caustic-pattern (with v2 rendering caustics is a matter of few minutes.) Now, you can use it for transmittance map for a "white material with a clip map". By tweaking it's color using brightness and contrast settings you can edit the width of caustics lines, so you can simulate caustics that will look very close to the real ones.
#313762
I like the real ones better. :mrgreen:
tom wrote:
m-Que wrote:I'm sure Tom knows how I did it.
It's cool you're an obedient young Jedi. :D

You've meant this? (a 25 min ML render)

No fakes here and all are real caustics, 1 pass, no post.

Physical Sky
Image

Room Lighting
Image

Room + Pool Lighting
Image

Pool-only Lighting
Image
  • 1
  • 38
  • 39
  • 40
  • 41
  • 42
  • 58
render engines and Maxwell

Ai actually can’t compete with a proper CA[…]