water

[Ros. Riha]

what is todays state of affairs?

I need to have nice water:

On one hand, I think of caustics, on the other I would like to use SSS - but it was said these two do not go along well - what approach to employ?

 

[George Ioannidis]

Sure I don't :) Here it comes... You can play with water colour. And don't forget (Roy's advice) NOT to tag water as thin.

[Alejandro Testa]

Nice material George... I agree with Akin

[Jorge Javier Lara Domínguez]

George, coincido con Akinlolu y Alex. Podrias compartir el material? Saludos!!

[Ros. Riha]

2nd sandpaper attempt - smaller scale, but the same disappointment - what am i doing wrong? What kind of the lighting do you guys use? (in my case it is sun+HDRI sky, but even if I play with channels, I am not getting there)

[akinlolu olugboji]

Why now for starters use the default exterior lighting.

[Marc Chaumier]

In my opinion, the first was the best. For so large pieces of water, I think the rubble is better, because there is several levels of wave. I would avoid too much transparency. In the case of pools, I put transparency to 1.0, because the deepness is not so high, so I consider that the water is pure, but in big volume, like in an ocean, the water is blured. I don't know what could be the caustic effect on a so big quantity of water. Perhaps increase a little bit the fresnel value. In case of pools, I like to see the bottom, more than in the reality, just an artistic prejudgement, a personal deal with reality.

[Roy Hirshkowitz]

The size of the pool is almost certainly the problem here in terms of getting any sharp caustics.  It may take thousands of passes or it may never happen.  I don't really know-- but we're definitely pushing the limits of the alogorithm here.

 

A few of the images exhibit artifacts that are due to coincident surfaces-- those we be the weird dark geometric patches on the bottom of the pool-- not the nice blurry shadows.  Using a region instead of solid for the water surface will get rid of those.

 

I would use transparency of 1.0 for the ocean, BTW-- just make it very dark.

[Marc Chaumier]

Roy, you consider that water is always pure, but when the water is troubled, dirty, or if it's "Pastis" (aniseed alcoholised french drink) ;o), you have to reduce the transparency or add some blurriness (dispersion? translucency? SSS?). What is the physical parameter to modify?

I discovered that when I move the blurriness slider in transparency tab, it move in the same time the sharpness slider in the reflection tab... right?

[Roy Hirshkowitz]

Yes-- blurriness and sharpness are inversely linked.  This is modeling a surface phenomenon.  It's more useful for modeling sand-blasted or etched glass (or translucent plastic.)

 

A glass of pastis exhibits sub-surface scattering.  It's similar to milk in that it's a high-albedo material, where the ratio of scattering to absorption is high.  These materials are very difficult to model.  I suspect our current SSS implementation will have a lot of trouble with it.  If you want to try, set the Scattering high and the Absorption to zero.  Use a nearly white color and perhaps reduce the transparency a bit.   Pastis and milk have sharp reflections, so don't change this.

 

A glass of pinot noir also exhibits SSS, but it's a low-albedo material.  To model this one set scattering to zero or just above, set absorption higher.  Sharp reflections as well.

 

Do remember that SSS takes quite a bit of extra calculation. It can be a while before you get a feel for where the rendering is heading.   It also requires a discrete volume so modeling the ocean will not work unless you also model a sea floor.  

Also, materials that scatter should have volumetric caustics-- the caustics should appear within the medium.  nXt doesn't do this-- not yet anyway.

 

 

[George Ioannidis]

Is there any specific reason you use HDRI Ros? I use for my exteriors just the sun and it turns out nicely in most of the cases. Please enlighten me more. What are the benefits of using sun+HDRI together?

[Ros. Riha]

Why to use HDRI+sun?

1. To me, sun itself does not allow for different "moods" of the image - although Roy did explain that sky changes its tone according the time of the day, the effect is insufficient for me
2. Try to render simple model using HDRI lighting only - tests with "stock nXt" probes attached
3. We use two channels - one for sun, the other for the sky... HDRI sky gives the mood, sun gives sharp shadows - you can play with sliders to obtain desired effect
4. For illustration, two renderings of the same model, only with different HDRIs and different sun intensities - it sort of worries me how different do look the same material definitions under different lighting: IOW, where should one start? twweaking materials or lighting?:-)

... the above described approach might not be the proper one, but it gives good results so forth. The main advantage is the versatility...

[akinlolu olugboji]

Error messages as to the .rar file.