VRay Sky Turbidity Ozone Quick Reference

[Jackson_cg]

Maybe these exist elsewhere for other versions of V-Ray, but I've been meaning to systematically test as many V-Ray sky turbidity and ozone combinations as possible for ages and thought the resulting table might be useful for others.

The model has a not-quite-white RGB 233,233,233 grey material applied to geometry and the infinite plane, is rendered using Physical Camera (Shutter Speed 1000, f-number 8, ISO 200), uses only the standard V-Ray sky for GI lighting (multipliers set at 1) with an RGB 255,255,255 white background. The skp model is set with a location of Malmö, Sweden, date August 2nd, time 13:30.

The table covers V-Ray's full ranges of Turbidity from 2 to 20 in intervals of 6 and Ozone from 0 to 1 in intervals of 0.2.

Image below is a thumbnail, full size with text can be viewed here.

[jrharvey]

Re: VRay Sky Turbidity Ozone Quick Reference

I cant tell a difference between the different Ozone levels at all. What does Ozone do?

[dalomar_cg]

Re: VRay Sky Turbidity Ozone Quick Reference

Its a slight difference from more yellow-ish @ 0.0 and blue-ish @ 1.0

One BIG thing to note about this is that the final appearance will depend largely on the angle of the sun vector. I've found that with a more vertical sun vector a higher turbidity prevents the image from becoming too blue, and that with a more horizontal sun vector there isn't as much of a need for a higher turbidity.

Also, White Balance can do a lot to tweak the final colors as well.

[Jackson_cg]

Re: VRay Sky Turbidity Ozone Quick Reference

Yep, the effect of the Ozone parameter using these settings is subtle to say the least- it's almost imperceptible when combined with high turbidity (at the yellow end of the table), but it is much more noticable when combined with low turbidity (at the blue end), affecting not only the hue, but also the brightness of the render (low ozone = brighter render).

I'd love to run some even more thorough experiments, combining different locations, dates and times with turbidity and ozone combinations, but it took much longer than I expected just to render, save and composite the above table.

One thing I've always wondered is if the location which is set in the SU model affects the hue or brightness of the GI lighting, other than via timezone, sun angle and rotation parameters? In other words are V-Ray's sky GI settings based purely on the geometrical relationship between sun, ground/objects and a simulated atmosphere or are there other factors taken into account? I ask as it is very noticable in the real world that for example, the light in the south of Sweden is generally much warmer/yellower compared to the colder/grey light in the south of Scotland, even though they're on almost the same latitude at 55°N and are just 10° apart by longitude. Furthermore, their climates and geography are fairly similar, both being fairly narrow agricultural land masses surrounded by water. Moreover I've observed this difference in light quality dozens of times travelling between the two countries at all times of the year so it doesn't seem to be simply a seasonal thing.

I just wondered if V-Ray sky algorithms take into account such phenomena, or are they (in the real world) simply a result of the same sun angle and rotation parameters which V-Ray already employs?

[dalomar_cg]

Re: VRay Sky Turbidity Ozone Quick Reference

I could see if I could automate something, but it might take me a while as the next few weeks are pretty hectic for me.

I'm not sure if those phenomenon are really attributed to latitude, or more about the kind of atmosphere in those areas. Although I don't think the sun sky system is really adept at catching that much subtlety, I think that turbidity and ozone would be part of getting a closer result for those cases.

If you're really interested in how it works, the sun/sky system is based off the paper that can be found at the link below

http://www.cs.utah.edu/~shirley/papers/sunsky/

[Jackson_cg]

Re: VRay Sky Turbidity Ozone Quick Reference

Thanks for that Damien, I always wondered what information the V-Ray sky algorithms were based on! I'll take a look at that paper asap.