Realistic sun & sky intensity

[allemyr]

Hi Chaosgroup,

I want to learn more about the intensity of the sun and different parts of the sky.

I read here: https://docs.chaosgroup.com/display/...em+%7C+VRaySun but there are quite many links that is old now.

I've used Gutries recent skys for a while now https://www.pg-skies.net/collections/recent-hdr-skies but realised that I have to know a bit more about the intensity relationship between the sun and the sky a clear day. I get nice result from his HDRis but want to learn more.
The second reference would be good to read from the Vray Sun Sky docs page but it's not availible ([2] R. H. B. Exell, The intensity of solar radiation, 2000 )

First I thought I should do my own not so scientific test that would consist of taking photographs with a 600mm equivalent lens with spot meter set in camera and ofcourse with some Solar ND filter infront of the lens, they often reduce the sun about 16-18 stops with that i could meter the sun sort of.
I understand that its very complicated stuff, but if every value I start with in the render is 90% right I'am happy, have no clue if the latest PGskies are anyware close to correct in terms of intensity values between sun and sky, I get realy nice result with his 1714 hdri thougt.

Won't write any longer that this for now but a link where I can read further about the intensity of the sun and sky would be superb.

Further thoughts I have is choosing right exposure to start with when building materials from that later after a sun and sky system is decided in the scene, and also a bit about colormapping.

Right now my workflow is based on these:

http://www.peterguthrie.net/blog/201...hting-tutorial
http://www.peterguthrie.net/blog/201...-for-interiors

Best Regards

Karl Allemyr

[glorybound]

Realize that those tutorials are 8 years old.

[allemyr]

Yes, but they are ok for my workflow.

The topic is more towards sun and sky intensity in real world.

[joconnell]

For a normal vray light it's up around 280 for the multiplier of the sun!

It's reeeealy hard to actually shoot a hdr with a full range of values since it's tough to get ND filters for fisheye lenses - most of the time people end up sticking some ND gel film on the back of the lens since there's no way to mount a filter to the front!

[allemyr]

Hi,

Thank you very much! Ok I understand.

Even tought I might read further about real sun and sky intensity and make a own test to see the difference in illumination between them to get the sort of right values between shadows and objects in sun, it might not give me much. I will start to use the Vray Sun & Sky I think instead or match the HDRi I'am using so it matches the value from a Vray sun sky setup. I understand that Chaosgroup have made intense tests on this and that you should leave values sun and sky values at 1.

I've searched the forum a bit and there are quite many interesting topics about this: https://forums.chaosgroup.com/forum/...-natural-light

That thread summerise the thoughts I'am having its about starting with some sort of correct values when you create a scene let's say 75% correct values, not 90 wrong values of sun sky intensity which you can end up with when using a hdri that's not very good.

OffT: Trying to create a photorealistic rendering is a huge statement, a camera already have 100% physical accurate world to start with and record and take a picture off. To do a photorealistic rendering you also need a 100% physical accurate world to do observations in.

I will continue to read the threads here on Chaosgroup about those topics, I see that Chaosgroup support isn't too present in those threads, I might understand why, its fine.

Anyways, I'am very happy with the render engine you are creating keep it up!

[vlado]

You can find the "The Intensity of Solar Radiation" text here: http://www.jgsee.kmutt.ac.th/exell/I...tySolarRad.pdf - we need to correct the link in the docs, thanks for the pointer.

Best regards,
Vlado

[allemyr]

You can find the "The Intensity of Solar Radiation" text here: http://www.jgsee.kmutt.ac.th/exell/I...tySolarRad.pdf - we need to correct the link in the docs, thanks for the pointer.

Best regards,
Vlado

Hi,

Can't access that university website from Sweden, or their might be a login to it.

[samuel_bubat]

I?ve written an inhouse tool that mimics Vray Sun and Sky with a direct and gradients. This was made to load and save lighting presets as well as beeing more flexible when it comes to sun angle, colors, soft shadows and so on. Sometimes the Vray Sun and Sky are not flexible enough. Anyway, I measured the values for the Sun (Hosek Sky Model) by aligning a white plane towards the sun,rendered it without GI and saved it to an hdr. From that HDR I could read the (unclamped) values of the white plane wich represent the intensity and color of the sun. For the sky colors and intensity I simply rendered a spherical hdr an measured the values at certain angles. Attached is a rendering that compares the Vray Sun with my Tool. It is not perfect as it for example doesn?t mimic the sun halo, but I comes pretty close. So I can?t tell you about the exact values of the real sun. But I can tell you the Vray Sun has a multiplier of pretty much exactly 77 at noon. Attached is also a screenshot of my tool that shows the environment colors and intensity as well as the sun color.

[MANUEL_MOUSIOL]

that looks like an awesome script! any chance you could share it with us?

[^Lele^]

I can tell you the Vray Sun has a multiplier of pretty much exactly 77 at noon.

While when viewed directly, the sun disc is around 3 million float (hey fireflies in reflections! ).
Both will however depend to a certain degree by the settings for turbidity (the values above imply the default of 2.5), which will lower sun and sky intensity the higher it'll get.