Sun Sky options hidden.

[GaudiGalopin3324]

viewtopic.php?t=8832&start=50#p41186 Scroll to the top.

Thanks for the link! Yes, Ceciliabr is a real artist. It is always a pleasure to study such works. And yes. I also prefer options with diffused light, without the participation of the harsh sun. It's not for nothing that photographers put a white cloth on their flash.

[GaudiGalopin3324]

Yafaray to use pathtracing, and increased the diffuse bounces to 3 or 4. Now this is quite slow to render, but the effect is pleasing.

Yes, this is the only magic mode that is needed. I tried to apply it to the interior earlier and couldn't. Simple pictures are obtained, but the interior rendering filled with objects and lamps lasts forever.. Your daylight method is probably also the only way to get an interior with reflected light. I'll try your method tomorrow, it's very interesting. So far, good AI models process everything quickly, but there is no complete control here, I often get angry..

[sjb007]

I've created test release builds of DobryDom3D that work just like SH3D release packages. I've already tested the Linux version. I don't know what you use, but it would help me confirm the Github builds are working if you can test it regardless of system.

For the package see: https://github.com/DobryDom3D/DobryDom3 ... .1.0-test1

For the instructions see: https://github.com/DobryDom3D/DobryDom3 ... r-Guide.md

@Xiste This also means that adding ozone should now be trivial, as I have Github compiling the YafaRay library, so both should see it added soon.

[sjb007]

@Xiste A little more on Ozone as a configurable parameter. It wasn't straightforward as I thought. Took a couple of rounds to get it right, and I had to do a bit of reading to understand it a bit more.

On earth the valid range is between about 0.2 and 0.5. with the default 0.35 being a solid general purpose value. There are various technical limitation of Sunflow and YafaRay renderers that mean all we can do is tint the sun colour. The effect is more pronounced when the sun angle is lower (sunset/sunrise) This gives a practical outcome at the far, extremes that looks like this:

Sunflow low (0.15) ozone:

test_trans_dd3d_sfoz_low.png (68.37 KiB) Viewed 335 times

Sunflow high (0.6) ozone:

test_trans_dd3d_sfoz_high.png (69.77 KiB) Viewed 335 times

Those low/high values are pretty much outside of what occurs on earth. From Grok:

• Global average ≈ 300 DU (0.30 atm-cm).
• Normal variation is roughly 200–500 DU (0.20–0.50 atm-cm):
  • Tropics: often lower (~250 DU).
  • Mid-latitudes: commonly 300–400 DU.
  • Higher latitudes / winter-spring: can reach 400–500+ DU.
• Extreme low values (< ~220 DU, and especially the Antarctic ozone hole averages around 100 DU) are uncommon outside polar spring and are not “everyday” conditions.

Yafaray in the followup message (seems to be an attachment limit kicking in)

[sjb007]

YafaRay low ozone:

test_trans_dd3d_yroz_low.png (70.12 KiB) Viewed 333 times

YafaRay high ozone:

test_trans_dd3d_yroz_high.png (70.12 KiB) Viewed 333 times

As you can see, the effect is not huge even at the far extremes, side by side. Is this still something you feel is useful after seeing the comparison?

[Xiste]

I'm away for the moment. What time is this? It seems to render som blue sky, at least. As you can see my illustrations, the Sun Sky was all bright yellow.

[sjb007]

This was 4pm on 31st of October, at GPS position somewhere in NE England. 5pm was pitch black, so it was within an hour of sunset with the sun just above the horizon, i.e. golden hour. I had the turbidity dialled back to 2.5 from the default of 6.0. A more realistic turbidity might have been 3.5.

Ozone has less effect when the sun is higher in the sky. At 60° (i.e. equinox midday in Cairo) ozone of 0.6 knocks red and green down by about 3% compared to 0.1. At 3° (sunset/sunrise) red and green are halved. But in both these cases the turbidity is much more influential over the full day. In Cairo the turbidity will be really high.

If you're worried about blue sky... It depends how you set it up (renderer options) location, date, time of day, turbidity, and where you are looking. Straight up can be a pure azure blue, but the horizon is usual near white. If your FOV is narrow, you won't see a lot of striking blue. Here is the same test file, you can see the options set (I'm not currently running the ozone fix, so this is at the standard 0.35. It's using pathtracing with 3 bounces in YafaRay. Because I'm using "Virtual visitor" here I still have to set the "Sky lights interior" otherwise I get degraded lighting.

Screenshot From 2026-11-06 23-12-58.png (89.81 KiB) Viewed 286 times

And if the Preetham sky physical modelling isn't blue enough for you, you can always use the Transparent sky option, and paste it over another image.

[sjb007]

As you can see my illustrations, the Sun Sky was all bright yellow.

I re-read this, and and I can't figure out if you are complaining about too little blue, or too much blue. I gave you midday to show you can get more blue, but I in looking again, now I think you are complaining that you want the whole sky a warm yellow. The previous images were deliberately not aimed at the sun to get a spread of colour. You can tell this from the direction of the shadows. If the camera points toward the sun you can get more yellow. If you want an even more saturated yellow you can boost turbidity. i.e. 7 is really yellow.

Screenshot From 2026-11-07 03-28-09.png (182.62 KiB) Viewed 277 times

Atmospherics are complex. The Preetham model is an approximation design for ease of computation. On top of that, I'm not convinced the Sunflow and Yafaray implementations are entirely correct, with a few shortcuts and limitations in the accuracy. Don't expect a 100% accurate physics simulation. There's professional tools for that kind of stuff.

[Xiste]

The ability to run automated batch renders or overnight video sequences, to see how the shadows fall is what I would like, that means defining the location to se exactly the position of the shadows.
So both illustrations are @07:00. on the 26/06/2026, The top has my created morning sky textured. The one under is rendered with Sun Sky. I can see a difference. The Sun Sky is wrong. The morning sky might look like this somewhere else in the world, but not here in Norway. There's always a blue sky up there, unless it's cloudy.

[sjb007]

Ah, OK. So a little digging, and playing with sky models in Blender, and I think I understand what is going on. The sky models in SH3D/DD3D are Preetham. This only accounts for single scattering of light in the simulation of atmosphere. Parts of the sky are blue, but near the sun it looks yellow. The newer versions of Blender now have Nishita sky textures (Preetham is based on the original Nishita research, but optimised more for performance.) Nishita in Blender used to be single scattering too, and gave the same yellowing around the sun. Since 4.x at some point Nishita got a multi-scattering mode, and this gives the sky around the sun a much bluer sky.

sungle-multi-scattering.png (1.31 MiB) Viewed 235 times

I'm afraid with SH3D/DD3D you are not going to get accurate skies with the current limitations. One option is design your space in SH3D/DD3D, then export it to obj files, import to Blender, and do your lighting and sims in Blender. Blender rendering is also waaaaaaay faster, with built-in denoising. DD3D has an updated export that preserves structure in the obj file, that works with Blender import options so the scene in Blender is not one massive uneditable mesh.

If you want more accurate lighting (but still not perfect) you need something like Blender, and Nishita. The next step again, there are Blender plugins that claim to have even more accurate lighting, i.e. Real Sky, Dynamic Sky, Pro Sky V2, Pure-Sky Pro, Sky Plus (some of which are paid) as just a few examples.

(BTW, there's also a lot of funky stuff the human eye/brain does to compensate when we look at huge dynamic ranges, colours, etc. For example... sun bouncing off the floor of a room should be blindingly bright, or the rest of the room should look dim, but our eyes rapidly contract and dilate to compensate. SH3D lighting is not realistic here - I don't think it was ever the goal.)