Thick Atmosphere Glow

[Tangled-Universe]

On those two images. The one on the left looks more realistic. the scale appears right. I think that the GI is causing this visual effect. More GI makes thing seem smaller.

Exactly. This is exactly what I meant with saying:

Atmosphere <-> GI <-> surfaces

These two images have the same settings. Only difference is a tweak in atmo density. That's it.

So, it doesn't need much effort/pushing to get the atmosphere affect your scene. It's *not* the GI.

The play between atmosphere and GI is probably correct. For example in the UK we generally have dull weather. You just don't see too much light bouncing around unless it's a very bright day.

That's not GI either (as a starting point). Dull weather = overcast weather. Clouds scatter and spread the light, resulting in one diffuse lightsource instead of a more pin-lightsource, as the sun usually is. In a way you can of course call this GI too.
*This- light is then further scattered in a GI manner, but it *starts* with the atmospheric condition.


In TG Atmosphere affects GI and GI affects surfaces, but atmosphere also affects surfaces.
The difference in saturation between the two images shows this.

[Tangled-Universe]

Because of me we're deviating somewhat.
I'm still interested in seeing your experiments to achieve a systematic approach on exposure vs low dynamic range output.
I have to admit I "just fiddle" with it to get the best result. No system.

[Oshyan]

Essentially the very nature of what you're talking about here suggests the use (and discussion) of various HDR processing and/or tonemapping applications. Fortunately in recent years digital photography has started to embrace higher dynamic range imagery to varying degrees, so we have an increasing number of tools, some dedicated to tone mapping tasks, others with such features simply integrated into an overall image editing system. Examples include Photoshop and Photoshop Lightroom, HDR Efex and Photomatix, etc. So I would look at some of the software the photographic industry is using (Lightzone being an example, unfortunately now defunct). You could also quite sensibly look to the film industry, however their tools tend to be much more expensive and often complicated, designed much more for dealing with such changes in motion, over many frames (e.g. compositors), so they're less applicable to the typical user here I think.

- Oshyan

[efflux]

Although you can edit HDR images in other apps (much more so now). I  have no idea what systems they have to handle the processing because I've only been using Lightzone. The beaty of Lightzone is that it has one algo called relight with a number of features to change the lighting. Both dark areas and light, distance of the lighting etc. It has zone mapping where you can move the levels of different zones so that the whole range isn't effected i.e. you can lock points. Probably there are such like things for Photoshop or even movie editing apps. There must be, especially for movies that have landscapes.

I'm not doing experiments with this yet. Maybe some other people can try it? The main thrust of the thread is that the TG2 output without any exr editing can look terrible but that's actually normal. I don't think a lot of TG2 users consider this and a lot of renders look like straight TG2 outputs. Sometimes this is OK but I really think that outputting to exr and editing that is absolutely essential to play with. Eventually I'll try doing really blown out glowing stuff that is not far from real world scenarios and I'll see how it works. I have one planet with some extreme atmo settings. It has valleys with very thick haze. When I go back to that one I'll do some tests. Shots into the sun have white skies.

[efflux]

That's not GI either (as a starting point). Dull weather = overcast weather. Clouds scatter and spread the light, resulting in one diffuse lightsource instead of a more pin-lightsource, as the sun usually is. In a way you can of course call this GI too.
*This- light is then further scattered in a GI manner, but it *starts* with the atmospheric condition.

I don't agree with this entirely as far as the solid surfaces are concerned. In a very clear atmosphere you get a lot of strong bouncing of illumination directly effecting solid surfaces. If the atmosphere is thick then the light scatters around the atmosphere more and surfaces don't appear to have as much light bouncing off them. Not as much light is reaching them. The sky could be quite glowy but the surfaces quite dull so increased atmosphere setting would quickly dull GI effect on surfaces. It's not really one diffuse lightsource instead of a more pin-lightsource. A lot of the diffuse lightsource is being scattered away from solid surfaces before it reaches them.

[efflux]

I am actually running some experiments with this in my last planet which is called Lonely Lakes in image sharing.

My findings are now leading me to believe that haze glow isn't too bad around the default settings (I'll experiment further with that) but I think haze has to be thicker and in particular in conjunction with more severe exponential in many environments to get a sense of space and distance into the atmosphere. Only other method is to build in low clouds. Looking at real world environments there is always quite a bit of atmospheric haze thickness. It usually only seems clear when you are close to things.

[efflux]

Maybe a relationship of more atmosphere less glow and vice versa.

[efflux]

I'm going to add new renders from my current planet in Image Sharing rather than here. It now relates to what is here because I'm tweaking atmosphere. However, in Image Sharing you will see what difference are occurring from the earlier renders which is important.