Kangchenjunga 1

[domdib]

A piece I've been working on for a while.

Thick snow from Lucio (http://forums.planetside.co.uk/index.php?topic=753.0)

Clouds from FrankB's excellent Cumulus pack

C+C welcome.

[FrankB]

Hey Dominic, you seem to have done something pretty strange with my clouds 
Somehow the details are all gone. Seems like the smallest scale has been increased too much, but you may have also made quite a few other changes to the preset.
I suggest you start again by just copy & pasting the configuration from the presets into your scene, and slowly tune it to your liking, but do not make more than one change at a time - these fat clouds are very sensitive to subtle parameter changes.

Cheers,
Frank

[domdib]

Well actually, I only changed one parameter, the Noise Octaves on one of the Density Fractals, because I wanted a softer, more billowy type of cloud  I guess it's not to your taste though?

[FrankB]

reducing the noise octaves is the same like increasing the smallest scale, effectively.
My taste doesn't matter, of course, I just thought something is not like expected, and you may have been looking for the reason why :-)

Cheers,
Frank

[Saurav]

Frank is right changing the noise octaves value changes the smallest scale figures, their directly related. If you wanted soft clouds your going to get better results playing with feature scale values.

Also is the real Kanchenjunga terrain (DEM)? It's a part of Nepal I always wanted to go to.

Saurav

[domdib]

Thanks for the tip Saurav, I may try that.

Yes, this is a DEM of Kangchenjunga and the surrounding mountains, taken from the Shuttle data. I was fortunate enough to see the real thing from Sikkim - though not quite so close.

[domdib]

OK, so I followed up on Frank and Saurav's suggestions re the clouds, and produced ones that are a bit more realistic (although I still have a soft spot for the billowy ones of the first image  ). Two different versions, with differing cloud densities. So what do you think? Anything else that could be improved?

[FrankB]

the second one would be my favourite. Now, you shoud think again about the cloud ambient color (or enviro light tint, not sure which it is), which is a little pink-ish. It seems that the pink-ish color is artificial in the sense that there is no light source that could produce this color i the scene. Usually, GI color on the upeer side of the clouds is coming from the blue sky color - so should be a pretty desaturated blue.
Lastly, I would think that the terrain would benefit from more small scale details. Try adding a few power fractals to the heightfield, stones and rocks etc.. if you still want to work on the image, of course.

Cheers;
Frank

[domdib]

Actually, the EL colour in the clouds is pure white, and the ambient colour is a mid-grey. The pinkish colour is being caused by the two suns, one pale yellow the other dark pink (the latter without a visible disk) that are both set in exactly the same position, 5 degrees above the horizon. The aim is to produce some of the unusual colours of dawn and dusk clouds and snow.

EDIT:

You can see the colour of the clouds and snow with a normal white sun in the attached snapshot.

Thanks for the suggestion re the power fractals. Curiously enough, I don't think I want more detail in the scene. Somehow, adding to an already existing DEM, even if it is low resolution, doesn't feel right to me - I'm a bit of a purist (although how do I square that with the weird cartoon clouds of the first image  I don't know)

[domdib]

Another POV, with a slightly lower sun.