WIPs: CTZn

[CTZn]

more like this. As it came out the maths were correct but the effect was just bad.

Not sure what the banding is, but I can make it worse...

[OnoSendai]

so cool

[pixie]

more like this. As it came out the maths were correct but the effect was just bad.

Not sure what the banding is, but I can make it worse...

Thats tha shit!

[CTZn]

Thanks ! Ono, your clouds are not beaten yet !

[OnoSendai]

Stepping artifacts are probably due to step_size being too large btw.

[CTZn]

yep they were. Maybe it's too small now... A short serie from the shader featuring a couple parameters. Sorry all are 25 spp only.

[yonosoy]

Some delay with maya ... hope than in few days we could see it in scale.

Silent edit: glad to hear that ... expectancy about the news.

[CTZn]

Maya is optional, the shader is scalable. New version, before only the top part of the clouds was shaped, here both top and bottom have a shaping of 50%, leading to "round clouds" sort of. Plain default, average shader values, no tweaking.

I am going to render someting bigger and with more spp's

[Mor4us]

WOW, this is simply stunning!!!
Is it possible to use this shader as volume noise for arch viz on plane scenes, or is this only usable on earth-scaled scenes.
Would love to test it in a huge scaled cube flying above an arch viz scene

best regards!
Markus

[CTZn]

Hey Mor4us hi, thank you !

Yes sure, you just tick the 'Box' parameter in both the absorbtion and scattering editors and you can set the top height and clouds layer depth in the respective shaders, toward the beginning of the code. The box shader uses the world ground as reference (at a height of zero meter).

It's real !

The one thing I forgot to expose is the very clouds scaling. It is the float value below (0.10), call it frequency. Smaller value = larger clouds.

Code: Select all

	#	some low frequency noise
	noize =  fbm01(pos * 0.10 , 8)

Mind you, I'm pretty happy wth the result since that's the only noise used in the shader. There's room for more randesomness !

Last but not least with the clouds frequency must one change both teminal multiplier for the shaders, namely 0.3 and 50.0. They're set against 10m wide clouds really.

[Mor4us]

wow, thanks for the files... but this reaaaally renders slow as fuck
6 SPP in 1hour......

why is that?

[CTZn]

That's the way it is, better off rendering at a small resolution with no supersampling for speed
You can try increasing the medium step size (double it) but steping artifacts will only be more visible.
Also it may enhance performance to scale the clouds sphere to the minimum required.
Lastly rendering a portion of the clouds within a box would be much faster than for the whole sphere.

I'm open to suggestions.

[fused]

Awesome shader!

[CTZn]

Thaaank you Yves !

[CTZn]

Also the scattering value of 50.0 might be too much, lower value may suffice and converge faster. Inversely, increasing the absorbtion factor will kill more rays and make faster, darker clouds.

Here's what I've got at 34 spp; power 2.0, offset -0.125, shape shift 0.25. What shape shift does is larger clouds, toward their base or toward their summit for low and high values respectively. 0.5 is for the so called round clouds.