Fire in 3Delight...

[Rez007]

I didn't get a chance yet to look at 3Delight, but I was wondering if anyone tried to make nice looking fire with it and if that is possible? - unlike the cottonball effect we get now with the current volume shader...

Thanks.

[ActionArt]

I tried it a bit but I don't think the volume shape tools are any different. I did notice that algorithmic textures aren't behaving like expected but didn't investigate further yet.

Did you take a look at the tutorials here:

http://usa.autodesk.com/adsk/servlet/it ... eID=123112

The rolling fire in this looks pretty good.

[Rez007]

Thanks for the reply, and somehow I missed those videos. I will have to take a look at that rolling fire - it does look nice.

Thanks again!

[ActionArt]

No problem! I found it quite useful.

[Mathaeus]

I didn't get a chance yet to look at 3Delight, but I was wondering if anyone tried to make nice looking fire with it and if that is possible? - unlike the cottonball effect we get now with the current volume shader...

Thanks.

yeah, it seems it still only possible, using a huge count of particles - and finding a way to apply fire-like motion to them.
Now if this one became available to SI....
Anyway, I believe that picture at bottom exactly describes what's missing, and it's recognizable in all FumeFx renders. 'Second layer' of details, they call it 'WT', wavelet turbulence, that's just 'drawn' over - of course, using a 'bit' more advanced mapping, than we are able to apply in render tree.

[Rez007]

Thanks for the reply, and that lower last image looked pretty cool.

I saw this http://www.youtube.com/watch?v=D0sV9uGJaJw a while ago, and no matter what I do, I cannot get this effect. The video link that ActionArt posted looks pretty good for me for the time being. If someone knows how to get the quality from my link posted and would be so kind to post what they did, it would be much appreciated.

Thanks.

[ActionArt]

I think the key is in getting your algorithmic textures just right and that can take some time.

It might help to have some more variety available. I recently purchased these:

http://www.shaders.co.uk/enhance_xsi/

I have to say they're great and the variety is amazing. I haven't used them to make fire yet but I think it would help a lot. The ones the come with SI are quite limited. I'm not sure what he used on your example though.

[Hirazi Blue]

While I agree that "Enhance:XSI" is an invaluable tool, in the case of the original question it probably wouldn't help much, as the "Enhance" procedurals aren't available for 3Delight yet. There seem to have been numerous requests already for an "Enhance:3Delight" but so far it hasn't materialized yet AFAIK.
;)

[ActionArt]

Ya, I just assumed it would be OK to get a fire effect any way possible

Has anyone had any luck getting procedurals to work in 3Delight (with volume shaders)? If not, the volume rendering would be kind of useless.

[Mathaeus]

what volume... here it definitively works with ICE particles, at least using Particle Renderer shader compound. Probably DNA team used just this interface for their stuff. I've tried fractal scalar texture. 'Cell size' parameter (in Particle Renderer) seems to has similar effect of 'precision'.

Volumic lights seems to work too, to some extent.

It would be interesting to try something along line of Em Render Point Cloud, using 'lightweight' particles and non-volume shaders (lambert or constant,) on point-particles...
Anyway, for believable animated fire, using 'reasonable' number of particles, I'm afraid SI users will need to wait for plugin or something, able to do mentioned wavelet turbulence layer.

If you mean volume 'inside' meshes or NURBS, I didn't sow any sl definition in rsl folder, with some of names of SI volume shaders (maybe I'm wrong). Still, it's possible to apply Renderman Shader property to object and use 'external' shader. But at this point, everything (including procedural texture) should be coded (or called) in that shader...

For now, I'm occupied with more basic stuff, getting reflective AO to work, getting 'private shadows' from infinite lights.... MR has a 'nice' feature to fit shadow map of infinite light to scene extent, and now, this 'genious' (and unique) feature is applied to 3delight too - even Redermans have a myriad of options for shadowmaps. With good old Affogato transator, this was just a XY scale on light..

[CiaranM]

It would be interesting to try something along line of Em Render Point Cloud, using 'lightweight' particles and non-volume shaders (lambert or constant,) on point-particles...

You'll have trouble using a lambert or something like that. The RiPoint primitive always has its normal oriented towards camera. So, for correct shading and self-shadowing, you have to write a shader to point N in a more useful direction.

[Rez007]

That looks really nice!

[Mathaeus]

You'll have trouble using a lambert or something like that. The RiPoint primitive always has its normal oriented towards camera. So, for correct shading and self-shadowing, you have to write a shader to point N in a more useful direction.

Thank you, I thought it's sometimes like that.

[CiaranM]

No problem. I can share a shader pretty soon.