Experiments..
Re:
http://download.autodesk.com/us/maya/20 ... =d0e422209OnoSendai wrote:I'm still thinking about how to handle the volume data formats etc...
Any more suggestions are welcome!
You may bookmark this one
obsolete asset
Re: Experiments..
Bumping this ancient but fun thread...
Volumetric scattering is now way more efficient.
Some new results: see attached pic, rendered in ~10 mins or so on a laptop.
Expect some new sky + cloud experiments soon.
Volumetric scattering is now way more efficient.
Some new results: see attached pic, rendered in ~10 mins or so on a laptop.
Expect some new sky + cloud experiments soon.
Re: Experiments..
Does this also result in faster SSS???
sky + cloud stuff is really something nice to see in Indigo.
I also wounder about the workflow and integration into the 3D Apps/Exporters!?
looks for me like having some boxes scattered around in the sky with some "cloud" shader onto it,
that does the work for soft falloff to the edges of the box...
Anyway, more examples please ^^
sky + cloud stuff is really something nice to see in Indigo.
I also wounder about the workflow and integration into the 3D Apps/Exporters!?
looks for me like having some boxes scattered around in the sky with some "cloud" shader onto it,
that does the work for soft falloff to the edges of the box...
Anyway, more examples please ^^
polygonmanufaktur.de
Re: Experiments..
Hi Zom-B,
It doesn't really speed up SSS, as it only speeds up participating media inside a medium of IOR 1, while all realistic SSS media have IOR > 1.
I'm not entirely sure how the integration and workflow will work.
For stuff like explosions, smoke etc.., manually placing a 'proxy' bounding object which then uses grid-based density data should work pretty well.
Automatic cloud generation is a bit trickier and will probably require a custom interface / parameters.
It doesn't really speed up SSS, as it only speeds up participating media inside a medium of IOR 1, while all realistic SSS media have IOR > 1.
I'm not entirely sure how the integration and workflow will work.
For stuff like explosions, smoke etc.., manually placing a 'proxy' bounding object which then uses grid-based density data should work pretty well.
Automatic cloud generation is a bit trickier and will probably require a custom interface / parameters.
Re: Experiments..
Good to know anywayOnoSendai wrote:It doesn't really speed up SSS, as it only speeds up participating media inside a medium of IOR 1, while all realistic SSS media have IOR > 1.
A misty room with sunrays visible from window will benefit, thats quite good news!
What kind of dataset(s) will be supported? Interesting to see if some compatibility to density datasets generated by some existing Apps will be given.OnoSendai wrote:For stuff like explosions, smoke etc.., manually placing a 'proxy' bounding object which then uses grid-based density data should work pretty well.
Yes, this will be obviously a job for the Exporter writers, maybe by some coded "Cloud shader" that feeds parameters easily into ISL. (like ISL Shaders for noise, gradient etc. in Cindigo for example)OnoSendai wrote:Automatic cloud generation is a bit trickier and will probably require a custom interface / parameters.
polygonmanufaktur.de
Re: Experiments..
We'll probably make our own volumetric format, and then publish the specs.Zom-B wrote: What kind of dataset(s) will be supported? Interesting to see if some compatibility to density datasets generated by some existing Apps will be given.
Re: Experiments..
Looks really nice Ono! Still, there is something about it, I can't put my finger on.. The top part of the smoke seems to be suddenly stop along a very noticeable line...OnoSendai wrote:Bumping this ancient but fun thread...
Volumetric scattering is now way more efficient.
Some new results: see attached pic, rendered in ~10 mins or so on a laptop.
Expect some new sky + cloud experiments soon.
- Borgleader
- Posts: 2149
- Joined: Mon Jun 16, 2008 10:48 am
Re: Experiments..
Well...it's possible that's the end of it's domain. I'm applying the concept from Blender, where the domain is defined by a cube and the simulation is calculated only for it's volume so if the smoke is "thick" at the edge of the domain, it abruptly stops and gives something similar to what we see here.
benn hired a mercenary to kill my sig...
Re: Experiments..
Well, imo it detracts from the realism. Im sure we've all seen that clouds don't have hard edges..Borgleader wrote:Well...it's possible that's the end of it's domain. I'm applying the concept from Blender, where the domain is defined by a cube and the simulation is calculated only for it's volume so if the smoke is "thick" at the edge of the domain, it abruptly stops and gives something similar to what we see here.
Re: Experiments..
Hi Neo0,
it's just because the density shader is contained inside a cube. A real cloud shader would gradually reduce density near the edge of the containing object, so you wouldn't get hard edges.
it's just because the density shader is contained inside a cube. A real cloud shader would gradually reduce density near the edge of the containing object, so you wouldn't get hard edges.
Re: Experiments..
hmmm... maybe the easiest way to go.... coding wise! But practically I see difficulties in the usage of this.OnoSendai wrote:We'll probably make our own volumetric format, and then publish the specs.
Only Blender with an integrated smoke fire volume physic may benefit, if the exporter writer does a code some on Export Converter to Indigos native format.
Other 3D Apps have maybe multiple 3rd Party plugins for doing so (I know 3 for C4D!). Providing support for multiple of such simulation plugins is quite illusionary.
But otherwise (if existing), the export to a genaral Volume Format and conversion into Indigos native format would be the way to go, providing one methode that works with multiple volume generators!
Also simple particle simulations could be converted (interpolated) to something quite interesting
We'll see how good this can be integrated into the 3D apps / Exporters! There is still a good bunch of more frequently used techniques then physically calculated smoke clouds of explosions etc. like clouds on the sky fog etc.
polygonmanufaktur.de
Re: Experiments..
I will most certainly attempt this at some point in the future, providing blender's API is kind to meZom-B wrote:Only Blender with an integrated smoke fire volume physic may benefit, if the exporter writer does a code some on Export Converter to Indigos native format.
Re: Experiments..
Blender will prolly be the first to implement the Indigo grid data format, but Maya will not miss its opportunity I'm sure.Only Blender with an integrated smoke fire volume physic may benefit
Limit surface + density grid = yummy
And as you said, Zom-B, even Cinema can do sim in multiple ways. But yeah, point cloud is another fun, and perhaps more accessible thing, through particles. I guess Ono is just focusing on shading concerns for now...
Very cool !
obsolete asset
Who is online
Users browsing this forum: No registered users and 22 guests