What EIAS need to have a volumetric rendering?

#9

Harry, we learned this great plug-in in all details. Be sure, a role of host is only auxiliary there :slight_smile:

#10

Psyclone is a volumetric effect that does ray-marching internally but it has limitations. First, some sort of geometry must be present in order for the shader to be called by Camera. Psyclone builds a cube on which to do the rendering but you’re not aware that it’s there. The resulting image is sort of projected onto the faces of this cube.

The second problem is that it doesn’t cast really accurate shadows because image is really being projected onto the faces of the cube. For example, if you turned on light rays and tried to shine the light from the top of the cloud pointing down and viewed the scene from the side, you’d see the rays begin from some invisible bounding box.

Psyclone also doesn’t respond to lights that exist inside the volume. It’s really meant to do clouds which are usually only illuminated by the sun which is very far away.

Technically, a lot of shaders exist in 3D space e.g. anything that uses Perlin Noise. If there were a way of creating volumes that are ray-marched and the ray-marcher called a shader, neat things would be possible.

#11

We don’t share the opinion about neat things possibility.

What rational volumes the host should provide? Gizmo shapes like in Max (simple primitives - sphere, cube, cylinder etc.)? Not a big help for developer. Note also that analyze inside/outside gizmo is a work for max plug-in, not for max host. Custom shapes (any polygon object)? Can be implemented with EI shader (though, yes, need to call RT to detect volume)

Second, Perlin noise and Volumetric are not good friends, sure. It becomes obvious if Perlin noise is used as a layer of clouds (as a shape - in c4d etc). The range of interesting volumetruc effects is very narrow, we’ve no illusions that ray-marching + shader would give even something approximately usable. Also straightforward ray-marching produces very poor and sharp self-shadows

#12

Hi,

What we need to have a plug-in like Shave and Hair cut?
Because its really dificult to have millions of hairs using geometry.
www.joealter.com

Thankkkkkkksss

Tomas Egger

#13

$, Tomas, $

#14

How much money?

#15

Sorry, can’t answer - that question is not for public discussion

#16

Well Joe’s Shave, I’m assuming, has a cost right? You have to license the technology. That’s what I was wondering Igors… Not your specific EIAS developers cost. Though I would be intrigued to know what that is…

privately.

Don’t get any ideas folks… I’m probably not that rich. lol

#17

How much? :drool: I remember Brian was in conversations with Joe Alter about licensing his technology… i suposse was too much for the posible EiAS market and nothing happened :banghead:

FelixCat

#18

Could we not build somthing like shave from “scratch” ?, or add to-improve existing tools ?, FiberForge ?

And if shave costs $399 then thats what i’d expect to pay for simlar tools/results.

Reuben

#19

IMHO EIAS has a lot of what normally should be in plug-ins/shaders (examples: blurred reflections, normal maps, all material’s edge settings, morph, skins etc). But not hair/fur - that should be in host.

#20

Given the intensity of such calculations… yah… you’re probably right.

#21

Again, I’m certainly no programer, but if someone’s working on a cloth plugin for EIAS, isn’t that just about as difficult as creating fur and hair? I mean Placer Deposit looks pretty good to me. Would a soft body simulator for EIAS make it possible to do hair too? – Kind of like taking what Placer Deposit does and making it flow?

Am I being naive here?

Thanks!

#22

speaking of volumetric effects, one of the most useful additions to EIAS would be also some kind of volumetric instancing like this:

http://www.happy-digital.com/instance.asp

have a look at the samples.

this kind of plugin is very interesting because ‘natural looking’ scene elements (like grass, fur, pebbles, trees, rocks, clouds, ehm …even teapots!) could be rendered in enormous quantities without actually handling geometry. rendertime increase for scenes with millions of instances should be much less dramatic than with real geometry instancing.

i did and i’m actually doing some projects with landscaping elements like flowers, grass and trees. in such scenes, even with camera a limit when operating with real geometry is fast reached.

is there any hope? in theory, with the current SDK, is such a plugin possible? with shadows, transparencies and reflections?

cheers

markus

#23

Well, maybe not naive, but too optimistic, really :slight_smile:

#24

No, Markus, we see no ability to implement such EI plug-in

#25

ok, thanks for dropping me back into reality :slight_smile:

#26

why can’t that kind of plugin be implemented? Problems with EI as a host? API limits? Could it be implemented if it wasn’t a plug but part of EIAS?

Thanks

#27

Markus isn’t the only dreamer here! :rolleyes:

#28

Yes, here is the place where EI shader/plug is helpless without host’s support. We’ve experimented on this way several years ago. The attachments are our virtual grass attempts (hmm… looks naive a bit :rolleyes:). The render time is quite acceptable (no RT though), but instead we’ve had 2 serious problems that we could not solved.

  1. Yes, without polygons we can fill unlimited areas with a grass. But we’ve lost a lot together with polygon/geometry limitations. And the material is the first and most painful. The shader should do all (absolute all) itself: color, texture, specular, bump etc. - no any side help cause all is linked to polygons we’ve not. There is no way in to force Camera to render a polygon “on fly” - the polygon should be passed to render engine and saved.

  2. Also very unpleasant: AA (maybe same problem is also in plug-in that Markus posted a link). As you know, Camera shades edges very carefully - a several “sub-pixels” are mixed together to obtain a smooth result. But with our “virtual geometry” the render cannot detect where more or less sub-pixels required. So, jaggy grass edges :sad:

And, hmm… when we see a next such cool thing (as this LW plug-in), we usually do following: first we try to forget about it. Then (next day) we open demos/samples again and see them (and their render quality) whith calm emotions. It’s veeery helpful :wink: