PDA

View Full Version : Instances with Skin Material vs Full object



mummyman
04-25-2016, 01:47 PM
I was wondering if anyone knows if this will be fixed in Lightwave Next... I have an object with the Skin Material applied, and it's instanced. Lighting does not look great. Wierd shadows appear, giving a very blocky effect in the background of the image. The bottom render, is made of all objects cloned in Modeler and saved as 1 object. I thought this was working properly. But apparently not. Lino? It'd be nice for the instances to render smoothly under certain lighting / surfacing conditions. The bumps on the instances don't matter in my image. That's just a result of world coordinates. Anyone?

133557

spherical
04-25-2016, 04:05 PM
It would be best to have a direct comparison to determine what is and isn't caused by bump. Difficult to tell without a scene to examine. More information would be great. Using GI? What kind of lights?

mummyman
04-25-2016, 06:22 PM
It's not the bump. They are essentially the same surface. It's the way instanced objects are rendered with Skin / Subsurface on them compared to full geo. No GI. Just a strong back Area Light, a front fill (distant) and a front spec-hit light. The scenes are identical in every way except a poly object vs a plane with instances of the same single cuboid object.

lertola2
04-25-2016, 06:37 PM
Is it the same exact texture on both? The instances look like they have the bump cranked up very high compared to the other. That is probably what is causing the choppy look on top of the instance version. But maybe the skin shader does not work with instances. I don't see that purple glow on the bottom of the instanced version.

spherical
04-25-2016, 06:55 PM
It's not the bump.

Great! Then show it to us with bump turned off. I notice some odd brightness in the gap between a couple of the foreground cuboids. Wondering what is causing that, too.

mummyman
04-25-2016, 07:17 PM
Well, I can do both without bumps, but the shape isn't exact as far as positions. I"d have to freeze the instances, which I didn't do/try yet. But cloning them on the same shape in modeler is pretty darn close...which is most likely (I'm 99 percent sure) will yield the same results

lertola2
04-25-2016, 07:19 PM
I just made a quick test and I wouldn't have thought anything was wrong if you had not brought it up. In the attached image the object version is on top and the instanced version is on the bottom. I do see a little difference in how the bump is applied. Can you post a scene that demonstrates the problem?

http://forums.newtek.com/attachment.php?attachmentid=133558&d=1461633300
133558

mummyman
04-25-2016, 07:40 PM
133561I'm not worried so much about the bumps...but the shadows. The light is doing something wonky to the instances. Almost like every other is darker, then lighter in my sample. Here's my no-Bump comparison. Hmmm.. have to check the scaling of my objects too.

spherical
04-25-2016, 09:15 PM
OK. That's better to be able to isolate the issue. While there is some difference in a few of the foreground cuboid texturing, it appears that the lighting anomaly is most pronounced on the far half.

mummyman
04-26-2016, 07:17 AM
OK. That's better to be able to isolate the issue. While there is some difference in a few of the foreground cuboid texturing, it appears that the lighting anomaly is most pronounced on the far half.

Sure does. It's almost like the light is still trying to work on where the instance is at 0,0,0 instead of where its placed. Got to be a glitch. Oh crap.. I rendered out a test sequence. Horrible flashing. It's limited to the background for sure. I have a falloff from front to back so the surface "swells" with a displacement texture moving up through the main floor plane (object with instances on it) It definitely doesn't like that. Damn. Should work, but it doesn't.

mummyman
04-26-2016, 07:22 AM
I just made a quick test and I wouldn't have thought anything was wrong if you had not brought it up. In the attached image the object version is on top and the instanced version is on the bottom. I do see a little difference in how the bump is applied. Can you post a scene that demonstrates the problem?

http://forums.newtek.com/attachment.php?attachmentid=133558&d=1461633300
133558

I appreciate you trying to replicate this. I don't believe I'm going crazy... well...maybe a little