PDA

View Full Version : Instances & GI


corella02
02-24-2008, 05:13 PM
The photon map seems to ingnore instances in MSP3.
Can anyone confirm that?

Also, metanurbs still doesnt handle triangles? :sad:

C.

Wegg
02-24-2008, 11:44 PM
The photon map seems to ingnore instances in MSP3.
Can anyone confirm that?

Also, metanurbs still doesnt handle triangles? :sad:

C.

I have used instancing and GI a lot and never had any problems. Can you take a screen shot of your render settings and post them? Or a test scene? Maybe that will help.

Metanurbs still hasn't changed. You still need to model in Quads and avoid more than 5 edges joining at 1 point. Its "old school" modeling but it works.

corella02
02-25-2008, 11:52 AM
Hi Wegg,
here is another example. I made a new scene, four boxes one sphere. I copied the sphere in the first case, al spheres share one material. As you can see, arund every sphere i get a more or less nice color bleeding as expected. than i deleted the copies and instanciated the sphere. Now only the original has the color bleeding.
By the way, the photon radius minislider defaults to 1, also the GI intensity minislider, which makes both of them useless.

Ulven
02-25-2008, 12:27 PM
Looks like there might be a bug with the photon map GI.
Meanwhile, try setting it to MonteCarlo, and then click the GI noise reduction. It will then use a sampled method for the GI. It will work with instances. There's a very quick runthrough in the General Rendering tutorial, but I haven't had time to make a full GI tutorial.

corella02
02-25-2008, 06:22 PM
Ulven, thanks for the Video. The GI-Noisereduction feature seems to work well if one knows how to use the parameters. Now heres another one: I tried to combine Caustics and montecarlo with GI-Nr. and i get this.

C.

corella02
02-27-2008, 12:39 PM
Metanurbs still hasn't changed. You still need to model in Quads and avoid more than 5 edges joining at 1 point. Its "old school" modeling but it works.

Thatīs just old school for Lightwave users. And some of the people i get models from arenīt.


C.

kursad_pileksuz
02-27-2008, 04:36 PM
good point

And some of the people i get models from arenīt.C.

kursad_pileksuz
02-27-2008, 04:40 PM
It looks like your distant light is using shadow map, not raytrace. I just checked your scene that you have attached


Ulven, thanks for the Video. The GI-Noisereduction feature seems to work well if one knows how to use the parameters. Now heres another one: I tried to combine Caustics and montecarlo with GI-Nr. and i get this.

C.

corella02
02-27-2008, 04:53 PM
Thanks kursad, i didnīt expect the solution in the shadow setting :).


C.

CGTalk Moderation
02-27-2008, 04:53 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.