Yeah. It was a worth it for me. Anyways, I can't seem to get Material8 working, it's complaining about undefined type but I made sure I included all the headers. I'm going to leave this alone for now, maybe try again this weekend.

Good work so far! Maybe it'll be worth adding this to PolyVox when it's finished?

And sorry for hijacking your thread, but it seems that there are no additional cubes rendered beneath the surface of your world. I saw that in my application every volume gets rendered even though I assumed Ogre would sort those out if they can't be seen. How did you solve that?

Are you using the CubicSurfaceExtractor? I am not using Ogre, and only the surface it's extracted, everything under it is just air.

Actually I was referring to beyzend
If I remember correctly you didn't post screenshots of your landscape. But I do remember that you were using OpenGL And yes I'm using the CubicSurfaceExtractor.

Oh that. I disabled generating my terrain from 3D noise, instead I only build a 2D surface and currently I'm assuming everything is rock below "ocean floor". So no caves. I will have eventually fix this in the future. Plus I plan to implement a simple occlusion culling algorithm in the future.

Oh man this mean I need to implement caving again to test my system...Oops.

Ogre will cull meshes against the view frustrum so that they won't get drawn if they are behind you or to the side. But it does not help in the case where they are both in front of you but one is occluding the other. In this case both meshes get rendered, but the depth buffer ensures the visual result is correct.

Thank you! Didn't understand this before... This is certainly good to know

Is this topic dead? It looked like one of the most promising additions to polyvox other then built in picking. Doing it in two dimensions seems like it would be a lot more tricky... are there publicly posted algorithms for doing this? Inventing your own seems like it would be fairly painful. I guess I can always look it up, but yea, I'd like to use RLE for sure.

Beyzend already posted code for a RLE surface extractor.
From what I read on his blog recently he wanted to create a surface extractor which merges faces on all three axes. If that's what you are referring to you have to bug him to work on it

As mentioned already, a surface extractor which merges the faces could indeed be useful. I especially think it could be useful for physics, and of course it would lower the memory requirements of the mesh.

PolyVox does already have some capabilites in this regard but currently it only works on the Marching Cubes surface extractor, not the cubic one. It works differently from beyzend's approach in that it generates the full detail mesh and then merges the faces as a seperate step.

I do intend to brush this up and make it work on the cubic surface. Until then it's hard to say how it compares to beyzend's aproach.