Hey Guys, I was recreating this effect from Konstantine Magnus’s youtube channel:
Totally unneeded to watch but I’ll post it here if anyone is interested, always good to credit.
I got it working I think as you can see from the video I posted, but there is sort of an issue I am confused about. When lowering the voxel size on the mesh to density grid all of a sudden has the effect failing. I thought maybe this would be because the density grid looses as bit of space, as it becomes closer to the mesh and not some fog generally fitting it’s form, but at the voxel size of 0.06 as it currently is at it’s not really different much different from a voxel size any lower.
There is probably something else at work at work but I just don’t have the volume knowledge to know. I suppose I am using subtract instead of multiply like in his video, but I found better results with subtract.
That video has 1800 views and I’ve never seen it. This feels like a pretty mean response
I had a look at it and it certainly uses the same method I did just with his own presets. I would like to keep in mind that math is just math, this fellow followed the math and method of Konstantine Magnus, and so did I. This is not a surprising result.
If you look at some of my other posts on Bartists I often look at houdini tutorials, especially from Entagma, in order to learn. In this case I got it working, but I have a question about something that seems strange to me that I don’t understand. That is what learning is about.
I barely understand what may be mean about sharing some resource i found and seems also following the same source as yours did. I thought that it might help you since i did found this one recreating houdini tutorial some time ago (but never hat the time to really look closely into it) and thought that this is interesteing. Therefore the :
But if this does not help than this is simply unfortunate.
(And i also don’t get what the view count or that you haven’t seen this has to do with this..)
Tone is hard to get over the internet and I somehow thought this was an insinuation like "sure you replicated it, I’m sure you didn’t just look at this video. I had just woken up and I guess I felt a bit sensitive about it because I try pretty hard to do things on my own in order to learn.
Ellipses and a winky face is a killer combo in my defense. But now I see I misunderstood and was really mean in return. I must apologize, that’s not proper of me to assume the worst.
However if we can put that aside, as I genuinely feel bad about it.
I actually went to the lengths to download this guys premade packages lol. But it seems to have the same issues as mine. The smaller the voxel count, the more it doesn’t work. It’s very confusing because the mask being created here with the geometry proximity and map range nodes should be the same regardless of how many voxels there are, right?
Voxel size is basically resolution of the grid… Changing the values in that node seems to affect points distributed within the grid. Distribution of points is number per unit volume. But I don’t see how this affects the rest of your nodes.
My suspicion is on Geometry Proximity and the map range node after it. Maybe the map range needs adjustment to reflect the new voxel sizes?? This is just a running theory, Will be interesting to break this down to figure out.
Yes I agree I am curious about that same area and most of my poking around has been in that area. Something I wish we could do is use the field min max node or Attribute statistics node to normalize the distance result, but they don’t work on volumes currently it would seem.
Ok I figured it out! It’s because the initial density grid was not voxelized. Because of this when the gradient width is low it will collapse many voxels into tiles. When the voxel count was higher I think it forced more voxels to read gradients and kept them from being tiles. When you voxelize the initial grid all the issues go away and you can work at any gradient width and voxel count!
If I’m understanding the core logic correctly (I haven’t seen the video yet) : The distributed points delete the “Grid” around it…Then a “Grid gradient” is used to get a direction, which offsets the points. From the new location more of the grid is deleted and repeat…And capturing the points at each frame is what gives the foundation to create the tubes
What I see every time it breaks is that, after going through a certain iterations, some points don’t get a new offset. Seemingly the gradient goes to 0, i’m guessing when it it finds no more changes in value, as you said when voxels collapse into tiles could be the cause.
That basically correct, the only thing I would correct is that technically nothing is deleted. Voxels aren’t deleted but their value is changed to 0. As long as densities above 0 are left, and the mask isn’t too wide, the points should continue to find paths to move to making those tubes.
And yea I think the inside just got condensed into tiles without voxelization at certain voxel sizes, so the inside was mostly useless info.
