I still don’t understand why they don’t make meshpart editor instead of messing with intersections. Intersections are only needed in games like destrucion physic or lumber axe etc.
CSG and mesh editing are very different workflows.
Depending how much experience you have with mesh editing you may or may not be painfully aware but while mesh editing makes some things much easier (deformation) it makes other things a lot harder too (keeping your mesh topology sane).
I find working with CSG waaay easier than any 3D modeling software I have used (Blender, 3DS Max, etc)
It’s super easy just to make something with a few intersections and combination of parts and to be able to see the results live in studio immediately. CSG is something I have been using for more than 7 years now
Chat, is this a W?
Yes. Yes it is - Dave give this team a big W raise
The difference from before is astronomical.
Here’s a good example of test benching this with a Sierpinski triangle pyramid fractal.
I was just working this before this came out for beta, only managed a few thousand instances before crashing.
Here, I am running just fine with over 390,000 unions (with exception to any hardware limitations).
That’s awesome! Also, do you think you may release this one day?
Sweet my models are only made with unions this will greatly benefit me; I’d like to see an update on the lighting system because it looks a bit outdate now.
are you using future/shadowmap?
Lighting looks fine if you edit lighting settings and add post processing effects
Just a test nothing special. Most people wouldn’t visit this, too memory intensive.
Counting down the minutes and hours until CSG for meshes is put into beta, great work team, you guys are definitely up there in the most impressive and consistent sets of updates released for one ‘system’ (rip particles, flipbooks and grass)
I meant as a dev resource, it looks like a beautiful way I’d like to test the new optimizations
But now that I think about it, players may like it too. Seems like a grander version of Lag Test
Nice, love seeing the Solid Modeling getting some love and attention. This and that upgrade with only rendering what’s in view will greatly help performance. Keep up the good work. Thanks for the update in regard to SMO on meshes as well, that’s going to be wild when it comes out.
changing the axis size will only double the size and is a short term solution. there are many solutions you can implement yourself along with not making a map that big. hope this helps
it makes other things a lot harder too (keeping your mesh topology sane).
I don’t understand why it would be more difficult? You just choose the dot and then moves this. That is much easier and quicker then intersecting.
Thank you for this update! Really love the new performance improvements that are coming up!
Yeah i know, right now im mostly banking on the fact that i keep making the UI parts larger the further you are from origin. It’s not a good solution but it works for now. Though yea i should make the map a bit smaller. But as ive said in the post, the UI’s shaking can occur pretty close to origin too! Player’s finding themselves 3k or more studs away from origin these days really isn’t very uncommon!
The trouble is having the right dot to move, that’s the topology aspect.
That is, there may not be a vertex to move where you want to move something, and adding vertices in the wrong way can make the mesh harder to do further edits to. Working in a way where you naturally end up with vertices and edges in favorable locations for the edits you want to do is not easy and something that takes experience to learn.
One of CSG’s advantages is that the system handles the mesh topology for you rather than you having to explicitly think about it.
Solid Modeling has a few advantages over direct mesh modeling. As @tnavarts said in some sense its simpler than direct mesh editing, because you don’t have to worry about the topology. It also only ever produces objects that have a well defined inside and outside, which guarantees that they can be simulated accurately.
That being said, we are also working hard on EditableMesh and EditableImage and we hope the community will use it to make other useful modeling tools with different tradeoffs.
Will you be overhauling the process on how solid modelling works when editable meshes are fully released? Surely it will make faster and easier combing / separating faces and vertexes.
We want to make the interop between solid models and editable meshes as seamless as possible, so folks can switch back and forth between vertex/face editing and solid modeling depending on what they need.