yeah i’m on Blender 4.2 now without any addons, it’ll still just remove the trees. I’ll see if i can upload a .blend.
Here’s a .blend with all the old trees + the new tree (called crown merged i think it was): https://www.dropbox.com/scl/fi/1gt60ia0y1ep49kcxgvzm/trees.blend?rlkey=7f05d3ea1ongn1twldl4qt9wv&st=slah6ke5&dl=0
Your problem is, that all those meshes are merged into a single object. So these 1000 trees are combined in ONE big mesh . And if you change that to the new tree whose mesh data is just that of a single tree you swap these 1000 meshes to just a single one.
And as the 1000 trees have their mesh location at 0,0,0 it even falls together with the position of the new tree. So it looks like they all just vanished, while there are in fact now two at the same position.
In principle you can separate them and set their origin to their center, but some trees consist of unconnected branches. So it wont be just two clicks to fix this.
So essentially blender did exactly this: replaced the wood
with one new tree
…
Trying separate by loose part and additional set origin to center of mass (or volume) at least get a position for every “wood-object”. Here’s small exampl (without leafs and i pickes some “trunks” )
Merging the wood vertices by 1m helps a bit but is not reallly a good solition.
So the quesion is:
How could these “clusters” of vertices get a better edit workflow ? Maybe selecting the branches
somehow by volume and delete them ? Or is there any way to “detect” the rotation and scalling of the different trees within the big mesh or while already separated… to be used as base for instantiation ??
( by the way zengine …hmmm Gothic the game ?? )
…for the selection different sizes maybe something like this could help:
Thanks! Yes that explains a lot. Now when i understand the issue i can ask ChatGPT to make a script for me, and i can prompt in a way that describes the issue. I’ve had so much luck with that for various very advanced tasks and just today an even better version is out so it’ll likely work out somehow.
Yeah i love the Gothic games This is the mod so far:
I hope to be able to make the rather low poly trees on the world mesh better. I’ve already replaced the separate flora meshes that the game uses.
Thank you! That explains why it didn’t work, or worked in the way it did rather.
I lost track of all the “changed” versions… but G1 and G2 are the only real ones…
Totally agree
ChatGPT’s solution was the same as OkiDoki’s, pretty much. But i was thinking maybe merging is a solution, so i asked it, and it came up with a script to merge trunk+branches by distance.
So yeah, as long as you can prompt it in a good way (you have a good understanding of the issue, and not just a vague idea) it can help you with pretty much anything.
It may be interesting how such a script looks like and what it can do so far… or more precisely i am interested in this.
I’ve attached the script.
I’ve now tried it out. It kind of works. For some reason it will not connect the largest branch that seems to go into the trunk. However i can up the distance and it will eventually connect it, but this large distance will sometimes connect another entire trunk, not optimal, but not too bad. i can work with that.
While i complained about this it came up with another solution though, to instead use Geometry Nodes, and it made a script for it. I’ve not tried it yet.
merge by distance.txt (1.4 KB)
That’s indeed a rough start… comparison of volumes came to my mind to make the distance parameter somekind of relative… but there is some suspense on my side how this could evolve.