We’ve always been asking for a way to union not only BaseParts, but MeshParts too. But did you know it’s always been possible? Today, I’m going to show you a way to turn multiple meshes into one, that not a lot of people know of.
Firstly, let’s say you have a bunch of MeshParts that you want to combine into one for whatever reason:
Now import the mesh via the Asset Manager with the configurations as seen below and click Appy All:
Insert the mesh back into the game and align it with the previous MeshParts.
That’s it! You can delete the old parts now. If you already knew this, well I’m sorry, but I believe there’s a lot of people which will find this helpful.
Yeah I’m pretty sure you’re right. I only use this for textureless meshes, so I didn’t think about that sooner.
But I think if you import the mesh into blender then you could just join (CTRL + J) them and/or somehow fix the texture before importing it back to Roblox. (?)
Exporting can cause issues with broken meshes, etc. Plus there is little benefit to actually combining meshes since the same amount of triangles exist either way. I wouldn’t recommend doing this unless you are seriously desperate for the extra frame of performance for a ton of extra work.
There’s no point my in my opinion. It doesn’t add any extra benefit by doing this, as you can achieve the same thing by just grouping the meshes together.
I agree. The triangles can be decimated in blender but I don’t think it’d look that way. The reason I posted this is, or at least I think it works that way, when you have a lot of parts, it impacts performance, not in the rendering way, but memory. And it also looks good while editing the game haha.
Well, the same can be said for normal parts, but Unioning them is usually better. It doesn’t change the Tri count unless there’s too much triangles.
Also,
The reason I posted this is, or at least I think it works that way, when you have a lot of parts, it impacts performance, not in the rendering way, but memory. And it also looks good while editing the game haha.
One problem I see with this is it interferes with the engine’s instancing behavior. Having many meshes can add to the memory footprint, even if they share geometry
Having lots of parts doesn’t necessarily affect memory if they share the and meshid, like I said in my post above, the engine will instance the mesh and each part will only have their own cosmetic property footprint(size, color, etc) the vertex data, which takes the lion’s share of the memory footprint, will only be stored in memory once. So unless you can reduce the non visible triangle count of an assembly by a large magnitude it probably isn’t worth it performance wise
