I have a rubber duck i created in Blender, to reduce the file size i have decided that when objects are symmetrical and if they are small, i would save the mesh as a half mesh, copy this in Unity and set the scale to -1 then snap together. This all works well but my questions is: Should the half mesh be a prefab before i join these two meshes or is it ok to import a mesh copy it and then create a prefab from these two or should i have two prefabs, one of the half duck and one of the whole duck containing two half duck prefabs.
You shouldn’t do that, you can run into weird issues with negatively scaled objects (I did) and your method only saves memory at a slight performance cost, usually you want the opposite. Also how big is your rubber duck’s poly count that made you consider this in the first place? Seems to me like you are going at this the wrong way.
3 Likes
You shouldn’t do it, because it’ll invert its face orientation and will result in weird lighting issues.
Also, unless it is a trillion polygons duck, trying to “save faces” this way simply wastes your time and achieves nothing useful.
GPU will be happier drawing one object instead of two.
5 Likes
Are the weird issues you get only on interactive objects like with rigid bodies etc or would you say never to use negative scaling at? and thanks for your answer its good to know early on.
Considering the amount of time I’ve spent chasing bugs triggered by this in other things I would say it’s better to just never do it. In cases where you need to mirror a mesh, I’d rather write a script that rebuilds a mirrored copy of that mesh with normal 1/1/1 scaling on the transform.
1 Like
Very Informative , Thank you everyone!