Blender Stack Exchange is a question and answer site for people who use Blender to create 3D graphics, animations, or games. It only takes a minute to sign up.
Sign up to join this community
I understand that, when possible, objects should be instanced from a master object for optimal speed in both rendering and memory consumption.
Does this hold true for the Mirror modifier as well? If I have a mesh that is symmetrical on an axis, do I gain any statistically relevant speed (interface FPS or pre-render calculations) by using a Mirror modifier instead of the full mesh?
This is especially relevant when using subsurf modifiers over level 3 or 4, for example...
No, as far as I know there are no optimizations in place for constructive modifiers like mirror or array. You do get some benefits in terms of workflow since you only need to do "Half the editing" and you may also get a benefit in file size, since only one half of the mesh is stored on disk.
Other than that I don't think there are performance benefits from those modifiers since depending on the settings the meshes they generate may not even be totally identical.
If you are looking to use a mirror modifier method with the benefit of the Alt+D method (from the links of Mr Zak in the comments), then consider using the Double Sided option in the Data Panel >> Normals section.
Then just like you would have to have your mirror origin on the mirror modifier, you can scale to the negative value (on whatever axis) after the Alt+D operation, with your manipulator set to the 3D cursor origin, and then Parent this duplicated mesh to the original or ideally, parent both to an empty.
