In my last model I spent a fair bit of time making my model so that it was “comprised” of one mesh. What I mean by this is when making the barrel of the gun I didn’t just inset a cylinder on the front of the gun and join the two but rather deleted the front of the gun, aligned the barrel with the gun, deleted the front of the gun, and proceeded to join every vertex individually.
After watching a few speed builds I realized that a lot of people don’t do this and after some refection realized that this method may actually make the model “easier to run” as in many cases there will not be as many triangles to break up a surface.
Are there advantages to taking the extra time to remove the unseen portions of a mesh, or am I just wasting my time in doing so. (It might make it easier to texture depending on the mythology, or it might just make the UV map harder to read.)
While I may be a newb, I am not a total newb so this is assuming that the model will not be rigged (aka a gun or tool that is not going to move). I don’t know how to rig myself, but I can imagine unattached portions of a model playing havoc with the rig.
Depends on what it is. Human skin is comprised of one part. Machines are always many parts. Unless it’s really advanced science. Then anything goes.
If appearances are all that matter then why should it all be one piece? On the other hand a rig needs to move within physics to some degree.
For games and animations, you generally want to keep the mesh as simple as possible in order to reduce the calculations involved in rendering each frame. It’s not so important with simple scenes, but once you start adding in a bunch of complex objects and lots of textures, computer resources get eaten up fast and things begin to run slower, so it’s a good idea to plan ahead and keep things as ‘streamlined’ as possible. It isn’t unusual to initially build something with just primitive shapes joined together, and then later on go back and make improvements to the mesh or completely replace it.
Alright I may be a bit late to the game on this question but I have a few opinions on this subject
First it depends largely on the meshes’ intended use. Are you using using the mesh for a game (with poly budgets) or are you using the model for a pre-rendered movie? Secondly it comes down to rigging issues down the line. Floating geometry can come off… well, out of sync or “floaty” if not rigged with care on organic models. Also you can introduce clipping issues with multiple meshes if not modeled and rigged with care. BUT with that said you can sometimes save both time and polygons just having your mesh being several pieces.
So let me get a little more detailed. I think hard edge models like guns, cars and machines really are fines with floating geometry due to their ease in rigging and the headache you save having to merge everything. this is because the parts of the objects are typically rigged to to bones with 100% influence… Or in other words they don’t require complex rigging where parts of the mesh deform under the influence of several bones. So visual errors are almost non existent in hard edge models… Also for the record… Things like guns and machinery have many parts so why shouldn’t you model them as separate parts when it looks right and is convenient?
For organic models like people and their clothing… The less pieces you have floating around… The easier it is to rig without weirdness or clipping… So like for an example would be modeling a pocket on the chest of a shirt would be better to be merged to the chest mesh rather than having it float on the model as rigging the pocket as a floating mesh could make it stretch in an odd manner and it could potentially clip into the chest. With that said if you are under a really tight poly budget you may have to rely on floating geometry like chest pockets and endure the headache of careful rigging to make it look good.
Clipping is a big reason to merge models… It is common practice to first make a body mesh and then cloth them and then delete all the skin the player won’t see under the mesh and then make the clothing and body one mesh to ovoid the body mesh clipping outside the clothing mesh. Similarly… Unless you plan on having a scene where the character is stripping his gun down to it’s components… You don’t need to model all the internal bits in a gun that aren’t seen from the outside.
TL;DR: Use floating geometry to save on geometry, or on hard edge models or to save time /headaches… But realize that multiple meshes make rigging organic models tricky. Not an issue if you are a wizard rigger.