Hi, is it important to reduce the number of seams, in order to optimise a mesh. or 3 islands equal more or less to 1 island in term of ressources. does it have a noticeable difference
It doubles vertices along each seam you place.
Rule of thumb is that you make as few cuts as needed to get a good result with the minimal possible strechting. But it often is a balacing act between reducing stretch and reducing seams and keeping the layout readable and flowing ( the latter is less important these days unless you still texture in PS or another 2D editor exclusively).
If it is for games, having a seam is also often determined if the geo needed a hardedge along the section. If you had a hardedge there, you can also place a seam there. It wont double vert count again, the hardedge already doubled it.
Performance wise: I wouldnt go paranoid about it, just use reason and understand principles.
There was a post in like 2009 of an Epic employee on Polycount who investigated the matter and shaved a few million polygons out of levels only to discover that he couldnt measure any performance difference at all. And that was in like 2009.
thanks. i’m not anglo-saxon thus i’m not sure i understood your last sentence correctly. you mean an Epic employee made the test back in the years, with, and without lot of seams. to notice there wasn’t finally no significant difference. is that correct.
i’m a kind of optimisation obssessed dev.
. so any tip regarding the optimisation will interest me. even if i started Blender only a year ago, at this point, without having made any organic model, i find the lot of islands technique much, much faster. even with V. for buildings who only use seamless textures, it’s just ‘scale, and next’. the only interest i found to seams so far, it’s to create continuous, uniform surfaces. like for example curved doorsteps, or roofs.
so this is why i ask. by doing only islands you cut the worlflow by 2 easy. but as i plan to make hundreds of buildings, i want to know if i’m doing well before going any further