Hello, does anyone know how to get a approx guess on how many tris your mesh can have before your better off trying to fake details in via texture?
Right now my case is sitting around a 4k texture set vs 45,000 tris.
with the 45k tris, i can use existing seamless mats cause the detail would be modeled in.
But if I use a hand painted 4k texture I can add all those details in but would look faked cause of the issue of well its faked in lol.
But I am not sure if this is a lot for a type of mesh that is a building part you use to snap around making bases.
If anyone knows of a way to try and get a debug or stat from unreal to compare, please explain or drop a link to some documentation on how.
If that is not possible, someones thoughts on my case would be nice.
Well optimization is not about a number but a “feel” that something is wrong and needs to be checked into. The more you edit and the more experience you accumulate the more you will get a sense the something is off and one needs to trust their gut and the most advanced tool ever their own eyes. As a number 4K and 45K tris is nothing as far as pushing visual assets through the GPU so you can cater to the problem down steam as part of the finalization and is not as much an issue as say per object collision that is CPU bound that if you removed collision that is not being used would represent a performance loss “bound” to the object.
In general draw calls are more expensive than mesh detail and complex materials, ie 45k versus megabyte textures = no brainier.
As for tweaking for performance.
In general it’s not the object that the root of all evil but rather the total per frame render that needs to be evaluated as to what is causing performance loss by extension as 4K and 45K tris by it’s self does not add up toe much.
