• Propper optimisation question
    10 replies, posted
I'm working on a hostel kind of thing with a massive part of it destroyed. What I've done is clipped the shit out of it, removing a lot of brushwork, then clipped the remaining walls into brick shapes, then randomly deleted "bricks" to make it look destroyed. It's a nice effect and I'm pretty happy with how it's turned out but I need some tips on optimising the remaining walls when I prop them. At current, I just copied and pasted them into a new file and compiled them with propper, using a heavily optimised collision model. However, they have hundreds of faces inside the model which are wasted and will probably create a shit ton of lag. Say this is my unoptimised, laggy wall: [IMG]http://i40.tinypic.com/11kvxut.png[/IMG] Should I do a similar method to this one: [IMG]http://i43.tinypic.com/5ey5iw.png[/IMG] Or this one? [IMG]http://i40.tinypic.com/25z2zyg.png[/IMG] Thanks in advance.
Why not make the stacks all one brush, and anything else a big brush.
second one
The inside walls are two bricks thick and the outer walls are 4 bricks thick. I probably should of said that in OP.
Yes, The second one would be better
[QUOTE=swampie;20872561]Yes, The second one would be better[/QUOTE] Judging by your thread history, and questions asked, if I was ZOMG, I would not trust your judgment.
Meh, That is how i would do it, It wouldnt make a game breaking diffence
Oh, wow. Before I actually tried them, I made all the faces that weren't showing nodraw. Before each one took about 20 - 80 seconds to compile. Now they're taking 1 - 4 seconds :buddy:
Lol, nodraw basically isn't there when you compile with propper.
I'm assuming that the spaces are for illustrative purposes only. If that's the case then why not consider these two methods? [img]http://img153.imageshack.us/img153/3833/brushcount.png[/img] But honestly, I don't think it should matter this much.
Compile however you like, decompile, delete uneeded faces in 3dsmax/xsi, recompile.
Sorry, you need to Log In to post a reply to this thread.