Hard Light Productions Forums
Modding, Mission Design, and Coding => The Modding Workshop => Topic started by: Omniscaper on February 26, 2005, 06:34:02 pm
-
I never understood why people kept bugging me about the uber-ness of my modelling tendancies. Since I first started modding for this community, starting with the Trek conversions, I dismissed all expressed worries reguarding model polygon counts and uber-sized multi-textures.
Right now I'm at my parents place for the weekend and I'm using my old donated:
1 ghz Athlon
756mb ram
Geforce 2 GTS
I got spoilled ever since I finished building my uber PC workstation. I am now suffering from my own ignorant embelishments. The 60-120 fps I'm now use to has been reduced to a 5-50 fps experience. I guess the complaints I got really meant something to them. Because being in their shoes has really humbled my ambitions for all things uber.
I now have a new motivation to include extensive LOD's for all my models as well as texture sizes that start at 64x64 to a max of 1024x1024.
Once again, I apologize for my ignorace for the technological lower middle class. I don't apologize to folks trying to use my stuff on a 486 PC though. :)
-
HA!! FEEL THE PAIN OF LOW FPS!!!!! HAHAHAHA!
Well, glad to know that you know how it feels like when your game goes from a sequence of slow frames to a single screenshots for several seconds. That and the fact that you'll be working on not-so-ub3r maps and lods for that ub3r stuff :D
:nervous:
-
::lowers head::
-
your an elite you have the right not to care
-
You don't have a darn things to bow your head about (though humbleness is a good quality :) ) Omniscaper. The things you work on are awesome.
Only now that I bought more RAM am I able to play with all of the current SCP features ('cept for the shinemaps since my cat's are > 4.4).
Game companies spend months optimizing their work. I say make it a show stopper, so that people look at your work say "I need to get that". Then when they come onboard to HLP they get recruited into being work minions to churn out low LOD's!!
It's a perfect plan... :drevil:
Cheers!
-
:: ponders a reversion to ignorance::
hmmm....
::contemplates the difference between ignorant or progressive motives::
-
Originally posted by Omniscaper
:: ponders a reversion to ignorance::
hmmm....
::contemplates the difference between ignorant or progressive motives::
In all honesty, I don't think we can ever realistically say is if there is a difference because the results of either could be bad or good, regardless of the initial understanding or will.
I'm not going to get into it because that's one of the aspects of this midterm paper that I've been trying to work on all week... trying being the operative word. And I don't want to get too off topic, as I tend to do.
See I'm starting already... :rolleyes:
-
Hahahahahaha, suffer!
-
it's ok to make models beond the capabilitys of most modurn hardware, because by the time you get done, everyone will have caught up.
-
Well, at least this system has a lot of ram. (I hope it's DDR ram though. ;) )
The gfx card ist not that bad. If it has 64 MB ram, texture sizes (as long as you use compressed maps) should not affect the performance.
Ok, here are my tricks to improve the overall performance:
-Use only compressed maps. (S3TC for the main map, DXT5 for the shinemap.)
-If your model has more than one map, try to reduce the number of total maps for lower lods.
-Try to avoid the ani format. It's bad for systems with a slow CPU.
Rather use EFFs (S3TC) for effects and glowmaps.
-Try not to use too many lods. Systems with with a small ammount of ram will rather run out of memory than gain performance.
-Use PCX for effects that only need around 256 colors.
-Some effects can look ok with S3TC compression too.
-Use compressed maps for skyboxes. If the compression looks horrible, double the size of the map. It will still be smaller.
(Double size = size*4 / S3TC = size/8)
-Use S3TC for big glowmaps rather than PCX.
-
In the end though, the entry-level PC spec will catch up with your models no matter how many polys you shove in them. It's just a matter of time.
-
Most of these things take months or years anyhow so computers will catch up faster than people can make the models. Its just a matter of pace, one person verses an entire multi billon $ industry