Wtf is up with linear falloff?!?!?!

pthomas1172

Come on guys?!? Stop using linear fall off with spot an point lights. Have u ever seen a linear falloff in real life.

Sorry...
Sorry.

It just pisses me off to no end to see a production game with all the lights with linear falloff

equil

if realism is your issue then light falloff should be the least of your worries.

i have no idea what response you're looking for but linear falloff might be used because it's cheaper or easier to control, and some engines don't even support quadratic or cubic falloff.

Swizzle

Who cares if it isn't physically accurate?

If it looks right, it is right.

malcolm

All lighting in games and pixar movies and 3d in general is faked. Volume lights + linear falloff for the win.

Noors

Games should use linear workflow before considering lighting decay.

r_fletch_r

-yeah, ignore this-

Mark Dygert

It really depends on the environment and how strong and how far light travels... I really don't know what you're carrying on about? You can get some amazing results that can be a lot quicker and cheaper (techwise) with linear, if you know what you're doing...

in short: Yea what Swiz said.

rebb

From what i understand, some ( usually older ) games can't really afford using proper quadratic falloff - because to get a similar look the light bounds would have to be much larger, causing a lot more fragments to be shaded and a bigger performance hit. ( at least for realtime lighting )

pthomas1172

But i feel as games get better its something that needs more attention

Will Faucher

perna wrote: »

the hell?

.

Swizzle

pthomas1172 wrote: »

But i feel as games get better its something that needs more attention

Linear and quadratic and other kinds of falloff won't be important things to worry about until ray tracing and ray tracing-based radiosity calculations start to be done in realtime in games. That's when games will actually start to become photorealistic, and that's when we'll need to start to worry about representing things like lights more realistically.

Until that time comes, it doesn't matter if we use linear, inverse-square or your sister's underpants to determine light falloff. I repeat: If it looks right, it is right.

Besides, there are more important things to worry about. Fluid, cloth and hair simulations, efficient skin shading techniques, more advanced tessellation technology, megatextures and procedurally generated content, AI, smoothing out kinks in animation and motion capture, lip sync technology, destructible models, more efficient implementations of camera effects like bokeh and per-object motion blur...

Those are just a few of the things we need to worry about before we'll ever start to have to worry about whether a light's falloff is physically accurate. And even then, falloff is only something that a hardcore realistic sim might make devs worry about; it makes no difference to the people who play the game.

commander_keen

If you want realistic falloff then you are going to need every surface in a level touched by every light. For static lights thats fine as long as you are also baking probes, but nonlinear falloff is hard to control. Quake3 used it for baked lights and it was super annoying because you can only specify a color and brightness, no radius.

r_fletch_r

commander_keen wrote: »

If you want realistic falloff then you are going to need every surface in a level touched by every light. For static lights thats fine as long as you are also baking probes

Similar to what lightmass does right? Creating a grid and doing directional samples of the indirect illumination.
Also I had noticed that UDK has a default falloff exponent of 2 for its static lights. Things seem to be in order, Am i misunderstanding something here?

gsokol

perna on 03-26-2011 10:24 PM
the hell?