I stand corrected. Damn it. How can I argue with logic like that? Mr Spock would be proud.

Pete

quote: Original post by LaBasX2
@panic: At the moment, soft shadows are impossible in a realtime applications. Perhaps the GeForce3 offers support for them.

Not at all. All soft shadows require is an accumulation buffer. Render the scene multiple times each time with only one light turned on, then accumluate the results. You could optimize a bit by only redrawing the polygons that the current light hits.

quote: Original post by FordPrefect
All soft shadows require is an accumulation buffer.

And an accumulation buffer is NOT supported on any consumer videocard (it would take too much bandwidth..)

So you can''t do it in realtime..

3Dfx supports some kind of accumulation buffe through extension.
nVidia GeForce3 support MS antialiasing that could mean they support ABuffer.

Even if they don''t support it in hardware you can do it yourself with other hardware features, think instead of complaining.

-* So many things to do, so little time to spend. *-

i know how to do softshadows in realtime
dont bother asking how though. ive got more important things to waste my time on.

http://members.xoom.com/myBollux

Simon Price is right! Carmack was INVENTING.

Dude, Carmack is THE MAN! Yes, he uses a lot of the general 3D math from olde text books and stuff, but a lot of the specific algorithms he has to work out himself. Carmack is always at the front, with us suckers trailing behind. Of course it''s easy for us parasites.

So I say, let''s all get together and do something new, instead of following tutorial after tutorial, and doing Quake BPS loaders to see if we can do what Carmack did a few years ago. Let''s get to the real work!

------------------------
CRAZY_DUSIK* pCrazyDuSiK;
pCrazyDuSiK->EatMicroshaft(MS_MUNCH_BILL_GATES | MS_CHEW_BILL_GATES);

Ok, if you come up with something new we can do I''ll help you devleop it! So far every thing I''ve ever thought of as a new idea I found someone else had already done it.

Hi all


The thing about new ideas in 3D Graphics is generally new ideas are not new, well that''s what I also have found anyway. Basically all the new innovations in 3D Games have come from papers that have been sitting around for 10+ years. Which when you think about it is scary. The theory seems to be years ahead of the hardware that can do the graphics in real time. So I think that the best thing to do is to go read a few hundred or thousand papers and try and find something that is cool and has not been done before then adapt it and name it after yourself. That’s a recipe for instant fame well maybe.

Got To Go I Have Some Reading To Do,

3DG


Ps just think in <15 years time we will be playing games that look like Toy Story, Its an exciting time to be in 3D Graphics, exciting indeed.

I didn''t see the 80/20 rule mentioned yet. It states that you write 80% of the code in 20% of the time, and vice versa. So given you get an engine up in 1 month, it will take you at least four more months to fine tune it. And this is not counting the actual game code (logic, AI, networking, art) or administration time (scheduling, meetings, etc). People act like this is a piece of cake after the engine, but it takes a respectable amount of time as well, if not more so.

And as mentioned, we have the benefit of many tutorials whereas many of the "big boys" are cracking this stuff for the first time.

Ut