Not logged in, Join Here! or Log In Below:  
News Articles Search    

Submitted by Anton Knyazev, posted on April 30, 2001

Image Description, by Anton Knyazev

Not much to explain about it, actually. Just real time attenuated light+radiosity+soft shadows combo in real time. Oh, did I mention it was real time :) ? Yes, I know that the scene is pretty simple (not to say basic), but at least soft shadow part is independent on complexity.

Anton Knyazev

Image of the Day Gallery


Message Center / Reader Comments: ( To Participate in the Discussion, Join the Community )
Archive Notice: This thread is old and no longer active. It is here for reference purposes. This thread was created on an older version of the flipcode forums, before the site closed in 2005. Please keep that in mind as you view this thread, as many of the topics and opinions may be outdated.
Anton Knyazev

May 01, 2001, 08:11 AM

Hmm. Does that ATI technique produce correct penumbrae (that take into account _changing_ light size and distances from light to object and from object to shadow plane)? If it does than I suppose it's T-buffer like multisampling. My technigue is not multisampling, since in order to get a really good quality one will require hell lot of jittered samples (each one requiring object re-rendering, btw), and with only 8 bits per color component even this will not help. Image resizing and converting to jpeg does not give a good impression of quality, but one can see it in a full version.


May 01, 2001, 09:02 AM

Looks nice, but I have a few doubts about it.

Out of interest, what frame rates do you get rendering the scene without the shadows? I'm interested in what percentage of the processing time this "realtime" technique takes. It strikes me as being quite a low framerate, considering the machine spec.

I think it's already been mentioned that 20-30 fps will hardly translate to realtime once the technique is used in an actual game. Once you have a more complex scene, animations, AI, physics, special FX, sound etc each eating up portions of your processing time, will there be any room left for this technique?

Also you mention that the method is independent of complexity. Does this mean it's independent of the complexity of the model casting the shadow, or of the scene that the shadow is being cast on to, or both? Also how do multiple light sources affect the processing time?


Anton Knyazev

May 01, 2001, 09:08 AM

Yes, I know that this technique is not practical for current hardware. And it depends on complexity just like projected texture approach should be, while possibly allowing using more low resolution model for casting shadows than in the hard shadow case.


May 01, 2001, 09:11 AM

But now you're contradicting your original claim that the method is independent of complexity??

Anton Knyazev

May 01, 2001, 09:25 AM

One can use a pretty simple version of original model for shadow casting, of constant complexity, for instance. Of course, it cannot be _absolutely_ independent on complexity, since it would mean it would be independent on model at all, and since it it's not dependent on model, shadows of 2 different models will took exactly the same, blah blah blah, well, you got the idea.


May 01, 2001, 09:34 AM

Fair enough.

Have I got it right that you're NOT using a projective texturing approach?

Anton Knyazev

May 01, 2001, 09:43 AM

I said I did use projective textures, but, obviously, not in a typical way :)


May 01, 2001, 09:50 AM

Actually, I think it's more of a warm, moist towel war.

Alex Herz

May 01, 2001, 09:56 AM

just 2 things...

first of all does look pretty cool and that everybody is urging you to open the details how you did it should be seen as a compliment from the site of the flipcoders..if it was just another landscape or what ever, nobody would give a damn how ya did it.
but if it is *just* an implementation of something explained in some paper you should rather say that than to pretend you made it up yourself (by saying nothing)..this would be just stolen..

secondly I think that generally spoken developers should work together where ever possible..if you hide away your idea from the rest of the world nobody can improve it or built up on it(apart from you) and as you're just one errorfull human(as we all are) you're likely to miss a lot of things that could be done with it in the future..
if pythagoras had kept his maths stuff for himself and all the former *developers* after him we'd still use our fingers to count appples we bought on the market rather than making cool 3d shit..

I know that this is not compatible to many firms out there sadly..
the first things you have to sign even before your contract is an NDA saying you musn't say anything to anybody sadly..
Here ppl should learn from carmack's way of working..

uh..and carmack should do something new by the way..I'm bored of all the bsp pvs lightmapped stuff floating around all over the net..
pvs is brute force..and no way elegant or anything...



May 01, 2001, 10:05 AM

In case I havn't made it clear, I find this IOTD to be very impressive. Indeed, I don't know what you guys are talking about when you say you don't see the radiosity. What were you expecting to see? Anyway, I also note that portions of the shadow that are closer to the caster are sharper and portions farther away are more blurred. Once again, very realistic. And it's not likley that this is being done with straight ol' blurring of a texture since that would yield a uniform blur across the whole of the shadow.

And I, for one, find 25fps to be very impressive regardless of the fact that it is not operable in an actual game. It is still a great step forward and illustrates that not long from now it WILL be possible in realtime in an actual application.

The only thing I see missing - and correct me if I'm wrong here - is self-shadowing.

Good job. And don't hesitate to publish a paper about it at your earliest convenience! ;-)

Anton Knyazev

May 01, 2001, 10:14 AM

I do believe that self-shadowing does not need visible penumbra effects, neither it needs to be smooth, so it can be done separately using, for example, stencil approach (since in this case its major bottleneck - overdraw - can be eliminated)

Anton Knyazev

May 01, 2001, 10:23 AM

No, I believe there's no paper on this technique.

And I think it would be boring if I just told the details. It's a site about games; IOTD is a game, too. It's funnier to guess the way then just read it, especially taking into account that it's not of practical value on current hardware. Besides, my management is undecided about whether I should just describe the method publicly.


May 01, 2001, 10:26 PM

What I meant by "I cannot see the radiosity" was just that there's nothing to indicate that there's any radiosity processing - i.e. there's no visible diffuse interreflections coming off of the objects. Hence, there's no way to tell that it's not just fancy ambient lighting.

This thread contains 73 messages.
First Previous ( To view more messages, select a page: 0 1 2 ... out of 2) Next Last
Hosting by Solid Eight Studios, maker of PhotoTangler Collage Maker.