Is Doom3's renderer revolutionary?

[Scali]

I don't think Epic has chosen not to use shadow maps in this case for performance reasons only.

I never said performance was the only reason. In fact, I also mentioned it would improve the perceived image quality

There's a whole bunch of potential special-case problems with shadow maps, even in offline rendering where we can adjust every possible parameter and cheat as much as we want.

What does that have to do with anything discussed in this thread?

Thus I wouldn't say that they've dropped stencils, rather that they chose the appropriate solution for each case after sufficient consideration.

How exactly is that different from me saying that stencils aren't the most appropriate solution for everything, and therefore the Doom3 engine is only a current trend, and next generation engines will start applying other shadowing techniques?

 

[Scali]

D3 technology was revolutionary when it was presented in 2001 - fully dynamic lighting and shadows in real time. Get over it

That depends on two things, I'd say:

1) Whether or not you've actually seen fully dynamic light and shadows in realtime before.
2) Whether you consider this revolutionary, or just a result of hardware that is now powerful enough to apply techniques in realtime that have existed since the 70s.

For me, the first is a 'yes' and the second is a 'no'.

 

[Luminescent]

Again, I don't see any difference between a tech-demo and a complete engine. Both have to have a complete renderer if you want anything to show up on screen, obviously.
And I think you got it wrong... if a tech-demo doesn't allow interactivity, that says nothing about whether the technique can be used interactively or not. It's about the technique that is being demonstrated. If the technique is not universally applicable, then obviously it is going to be less useful for interactive games, but it could still be revolutionary in some way. That depends on what it is going to be used for. Perhaps it is going to be used for medical imaging, and it's a real breakthrough there.

So in short, things like interactivity, universally applicability, unification and revolutionary have very little to do with eachother.

My point is, a technique does not a renderer make. If everything else in techdemo's renderer is traditional, except for a few new methods, whether they be novel or not, the specific technique(s) is(are) revolutionary, not the renderer.

What differentiates a relatively fully interactive 3D world from one that isn't, is the number of scenarios permitted to test the validity or utility of any 1 solution, and not just any specific techiques, but the conjuntion of all the techniques that compose the renderer. If any one technique/solution breaks down, then what use does it have in the renderer.

If Sandy developed a new way of tracing the path of light to the viewpoint, took into account an infinite amount of rays, and got it to work for a single or a few frames, but nothing else, would you call his renderer revolutionary, or his singular technique?

 

[Scali]

My point is, a technique does not a renderer make. If everything else in techdemo's renderer is traditional, except for a few new methods, whether they be novel or not, the specific technique(s) is(are) revolutionary, not the renderer.

I think both sides can be defended. I would lean towards 'some revolutionary techniques == revolutionary renderer'. As in, the renderer is a means to demonstrate these revolutionary techniques. It is part of the revolution. It's no use having something revolutionary but never actually get it working.
Besides, the rendering process is a far too large and complicated process to just completely replace in one go. Even if most of it is traditional, a few changes could already make a revolutionary difference in performance or image quality or such.

What differentiates a relatively fully interactive 3D world from one that isn't, is the number of scenarios permitted to test the validity or utility of any 1 solution, and not just any specific techiques, but the conjuntion of all the techniques that compose the renderer. If any one technique/solution breaks down, then what use does it have in the renderer.

But this doesn't depend on whether the demo is interactive or not, but on whether the technique allows interactivity or not.
3DMark03 is a fine example of that, I suppose. It is not interactive, yet the renderer allows full interactivity.

 

[Luminescent]

I think both sides can be defended. I would lean towards 'some revolutionary techniques == revolutionary renderer'. As in, the renderer is a means to demonstrate these revolutionary techniques. It is part of the revolution. It's no use having something revolutionary but never actually get it working.

I can agree with that.

Besides, the rendering process is a far too large and complicated process to just completely replace in one go. Even if most of it is traditional, a few changes could already make a revolutionary difference in performance or image quality or such.

If these "changes" are robust enough to keep up with the renderer as a whole.

 

[Scali]

If these "changes" are robust enough to keep up with the renderer as a whole.

Not really. Over the years we've used plenty of rendering methods that weren't completely robust or accurate or such (baked lightmaps, shadowmaps, bumpmaps, occlusion term, spherical harmonics, etc, etc, etc). That doesn't mean that these methods are useless, or that they can't make a difference in interactive realtime applications.

 

[Reverend]

How much do we more-or-less know about certain things/implementations in Carmack's next engine? With the answers, how much of it is "revolutionary"?

Should we re-visit this when id shows a demo using the next engine from Carmack or when a game using this next engine from Carmack is released?

Are we going to see another poll like this one?

 

[Luminescent]

If these "changes" are robust enough to keep up with the renderer as a whole.

Not really. Over the years we've used plenty of rendering methods that weren't completely robust or accurate or such (baked lightmaps, shadowmaps, bumpmaps, occlusion term, spherical harmonics, etc, etc, etc). That doesn't mean that these methods are useless, or that they can't make a difference in interactive realtime applications.

Hmm, I guess the ability to dynamically link a series of techniques to essentially create a "robust" solution, via many specialized techniques, slipped my mind. You mentioned HL2 as an example of tackling a variety of individual issues via specialized solutions rather than a singular robust one. One advantage of having a robust solution, though, is that it allows one to shift focus on other aspects of a renderer without having to deal with a variety of specific cases.

 

[Scali]

One advantage of having a robust solution, though, is that it allows one to shift focus on other aspects of a renderer without having to deal with a variety of specific cases.

Yes, but as someone mentioned before: at what cost is such a solution?
This is ofcourse a very important question in a realtime environment.
The main issues are to make it look good and be fast.
If it can look better and/or be faster with a set of less robust or flexible methods, then use that.
Robustness and unification are luxuries that we cannot always afford yet. Not even Doom3, and not even UE3.0.
But we've come to a point that they no longer affect the realism in a big way.

 

[Reverend]

The main issues are to make it look good and be fast.

So how revolutionary is the z-fail (=Carmack's/Creative's reverse) method in the Doom3 renderer? How important is the z-fail in the "grand scheme" of the Doom3 renderer?

 

[dksuiko]

You may not have outright discredited him, but with comments like "stupid little outdated game," it's not too difficult to get that impression.

Why does everyone seem to take that literally?
And what's worse, why does everyone feel like nagging about that on and on? I'm sure if it was about Bill Gates and his 'stupid little outdated OS', that hardly anyone would be bothered by it.
Which is my point exactly.

Why is my point as well. If someone called Windows nothing more than a "stupid little outdated OS," it obviously wouldn't be a technical discussion would it? You're trying to bill this as a technical discussion, but the whole point of it is to do nothing more than troll the carmack zealots. (Which you admitted yourself: "I am not aiming at Carmack/Doom3 itself, but at the people who unrightfully place them above other great developers/games.")

Carmack's Doom 3 engine is not the "Ultimate" engine, fine - it's guilty of making compromises like other engines. We get the point. Actually, I don't think anybody ever thought the contrary to begin with. You just continually insist that we see him as some sort of god, projecting some sort of image Carmack-zealots onto participants of the thread. Why? Who knows, because I for one haven't witnessed it until you made it a point to show that they were here. In fact, I still haven't seen any Carmack zealots here.

Far Cry was given equal praise on these boards. Along with EQ2's engine, Source, and UnrealEngine 3. And Doom 3 is no exception to that list. You say that the only difference is that "Doom3 is seen as the ultimate game engine, and Carmack as a god." To whom and where? This messageboard? Sorry, but I don't think so. Many of Carmacks own comments regarding various subjects have come into question in threads of this forum. And I believe there was a mutal sentiment going around here that Far Cry had rained on both the Doom 3 and HL2's graphics parade. Yet, you still wish to color us zealots.

In any case - no, it isn't fair for people to claim that Carmack did it all (who exactly here is doing that?). On that same note, is it fair for people like you who claim Carmack did nothing whatsoever?

I never claimed any such thing. Are you perhaps blinded by that same emotional hangup about Carmack that I am talking about?

See what I mean? So put down the pitchfork, the Salem Witch Trials ended long ago.

 

[Laa-Yosh]

Is someone here acting like an ass?
yes
Am I drunk?
yes
Should you check your email after your pal's wedding?
no

 

[vrecan]

It just took doom3 to long to come out, basicly everything that I saw in doom3 had already been done before and basicly all of it was done in the RIDDICK engine for the xbox.

It just seems that doom3 was to late to the party to be a revolutionary engine...

 

[AlNom]

It just took doom3 to long to come out, basicly everything that I saw in doom3 had already been done before and basicly all of it was done in the RIDDICK engine for the xbox.

It just seems that doom3 was to late to the party to be a revolutionary engine...

The flashlight in riddick is so much better than the one in D3 too...and it has a body view and a proper 1st person shadow.

 

[Scali]

Why is my point as well. If someone called Windows nothing more than a "stupid little outdated OS," it obviously wouldn't be a technical discussion would it?

I have given plenty of technical arguments, both before and after I made this exaggerated point. Get over it.
My point was, Bill Gates is not seen as someone holy, so anyone can flame him, his company, and their line of products (which are obviously very good products, else they wouldn't have lasted this long), and nobody cares. But even if you have well-founded criticism on Carmack's work, it's considered pure blasphemy.

Actually, I don't think anybody ever thought the contrary to begin with.

I do, look in other threads like this one: http://www.beyond3d.com/forum/viewtopic.php?t=16062
You will find people claiming things like that the entire lighting system is unified, or that everything is dynamic and realtime, and nothing is precalced, etc.

See what I mean? So put down the pitchfork, the Salem Witch Trials ended long ago.

The problem is more that I got attacked brutally from every angle, when I gave any kind of criticism on Doom3. Yes on this board. So who has the pitchforks?
And LOOK what I said in that quote.
There is someone who apparently thinks I said all kinds of things that I never actually said. Why is that? Well the only explanation I can give is that he is no longer capable of rational thought as soon as I give some criticism, and he just completely sees everything black and white.

 

[Circastes]

The poll makes it too easy for anyone anti-id to pay the renderer out. Pretty silly.

 

[mustrum]

Blocky cahracters combined with low res textures (partly) and bump maps. Throw some nice real time shados in there that eats 3ghz CPUs for breakfast and you got your revolutionary engine.
I am extremely disapointed in the D3 engine.
Revolutionary? No - Not at all.

 

[T2k]

The poll makes it too easy for anyone anti-id to pay the renderer out. Pretty silly.

LOL - talk about being silly...

 

[dksuiko]

Actually, I don't think anybody ever thought the contrary to begin with.

I do, look in other threads like this one: http://www.beyond3d.com/forum/viewtopic.php?t=16062
You will find people claiming things like that the entire lighting system is unified, or that everything is dynamic and realtime, and nothing is precalced, etc.

Yes, and thats when you promptly began calling people stupid, shouting "wrong wrong wrong!", "it's not the ultimate engine!" and simply began venting. Yet, you wonder why you get an emotional response rather than a technical discussion? The real question should be, why are you getting all emotional? Which goes back to what I said earlier, "the emotional responses are not in defense of Carmack, but rather the attitude you come at this issue with."

Why the emotional response? 'cause you started it!

 

[Scali]

Read the thread more carefully, and conclude that the emotion is annoyance, as a result of getting the same kinds of answers time and time again, and having to repeat oneself numerous times.
It's not like my first post was like that. It's not even like most of my posts are like that. If you can't see that, you are obviously blinded by something.

And I have to reiterate this time and time again aswell.
Just give it up. I have presented solid technical arguments and facts, and if you want to nag about some strong wordings I used here and there, it's obviously because there is nothing better you can find to argue.