Unreal 3.0 and Dynamic Radiosity Lighting

[Scali]

So the renderer is the easiest/shortest/smallest part of the total workload in a game development?

That's not what I said, and that highly depends on the game, and the team working on it. But yes, in general you'll probably find more people working on content than on code, for games. And oftentimes you also find that content is what takes the longest, not code.

 

[Scali]

Scali if you are so important why do you have a morbid fear of saying who you are? Your arguments rely heavily upon your supposed expertise and opinions, so it is only logical that people might wonder about your experience. Why not show what you actually coded years ago, or tell us who you supposedly are so we can all see how right you are instead of just blathering?

Is this a discussion forum or a job interview? Geez.
What does it matter who I am, if I am making technical statements anyway? If the technical statements are valid, and I can give in-depth explanations of them, then the expertise in that area should be obvious anyway.

 

[Scali]

Can't you read? I wrote that Doom's preprocessing is not about lightning, it's about culling.

How exactly are lightmaps not about lighting?

Lightmaps are not the only kind of preprocessing that is done by real-time engines like doom 3.

You said the processing was NOT about lighting.
I never said that the lightmaps were the ONLY preprocessing. I just said that there were preprocessed lightmaps, so at least part of the preprocessing is about lighting, see?

 

[Scali]

It's all a matter semantics and how you define revolutionary. IMHO Doom 3 is a revolutionary game engine; it may not be the opinion of a well-versed engine programmer, but of an avid enthusiast who can understand and interpret the direction/development of real-time rendering in retrospect.

I think it's more about making a distinction between an engine, a game and a rendering subsystem.
If you break it down, it's something like this:
Doom3 the game: Immersive experience because of novel use of light and shadows. Revolutionary? Perhaps.

Doom3 the engine: Limited in its applications, but at the moment the most advanced one for Doom/Quake style FPS. Weak AI, physics not as nice as for example Max Payne 2. Not revolutionary.

Doom3 the rendering subsystem: Not significantly different from the ones used for Tenebrae or 3DMark03 for example, not such a big deal.

Doom 3 may use hacks/techniques that were around long before, but it was a bold move in that no other game before it used the phong shading model for every pixel in addition to a robust (robust meaning a singular implementation for all but 2 game instances) shadow algorithm for the dynamic shadowing of pixels based on light occlusion data. It is real-time scanline rendering's priliminary attempt (or parody, however you may see it) at dynamic global illumination. In this attempt, inderect lighting was unaccounted for, probably because it would'nt have been dynamic. Therefore Doom 3 represents, to some degree, how accurately rendering can be, at this point in time, when rendering everything dynamically and scalably is paramount.

As said before, it does stuff that was possible on GeForce256. Doom3 represents what was possible back then, we can do somewhat more today. If you look at HalfLife2, UE3.0 and 3DMark05, you see some of the things that today's hardware is already capable of, but that Doom3 doesn't use yet... Some examples would be softshadows, HDR lighting, offset mapping...

One can claim that level design was artistically taylored to the rendering engine, which is true to some extent, but the gist of having a unified lighting and shadowing algorithm was there. If Doom 3 would not have made the move, we would be continuing to see games which offered dynamic lightmaps on characters with precomputed static lighting on many models and game environments.

If Doom3 would not have made the move, others would have. And others did, before Doom3 was released. Most notably Tenebrae and 3DMark03. The main technology has been around since the 70s, we've just been waiting for the hardware to arrive that makes it happen (in fact, 3DMark99 already contained a scene which applied shadowvolumes and a projected spotlight).
That's my entire point about Doom3 not being revolutionary. It was the obvious next step with the programmable hardware of recent years.

The same thing with shadowmaps... They have also been around for ages, but weren't feasible on hardware until recently... Now if we would get the first completely robust algorithm for shadowmaps, that would be revolutionary. I'm not going to hold my breath though.

 

[Neeyik]

So the renderer is the easiest/shortest/smallest part of the total workload in a game development?

That's not what I said, and that highly depends on the game, and the team working on it. But yes, in general you'll probably find more people working on content than on code, for games. And oftentimes you also find that content is what takes the longest, not code.

I know that's not what you said - stop being so defensive. It was a question, that's all. I was aware that the biggest workload is game content but I just wondering, compared to other elements of the complete engine, just how much does a renderer account for. Clearly it depends on the game but surely there must be market space for a functional, scalable and stable renderer that can be licenced out?

 

[Scali]

I was aware that the biggest workload is game content but I just wondering, compared to other elements of the complete engine, just how much does a renderer account for. Clearly it depends on the game but surely there must be market space for a functional, scalable and stable renderer that can be licenced out?

I don't think that a renderer by itself will have a big chance of survival. In today's world, most engines are still purpose-built, and this includes the rendering system. As I said before, you can't run FarCry on the Doom3 engine.
And if you write a renderer for Doom3-style games or FarCry games, you would have to compete against complete packages... I don't think people want to do the physics or AI themselves if they don't want to do the rendering either.
Actually... if there was a market for renderers, I think FutureMark would be the first one in it.

 

[Luminescent]

If you look at HalfLife2, UE3.0 and 3DMark05, you see some of the things that today's hardware is already capable of, but that Doom3 doesn't use yet... Some examples would be softshadows, HDR lighting, offset mapping...

Yeah, but in the case of HalfLife2, the local lighting system is not on par with Doom 3's per-pixel phong approach. In HL2 dynamic objects get per-pixel material properties and shading, but many elements in the environments do not. Sometimes it seems like models don't belong in their environments. What is missing is uniform lighting.

I think UE 3.0 merges the approaches taken by Doom 3 and Half-Life 2 into one. It takes Doom 3s local lighting + stencil approach and adds it to something that is equal or better than Half-Life 2s indirect lighting solution along with its HDR and widespread use of shaders. On top of this, it adds jittered shadow maps.

It is cool to have widespread shaders in the sense that HL2 presents, but I see as the best of the old with newly tacked-on features (HDR, etc.) that challenge new hardware, but these changes do not affect the world uniformly, which is why there is a visual discrepency between it and Doom 3.

Finally, back to the topic at hand. What kind of results should we expect from pre-computed shadow masks for indirect lighting? Are there solutions (even simple ones) that would not require offline processing

 

[Sxotty]

You faulted people for not knowing who you are perhaps you should therefore tell them.

You don't even know who I am. Perhaps you've been playing my games for a long time now?
That's another stupid thing about human nature. They assume that people they don't know are not capable of anything, and can't possibly be smarter or more experienced than they are.

You made a point to say you had made a renderer years ago that did the same thing as Doom3 as a supporting argument to show it is simple and easy. If that is true then why not show some of your work... Furthermore if it is so simple and easy then you are effectively telling us it took no talent to make and it doesn't even show you have any experience and skill to begin with. You have argued yourself into a hole either it is difficult and therefore proves you have the technical knowledge to back up your statements which refutes your statement that it is simple, or it is easy which destroys any value that can show for your experience.

 

[Luminescent]

Scali probably exhaggerated, not thinking anyone would take his statements at face value.

 

[Cryect]

No it does actually use preprocessed lightmaps, as stated many times before. And these lightmaps introduce limitations on the lighting and animation system.

What are you talking about? You talking about the projected textures you can use for lights if you so choose to? Thats a purely artistic choice if you want a fancy design on your light or whatever. Since its using stencil shadows of course it can't go through a translucent surface to change lighting and this is fine for majority of situations to use a projected texture instead. I really wouldn't call an artist making a texture preprocessing.

The only damn preprocessing it does is compute the BSP and compute for stationary lights the shadow transformations from what I understand. Theres a little visbility set computation if portal textures have been used. At least as far as map preprocessing is concerned (can of course discuss preprocessing for compiling of shaders, creation of normal maps from heightmaps or whatever).

 

[Laa-Yosh]

I think that someone here needs to get a few weeks off...

 

[Sigma]

I think I know who Scali is! 8) It's Gabe Newell... Or Sweeney...
Just kidding...

 

[dksuiko]

Scali, if you won't give credit to Doom 3 for anything it does, just what recent engine do you credit as something worthy of praise and why? And whatever engine it is, I certainly hope it's not one which applies previous theories into a practical real-time application - as that is one base you fault Doom 3 on.

 

[T2k]

Is everyone intoxicated by Carmack's stupid little outdated game and his stupid WRONG buzzword talk? It disgusts me.

What exactly is your problem? Is this some kind of jealousy that Carmack is rich, succesful and you are... well, a no-one who constantly resorts to insulting people on bulletin boards? Get a grip.

Diplo: stop doing this, please. It's really ugly and it's not even true!
It's like a bad mantra when somebody darn to call Honourable St. Carmack virginity in question...

I've got the same when I was explained my bad mouth taste after playing this half-baked bread, called D3 and expressed my deepest doubts regarding its engine's capabilities - similarly to Scali.

Honestly, I'm VERY impressed by Scali: he's supertolerant, hyperpatient... so many utter silly question, again and again, nobody wants to admit his points.

PS: You can challenge him on anything but he's definitely NOT the one who constantly resorts to insulting people on bulletin boards?... you know:

What exactly is your problem? Is this some kind of jealousy that Carmack is rich, succesful and you are... well, a no-one

 

[T2k]

I think that someone here needs to get a few weeks off...

*shrug*

May I ask who and why?

 

[T2k]

I think I know who Scali is! 8) It's Gabe Newell... Or Sweeney...
Just kidding...

That would be cool, atucally.

 

[Bjorn]

Doom3 the rendering subsystem: Not significantly different from the ones used for Tenebrae or 3DMark03 for example, not such a big deal.

The Doom 3 engine was done 3+ years ago according to Carmack. But it takes time to make game content these days.

 

[Reverend]

In a PM to me, Scali has alluded that I may have actually played games that he has contributed to. He could be right. I mean, how many thinks I may have played a part in any of the games you guys may have played?

In any case, let's leave Scali (and some others) alone with his opinions on why the Doom3 rendering engine is jurassic-like in age, on why Carmack should not get too much credit for the Doom3 rendering engine (since Scali came up with a similar engine 2 years ago), on why the Doom3 rendering engine (which has appeared in a game we can all buy) couldn't have been as hyped as it is if all the other id staff's input to Carmack wasn't taken into account by Carmack himself, and on whether, really, if Scali could have released that similar-to-Doom3 rendering engine of his and have it implemented in games that were available 2 years ago without any of the prevalent DX8 or DX9 hardware currently existing in the market today.

Let's just focus back on topic (what wa sit again?... UE3, is it?) and avoid criticisms of forum participants. Otherwise, I'll lock this thread.

 

[T2k]

Doom3 the rendering subsystem: Not significantly different from the ones used for Tenebrae or 3DMark03 for example, not such a big deal.

The Doom 3 engine was done 3+ years ago according to Carmack. But it takes time to make game content these days.

Ahem... linky, please.

Jan 29, 2003
------------
NV30 vs R300, current developments, etc

At the moment, the NV30 is slightly faster on most scenes in Doom than the
R300, but I can still find some scenes where the R300 pulls a little bit
ahead. The issue is complicated because of the different ways the cards can
choose to run the game.

The R300 can run Doom in three different modes: ARB (minimum extensions, no
specular highlights, no vertex programs), R200 (full featured, almost always
single pass interaction rendering), ARB2 (floating point fragment shaders,
minor quality improvements, always single pass).

The NV30 can run DOOM in five different modes: ARB, NV10 (full featured, five rendering passes, no vertex programs), NV20 (full featured, two or three rendering passes), NV30 ( full featured, single pass), and ARB2.

Remember?

How many rendering path do we have now?
So what kinda 3 years we're talking about?

 

[T2k]

In a PM to me, Scali has alluded that I may have actually played games that he has contributed to. He could be right. I mean, how many thinks I may have played a part in any of the games you guys may have played?

In any case, let's leave Scali (and some others) alone with his opinions on why the Doom3 rendering engine is jurassic-like in age, on why Carmack should not get too much credit for the Doom3 rendering engine (since Scali came up with a similar engine 2 years ago), on why the Doom3 rendering engine (which has appeared in a game we can all buy) couldn't have been as hyped as it is if all the other id staff's input to Carmack wasn't taken into account by Carmack himself, and on whether, really, if Scali could have released that similar-to-Doom3 rendering engine of his and have it implemented in games that were available 2 years ago without any of the prevalent DX8 or DX9 hardware currently existing in the market today.

Dirt cheap, Rev. Dirt cheap.

Let's just focus back on topic (what wa sit again?... UE3, is it?) and avoid criticisms of forum participants. Otherwise, I'll lock this thread.

What a surprise.