Jump to content
Search In
  • More options...
Find results that contain...
Find results in...
hardcore_gamer

What do you think the community will do with the Doom3 source code after its release?

Recommended Posts

As some of you might already know, the source code for Doom 3 is being released later this year:

http://www.gameinformer.com/b/news/archive/2011/08/05/doom-3-source-code-will-be-released-this-year.aspx

What do you think the community will do with it? There have already been some mods that improve the graphics a lot even without the source code, what do you think people will create WITH the source code? Do you think we will see ports/mods that will dramatically change the game as was the case with the original Doom?

Discuss.

Share this post


Link to post
hardcore_gamer said:

There have already been some mods that improve the graphics a lot even without the source code

What and where are these graphics-enhancing mods you speak of? All I've seen is a few high res texture mods which manage to ruin the look and feel of a lot of the "cleaner" Mars City textures, and a frankly horrible HDR mod which makes the game look like this.*

As for the source code release, I can't see how we're going to see a sudden influx of revolutionary mods or ports or brand new projects. But I guess we'll see.


*EDIT: Perhaps a more architecturally interesting area of Doom 3:-

http://img515.imageshack.us/img515/9941/shot000030006.jpg

The effect of the mod still looks like ass though.

Share this post


Link to post
Gez said:

Yeah but this thread has a more direct question. The answer: dunno, probably the graphics-obsessed goons at Doom3world will grab the source code to do more of that shit, all while I prefer to see Doom 3 MAPS instead.

Share this post


Link to post

Again I'll reiterate my own opinion that the most a DOOM port will probably find useful in the idTech 4 codebase would be its scripting language, and even that would require significant work to decouple it from DOOM 3's model animation system - to express it in terms of an analogue with which everyone should be familiar, it's like DECORATE and ACS in one package.

The same thing would be if EDF and Aeon were one and the same in EE (although in fact Aeon *will* be embeddable inside EDF thanks to HereDoc Technology (TM) (R) (C))

Share this post


Link to post

Performance, Performance, Performance...

All of Doom 3's weaknesses with regard to the performance of it's renderer are now well known. I would be VERY surprised to find that major performance improvements are not added in short order. Changing to a deferred renderer alone should about double the performance in GPU limited scenarios.

It would be funny if John Carmack added every wish-list extension to the engine before release leaving coders and modders with nearly nothing to improve though. LOL

Share this post


Link to post

Skulltag 3... followed by people releasing their own games with a modified Doom 3 engine.

Share this post


Link to post

I'm relatively confident that the ambitious Doom 3 modding projects that are still going on will take advantage of that if they have good C++ coders in their teams to make their life easier by adding some features.

What remains to be seen is whether a few reference source ports will emerge.

Share this post


Link to post

It would be so great if somehow the source code granted access to the original high-poly models from which the actual in-game models as based. Well, now that I think of it, I can't come up with a single reason why that would be true. But dammit, it sure would be cool. I dunno, if I was John Carmack, I would have allowed for the full high quality meshes to be used in-game, since people would still be playing the game years after its release, on faster computers that can handle that much geometry.

Share this post


Link to post

if I was John Carmack, I would have allowed for the full high quality meshes to be used in-game, since people would still be playing the game years after its release, on faster computers that can handle that much geometry.

I totally agree! In fact they should've done that for the original Doom. No sprites or pixelated textures, just high-poly models everywhere! Nevermind that nobody would've been able to play the game at the time of its release, just think of how much fun we would have today if that were so.

Share this post


Link to post
GoatLord said:

It would be so great if somehow the source code granted access to the original high-poly models from which the actual in-game models as based. Well, now that I think of it, I can't come up with a single reason why that would be true. But dammit, it sure would be cool. I dunno, if I was John Carmack, I would have allowed for the full high quality meshes to be used in-game, since people would still be playing the game years after its release, on faster computers that can handle that much geometry.


What high poly models?

Screenshots/links?

Share this post


Link to post

What high poly models?

When models are made for video games, they first create super high-poly models and then use programs to reduce the number of polygons on these models to a number that can actually be used in game. Typically the original high-poly models have polygon counts that range in the millions. After reduction, the usable models usually have only a few thousand, if that.

By the way, in case nobody caught it: the things I said in my last post were totally laden with sarcasm. Putting the original high-poly models in Doom3 would've been a completely bonehead move, since nobody (and I mean, nobody) would've been able to play the game at the time of its release. Few people, if any, would be able to play the game today with such high-poly models. Basically it's a just a really, really bad idea.

Share this post


Link to post

I wouldn't expect any major engine or gameplay breakthroughs, and as others said, any extensions will be more bug and performance fixes, and perhaps some mods using a slightly modified engine.

TL; DR version: compare to the extent of (user-made, non-commercial) source modding that Quake 1-3 went through. Not even comparable to Doom's, is it?

Comparing the Doom 3 and Doom source releases is a bit like comparing a modern car laden with a ton of electronic gizmos (or even a small passenger helicopter) that require a super-specialized and specifically trained professional to even dare touch under the hood, vs a VW Beetle that any backyard mechanic can mod to incredible levels, exactly because it's so much simpler and the mods usually leave little of the original car.

It's not like you can e.g. change the renderer dramatically (unlike Doom's, which was brought from low resolutions, to higher ones, added previously inexistent effects etc. and that was even before OpenGL and 3D accelerated ports were introduced. This was also done with the first Quake, to a degree, but much less so with the second one, and practically nothing of relevance happened to Quake 3 (outside of commercially licensed "mods").

There's no reason to believe that Doom 3 will be any different.

Share this post


Link to post

I'm in total agreement with you, Maes. It's a bit refreshing to see someone who really knows what they're talking about here.

Share this post


Link to post
Maes said:

I wouldn't expect any major engine or gameplay breakthroughs, and as others said, any extensions will be more bug and performance fixes, and perhaps some mods using a slightly modified engine.

TL; DR version: compare to the extent of (user-made, non-commercial) source modding that Quake 1-3 went through. Not even comparable to Doom's, is it?

Comparing the Doom 3 and Doom source releases is a bit like comparing a modern car laden with a ton of electronic gizmos (or even a small passenger helicopter) that require a super-specialized and specifically trained professional to even dare touch under the hood, vs a VW Beetle that any backyard mechanic can mod to incredible levels, exactly because it's so much simpler and the mods usually leave little of the original car.

It's not like you can e.g. change the renderer dramatically (unlike Doom's, which was brought from low resolutions, to higher ones, added previously inexistent effects etc. and that was even before OpenGL and 3D accelerated ports were introduced. This was also done with the first Quake, to a degree, but much less so with the second one, and practically nothing of relevance happened to Quake 3 (outside of commercially licensed "mods").

There's no reason to believe that Doom 3 will be any different.


As far as game-play changes, those can already be done in the current SDK which already provided all the Source Code for AI, and Game-Play areas. There simply hasn't been much interest in Doom 3 as a platform because the Art Assets are all so specific to the game. Nobody but a few dedicated Mod teams wanted to make boat-loads of new textures and models for their mods and the community wasn't too interested in re-using the cramped Alien themed environs of Doom 3 they way they LOVED re-using HL2's environments and art assets.

Now, performance is another stumbling block for this engine but you must realize what you have here. A fully dynamically lit engine. We did not see another until UE3 and even there it has lots of gotchas and problems to the point where most UE3 content is light-mapped. With the Source Code release, there is the chance to make Doom 3 a real competitor to UE3.

It needs the following ingredients:

1) Deferred rendering
2) Object LOD and Occlusion tech (better Mesh management)
3) GPU calculated shadows

All those are well known tech and have public academic example implementations.

The real question is not whether Doom 3 could fly with some neat community added rendering features. The real question is whether OpenGL will actually render those features with good performance. John Carmack himself was gonna use Shadow Maps for the shadow tech instead of Stencil Shadows but OpenGL's poor driver support for buffers caused Shadow Maps to be so slow they were worse than Stencil Volumes. So the same might be true of other OpenGL features. The community could try adding Tessellation, Geometry Instancing, etc and find that these features perform horribly.

It is fortunate that games like RAGE are coming soon so that ATI and Nvidia have reasons to pay more OpenGL driver staff to fix stuff. But there are many questions... I suppose Chronicles of Riddick DA's engine has helped there too.

Quake 3's engine was majorly overhauled in the form of Xreal. Not many game developers have jumped in to build for it nor has the Quake 3 mapping community committed to a full source port. It seems that those interested in the Q3 engine took it and extended it without any care for the Q3 game. Nobody did the nasty (tedious) work of making sure that engine extensions were compatible with stock Quake 3 assets. The Quake 3 community is now focused on some extension tech called "Entity Plus" but they really need someone to re-write the engine or port Q3 to Xreal so that it doesn't live with old CPU bound limitations.

Doom 3 may see more care than that. If you think about it, Sikkmod already is a radical upgrade to the Doom 3 engine that works for the existing game. If the same changes were made without the poor performance of having to use hacks like z-buffer capture it would already be a bigger step up than Quake 3 was ever given.

Share this post


Link to post

First off there'll probably be an SDL port to just about every platform then I'm sure they'll be a ioDoom3-like project.

Share this post


Link to post

First off there'll probably be an SDL port to just about every platform then I'm sure they'll be a ioDoom3-like project.

Most likely yes. As the last few posters said, this source code release does not mean that we can expect to see any significant changes to the original game, either graphic-wise or especially gameplay-wise. In fact it's quite the opposite: in terms of the Doom community, both here and beyond, this source code release really only bears significance to the mod teams that could use it to make internal improvements to their own projects.

Share this post


Link to post
Redeemer812 said:

Only if a few years = 10-15 years. Seriously.


I think you may be underestimating the exponential growth of processing power, combined with recent breakthroughs in quantum computing.

Share this post


Link to post

I think you may be underestimating the exponential growth of processing power, combined with recent breakthroughs in quantum computing.

Video games are optimized nowadays so that an entire rendering cycle will process absolutely no more than a million polygons, if that. And even with those optimizations, you need a good computer to run those games favorably with all the fancy visual effects turned on. With that scale in mind, I don't really understand how you can say that in just a few years we'll be able to put ten times that level of polygonal detail into a single character.

Oh yeah, and about the quantum computers: quantum technology is cool stuff. Nevertheless, that technology is not used *anywhere* in modern consumer hardware. Nowhere at all. Just recently, one company designed and began manufacturing a quantum computer for business purposes. But the things are the size of refrigerators, just like our current technology used to be some decades ago. If history has anything to tell us, it will be decades before we can build quantum computers as small as the PCs we have now.

I've been a video game programmer for ten years. I study this stuff. Take my word for it... this is just not feasible, and it won't be for at least another decade. That's being generous.

Share this post


Link to post
Redeemer812 said:

Take my word for it... this is just not feasible, and it won't be for at least another decade. That's being generous.


I don't get it what would be the point either.

Modern gaming graphics are already close to being photo realistic anyways. What difference could some kind of a super computer possibly make anyways?

Share this post


Link to post

Modern gaming graphics are already close to being photo realistic anyways. What difference could some kind of a super computer possibly make anyways?

The benefits that newer hardware will afford us will not be the ability to throw millions more polygons onto something, but rather to shrug off cramping optimizations that really limit how a designer can make a game.

As an example, Voxel technology has always offered much more dynamism and flexibility than polygons, which have no volume. But Voxel technology is incredibly resource heavy and this is really as good as it can look on current technology if you don't want a framerate of 3fps.

As John Carmack has said multiple times before, the holy-grail of all virtual-reality enthusiasts (such as himself) has always been the Holodeck. While PCs really just offer the consumer a window into such a dynamic and responsive world, new technology might one day give us the opportunity to experience such a thing for ourselves. :)

Share this post


Link to post
hardcore_gamer said:

I don't get it what would be the point either.

Modern gaming graphics are already close to being photo realistic anyways.

Uhhhh, you're crazy. Not even Crysis or Crysis 2 DX11+High res texture are that close to photo-realistic, not to mention the majority of modern gaming graphics are tuned to console hardware instead of something with actual power. Show me an actual game that renders the graphical fidelity of a movie like James Cameron's Avatar (fidelity, not aesthetics because fuck that movie) in real-time, or the holodeck, and maybe I'll believe you.

Share this post


Link to post

http://youtu.be/tuCqLJtpuOs?t=20s

I wouldn't call this ENB series mod for GTA IV photorealistic in the same way that Avaturd is, but it's about the closest I've seen. Yes, you can still see angularity in the edges, and not 100% of the shadowing is represented. But watch this and try tell me that there aren't compelling moments where it looks damn near believable.

Share this post


Link to post
GoatLord said:

But watch this and try tell me that there aren't compelling moments where it looks damn near believable.


I'm sure people said the same thing about the original Half-Life when it was released.

Share this post


Link to post

I see what you're saying, but there's not a whole lot preventing games from looking photorealistic. We have the tools to do it, just not the horsepower. We're still more than a decade away from being able to trick someone into thinking a simulation is real, but you can't deny the enormous progress that has been made since the onset of mainstream 3D gaming in the mid 90s.

Whenever I look up screenshots from new games, the details are more and more refined, to the point that everything is shadowed with penumbras, depth of field looks cinematic, normal/displacement mapping looks much less plastic/ceramic and polygonal edges are less obvious. There have been a few times when, upon entering an electronics store, I was briefly fooled when seeing a sports game playing on demo mode on one of the TVs.

Share this post


Link to post

IMO it's not the tools or the hardware that we really lack, but the skill. Even if you dedicate all the processing power of a super computer towards rendering a hyper-realistic, motion-captured model of a human being, anyone could look at it and see that it's just a model. You might be momentarily fooled by it, but the fact is that there is always a giveaway.

Share this post


Link to post
hardcore_gamer said:

Do you think we will see ports/mods that will dramatically change the game as was the case with the original Doom?

No, I don't think so, but it would be great to see some new mods for Doom 3. Really, it would be awesome, in the past we only had a few great mods here.

Share this post


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×