Slashdot Mirror
Doom 3 vs. Source: Comparing Engines
Tom V. writes "DevMaster.net has an article that outlines some of the technical differences between Half-life 2's Source and Doom 3 engines from various game development aspects such as graphics, A.I., physics, networking, etc. According to the author, the winner is the Source engine based on its 'completeness' as a game development package. However, in terms of graphics, the clear winner is Doom 3."
The reviewer's methodology seems to be:
1) Play both games.
2) Compare graphics/sound/AI etc between the 2 games.
3) Assume that those comparisons are solely founded on what the underlying engine provides, rather than some of what the games add on top.
It's disappointing that a site with 'dev' in it's name didn't actually review the functionality, performance, and extensibility of the engines outside of games that might use it. Doom 3 is not just map and sound data fed into Id's engine, and Half-Life 2 is not just map and sound data fed into Source.
The title should be "Doom 3 vs. Half-Life 2: Comparing games". This has the potential to be a fascinating topic, but this article is sophomoric.
Last night, I built the incredible mattress-car - basically, just a mattress with a (powered) car wheel at each corner. It writhed and wriggled in a gloriously disgusting manner, and somewhat disturbingly started following me around. I tried shooting it but that didn't help, so I tied a fridge to it, set it on fire and chucked it into a lake...
Doom 3 might have a basic physics engine, but I'm really looking forward to what modders can do with Source's network-friendly version of Havok.
Doom 3's sound engine is awful compared with the original Half-Life, let alone Source. I've got a below-minimum-specs PC with a cheap sound card from 1998, and in Half-Life 2 I get real-time, room-specific reverberation and sound occlusion. I once walked off while a character was talking, and his voice became muffled when I went round a corner. It sounded real. Plus, the gun and bullet sounds are physically modelled - notice how they vary with distance and surroundings? The only things I haven't noticed it simulate are the speed of sound and proper Doppler effects (which Halo does!), but still, Doom 3's sound playback just seems bland and flat in comparison.
Doom 3's graphics might be the first of a new generation of engines, but Source, while primitive in some areas, is an old-school engine taken to the logical extreme. Which is why I like it so much...
Tedious Bloggy Stuff - hooray?
I RTFA, and this guy is clueless.
Quote: "The Source Engine's main lighting system is real-time radiosity lighting."
There are no games using "real-time radiosity," period. Radiosity (or more generally, global illumination), almost by definiton, is too slow for real-time.
This should probably read "pre-computed radiance transfer." It's pre-baked radiosity, cooked as a variant on spherical harmonic lightmap encoding. In other words: no real-time lighting, just PRT, faked dynamic lights (which EVERY other game does) and projective shadows. There is also no real HDR (high dynamic range) rendering in Source, just the same clever faking everyone else has.
This stuff is old hat. Relatively speaking, Source is not technically advanced at all. The only place it consistently (purportedly) wins is the content tools.
The big point that is NOT mentioned in the article is performance. Anyone who has played a lot of HL2 and CS:S can tell you that Source is just sloppy, on any hardware configuration. It is prone to periodic chugging, studders, fps drops from particle effects and physics lag online, etc.
D3, comparatively is just tight. The unified surface shading model (lighting and stencil shadows) rocks, and iD knows exactly what they're doing. Valve apparently can not compete in the brainpower department.
Game-wise, I personally preferred Doom 3 to Half-Life 2, old school playability vs. hype-tour-04, but that has nothing to do with the technical content.
Valve can only hope to win by being the preferred mod platform. Their SDK uses tested and proven, centuries old, Elaborate Puppet Theater(TM) technology, so naturally everyone adores them for maintaining the traditional status-quo. Hooray for Valve.
Those who want to see how well Doom 3 can pull off Half-Life 2 style environments should check out the Doom 3 Can Do It Too project. I don't think they have a site but their forum is on http://doom3world.org. Anyway, if you just want to see results here's their latest video: http://www.pcgamemods.com/9875/ (still a work in progress, obviously)
Seriously, if he's talking about engines, why is he making mention of the character models?
The monsters however seem much more lifelike with their detailed skin combined with many details. The Source Engine has better models, especially human, but combined with the lighting and shadowing, the Doom 3 Engine creates amazing textures.
Though it is important to note how the engine renders the models, but the models themselves are not part of the engine! It's possible to export the human models from HL Source and stick them into Doom3. It's as if he's comparing the artists, and not the engineer of the engine.
He also says something to the effect of Source not having cut-scenes. Last I remember from playing Doom 3, the cut-scenes looked like they were being rendered by the engine, and not pre-recorded.
It looks like he's confused engines for games. If I were going to compare the engines, I'd create my own levels and models, render them with both engines, and base my results off of that.
You need to restart your computer. Hold down the Power button for several seconds or press the Restart button.
John Carmack admitted that the Doom 3 was so dark because the game couldn't handle light on the engine level. There is, apparently, some pixel bleed-through and seaming of textures when light-entities are placed more generously. He said, on his blog I think, that he essentially made a game to suit the engine, but that the engine would develop with each new game release
Unfortunately, I do not have a link on-hand.
Valve explained this once. The lake is the source, the filled glass is the game, and the faucet is the valve.
Thank you for reading my article and taking the time to comment on it, even if those comments were sometimes a bit harsh.
.plan entries, which I submitted to DevMaster.net. I'll admit looking back on it I made a few errors like saying the original Half-Life used Quake 2 when it mainly used Quake 1 code with some Quake 2 code. Someone mentioned this is a common myth, which I somehow adopted at the time. I was aware that the original Quake Engine was used at the beginning of development, but thought Valve switched to Quake 2 after deciding to redo most of the game. I wasn't able to listen to the Half-Life 2 sound, because I didn't play it at a computer with speakers. Since writing the article (I wrote it in mid-November) I've listened to the sound and still agree that are equally good in the sound category. I tried not comparing the actual sound clips, but instead the way the sound is able to resonant.
As the author of this article I was happy to see my first article Slashdotted. Some of you came off as pretty harsh, but I can see where some of your criticism is valid. I'm not a crazy fanboy of either company, I'm not getting paid by Valve to write this, and picking at my spelling mistake of using "then" instead of "than" are ridiculous. I simply submitted the text article and the nice editors at Devmaster added the pictures, title, and summary. One of us should've probably caught the mistake, but it's easy to overlook. Whether I picked the Doom 3 Engine or the Source Engine I would've been flamed. The debate is similar to the Republicans and Democratic parties where you just can't win. I don't think anyone can rightfully say with an unbiased view that the Doom 3 Engine is a more complete engine then the Source Engine. Once you get past the graphics the rest of the engine just can't compete with what's out there today.
The article was originally one of my
I tried my best to compare the engines as best as possible without comparing the actual game's content. I used the SDKs to try and do this, but it still came down to in-game content for stuff like character models. Someone said that Doom 3 can achieve Half-Life 2 quality models, but that is completely untrue. The engines use different methods for creating character models, which gives each engine its own distinguishable type of model look.
I only compared the Source Engine and Doom 3 Engine, because those were the two people had been taking about. The discussion has been up for debate on forums all over the net, which is clearly seen by reading these comments. The Unreal 2004 Engine is a great, flexible engine, but it wouldn't have been far to compare it with the others. I did write a FarCry article, which should be published at DevMaster.net soon. If you want to read it now it's up on my site.
I appreciate the constructive feedback, because it lets me know how to revise my writing style for my next article. Like I said before this was my first article, and I made a few mistakes. I'll make sure to not repeat those when writing my next article, which will compare Unreal Engine 3 with Oblivion's graphics engine.