HalfLife 2 engine and Doom 3 engine comparison

I'm an ultra noob, but would like to pose a question. The fact that Doom3 has less of a polygon count doesn't mean the engine isnt capable of more polygons, just mean they chose to use less because current cards can't handle it right? So if games in the future using D3 engine impliment higher polygon models then the look of the game should be superior to that of the source engine right? Please don't make fun of my lack of knowledge for the love of all that is pure and holy.

 

I wish all 'noob's' would think as lucidly as you do since you are quite right. There is no inherent limitation of the polygon count in the Doom3 engine - it is limited by today's hardware capabilities. It is actually quite easy to separate a Doom3 engine limitation from a Doom3 game design/implementation decision.

 

Correct - at the moment processor power is the limiting factor. The more surfaces there are the more calculations the lighting routines have to perform.

Well, this is once again a fairly subjective issue, since how something looks is to an extent a personal preference. However, I personally think that higher-poly count and higher-res textures would give the D3 engine the potential to look better, yes.

 

But who knows what features are hidden in the Source engine. Just because HL2 uses a bunch of old tricks doesn't mean that other games are limited to light maps and static shadows. Besides, licensees tend to modify an engine to suit their specific needs.

 

Okay, I'll bite. The HL2 engine lighting model as presented in the PDF is arguably more complete than the D3 engine lighting model as presented in the D3 game.

The latter seems to ignore indirect lighting altogether and this makes some scenes in D3 look really bad, especially the one where there is one single light in a small room and the light is swinging to show off the pitch-black shadow volume attached to its upper side (I think a small touch of ambient light might have made it look less unnatural without playing havoc with the bump maps).

Valve's model does account for indirect lighting via precomputed radiosity maps as usual (at least for static lights on static geometry) except that their technique extends the lightmap idea to include directional information in order to make this work in conjunction with bump mapping. How well the interpolation for non-static objects works remains to be seen but I guess it will be less obtrusive than the fake shadows. All in all I think this allows for more atmospheric lighting in scenes, because the static portions are by and large lighted correctly (thanks to the radiosity precomputation).

Except for scenes where darkness and shadows are important, of course ...

 

Radiosity doesn't light things correctly. I'd rather have pre-computated noiseless path-tracing or pre-computed photon mapping. Radiosity calculates lighting perfectly diffuse. It's similar to ray-tracing which calculates shadows perfectly hard and lighting perfectly specular/direct.

 

Doom 3's renderer is superior to Source's renderer, its a fact.

Now as to which game looks better, its entirely different.

 

Roger that. I think Doom3 looked great and I greatly enjoyed playing. Half-life2 also looks awesome and I'm sure I'll have a blast playing it as well.

 

You mean you will actually enjoy BOTH games without resorting to petty shots at the game that does not favor your favorite IHV? I shudder at the thought!!

 

Doom 3 never has shadow mapping. The shadows look like they're floating in the air but if you look closely then you will find out that the shadows are indeed affected by the geometry of what the shadows cast on.

 

True, but look at the minumum requirements you need to play Q3, and then look at the min requirements for Call Of Duty (MoH didnt have many "true" outdoor environments). Imagine the same difference between D3's minimum requirements, and a game using the D3 engine with large outdoors.

 

Precisely, he was making the point that it was not an engine limitation. You have just supported that by indicating that the hardware requirements are the limiting factor.

 

Ahem, it's pure BS: Q3 had large maps, outdoor maps already, even not as large as later titles had.

Rest of the topic simply funny. D3 as the future way to go? ROFL...

Gentlemen... as somebody said somewhere on this board (I'll look for it), probably even JC would admit that the engine (NOTA BENE: not the approach, the unified lightning - this engine!) a no-go for the future.

For the polygon counts: WTF are we talking about? D3 has such a low polygon count with (nowadays) already really low fps number. Perfectly shows this engine, as it is now, is incapable for 'real', 'normal' polygon count for years without other trade-offs (check the crappy environment in D3). And by the time it'll be an obsolete technique anyway - we will call this a nice dead-end, believe me...

Anyway, this "D3-is-the-future-over-UE3/SE/X-Ray/FC/Source/anything-even-though-we-don't-know-shit-about-half-of-them" kiddie BS really made me really laugh...

 

Neither HL2's nor D3's rendering engines are paving the way towards the near future of engines.

D3 uses many hacks and few advanced pixel shaders. Stencil shadowing is most definatelly not the shadowing technique of even the very near future! They have a razor edge and that can't be fixed without an advanced post-effect or re-rendering the stencil several times per shadow and blending them. But by then you've gone way overboard with your fillrate limitations. At least the advantage is that yes it is a unified lighting system. Your terrain, characters, vehicles and cola can shadows will be rendered by the same one system all over, reducing the complexity of the engine.

HL2 uses many different hacks and techniques for it's shadowing. It looks a lot nicer than hard-edged shadows IMHO, but too complex and will be prone to clitches and inconsistancies. I doubt HL2 will support dynamic shadow casting for every poly in the game like D3. Hl2 seemingly uses far more advanced pixel shaders than D3 (remember, D3 was made to run on very low-end pixel-shader cards, and even lower), enabling advanced effects with more efficiency. Also, because it doesn't use stencil shadows (who's weakness is that it needs a low-poly model in order to run well enough) the characters and general shadow-casters can sport a much higher polygon count. Finally, call me crazy, but I much prefer a blurry soft shadow with some pixelation here and there, and not truly dynamic for large objects such as buildings, than a game composed 95% of hard shadows (although dynamic). Personally, I find it much nicer to look at, overall.

I quite like the direction U3 is going, engine-wise. I don't know all of the details, and I'm dissapointed with their willingness to use stencil-shadows (if that is in fact true), but I was really impressed with the quality of the materials, and the soft cubemap shadows that would get blurrier and eventually dissapear as the light source would get further away.

 

Low shader complexity is an attribute of Doom3 the game and not the engine...why is it that people are having so much trouble with this distinction?

Me too

 

Well low polygons and low res textures is a result of the engine .

Performance with stencil shadows was extremely bad even with the low polygon models and low res textures. Because of design choices .

In the future polygon counts can increase and texture quality can increase. However they will allways be low polygon compared to other engines of the same time .


As for the out doors . You don't want to play a doom3 engine game that takes place at say d day . The main feature that the engine is meant to make look great would be of very little use (shadows in broad day light on a beach with few things to cast shadows ? ) And performance would be horrible .


The engine is not the way of the future. It was thought up over 5 years ago . Source engine isn't either but it uses newer ideas than the doom3 engine.


In the end it doesn't really matter. I predict that the source engine will have a few more games when all i said and done. But in another year or two we will be getting dx 10 and all the engines for that .

 

Perhaps because its complexity related to how the engine performs on today's hw?

 

Yah

This is kind of a weird discussion. Everyone has already made up their mind before they participate, and it doesn't seem like anyone is willing to budge. May as well be arguing which colour is better: green or blue.



Obviously, the answer is green.

 

I'm thinking HL2 might be better than D3 in gameplay but its engine isnt as advanced.