Source vs UE3

[Crusader]

The only think that worries me about the upgradeability of Source, is the BSP system.... That has always seemed kind of restraining to me. Not currently, I think it's actually a good method for these Next Gen games, but for the next Next Gen games I wonder if the BSP system will struggle.

But then I am no expert on the matter, I just have experience coding and mapping for HL, and the convex brush system and entity system seemed restrictive.

 

[Dr. Freeman]

There is absolutely no comparision. UE3 is a next generation engine. The Source engine will be improved and upgraded constantly and they'll use it to power Half-Life 3. When that happens you'll be able to compare them.

quoted for emphasis.

seriously.. using common sense, don't u think its a bit foolish to be comparing an engine that will be availabel 2-3 years from now to an engine that is being used currently? certainly one would think that a level playing field by comparing engines made in the same time frame would be something that is natural.. :|

 

[Play4Fun]

That right there = Awesomenessticity.

I still think Gollum looks better

 

[guise]

http://www.sonsofvanu.com/dl/pafiledb.php?action=file&id=24

^ 170 meg hires movie of the E3 demo of the unreal 3 engine for those who havent seen it.

it may be a few years away... but the technology is around NOW to beable to do a tech demo which i find pretty amazing.

i suppose technically it cant really do much that source cant, but to me it still LOOKS a hell of a lot prettier.

and to the guy that was saying it might be a good idea for epic to start thinking about some new IP and different ideas for games, couldnt the same thing be said for valve?

 

[SnowBall]

Theres no doubt that UE3 > Source right now, since like many have said, its a next gen engine. Source is kinda like the father, old, but good, and UE3 is like the son, younger, better, stronger, faster... you get the point. However this might change as Valve begins to upgrade Source. Source just needs to work on the lighting system more as it seems thats its weakest area.

 

[guise]

source will NO DOUBT be upgraded. i think the only reason why the features weve been seeing so far may be seen as sparse by some people is becuase vavle was as many people as possible to enjoy hl2... and so to get it to run on an "average" system, sacrifices had to be made.

im sure valve will be able to equal if not beat the tech demo of the UE3.0 by the time its actually being used in games.

 

[Rupertvdb]

and to the guy that was saying it might be a good idea for epic to start thinking about some new IP and different ideas for games, couldnt the same thing be said for valve?

Yeah, you're right, making the second of two games is kind of rpeetitive, wait i have to go play Theif 3 or Doom 3 or any star wars game ever.....

 

[TIIMMYY]

if they are showing off the engine now, why not use it in a game now? the computer they did the demo from must have been able to run it ok, and thats made from todayse technology.

 

[The Dark Elf]

if they are showing off the engine now, why not use it in a game now? the computer they did the demo from must have been able to run it ok, and thats made from todayse technology.

That wasn't a full completed engine + game in the demonstration though, by the time its done it will barely even run on current equipment.

 

[EVIL]

Theres no doubt that UE3 > Source right now, since like many have said, its a next gen engine. Source is kinda like the father, old, but good, and UE3 is like the son, younger, better, stronger, faster... you get the point. However this might change as Valve begins to upgrade Source. Source just needs to work on the lighting system more as it seems thats its weakest area.

LOL.. excuse me.. I think source has the best lighting IMO.. fock all the dynamic lighting and shadowing shit doom3 has, it make everything look fake.

 

[EVIL]

if they are showing off the engine now, why not use it in a game now? the computer they did the demo from must have been able to run it ok, and thats made from todayse technology.

Have you seen the demo? once they showed shots from the outdoor level it got such a massive frame drop, and the card they used for it was nvidia's latest card, a overcloaked Geforce 6800 ultra I think it was. If they would release a game that looks like that now. You still have to wait atleast 6 months before the fastest retail card can run it at a steady 30 fps

 

[SnowBall]

LOL.. excuse me.. I think source has the best lighting IMO.. fock all the dynamic lighting and shadowing shit doom3 has, it make everything look fake.

Thats its problem though. Its not very dynamic in HL2. I think it also only uses Per polygon lighting instead of per pixel. Correct me if I'm wrong. Shadows only point in one direction in maps in HL2 as well. As for DOOM3, I dislike its lighting as well. Its overhyped. It uses ugly pure black shadows that dont blur. W00t! Go Open GL... :rolling:
But of all of Source's areas, lighting is definatly its weakest point, even though it still looks good.

 

[Someone]

I love the look of the UE3...but as everyone has said: it's years away. So there isn't any real engine right now to compare it to. But, there will surely be other engines which are on the same level very soon. At the time it finally is IN a game, we may very well be saying, "Wow that's pretty average by today's standards isn't it?" We just never know since it's so far away.

I guess a bigger thought (for me anyway) is just imagine what will be out in 4-5 years and beyond?

 

[NeLi]

Photorealism in 8 years.

 

[The Dark Elf]

Photorealism in 8 years.

photorealism has been possible for years. There's been absolutely nothing stopping programmers writing engines than could simulate everything and do everything and have millions of polygons per object... There's only one problem.. Hardware needs to catch up before an engine like that could be run.

So when you look at it that way, the UE3 engine isn't all that impressive atall, neither is Source, or x-ray or any of the others. Any good coder could go out right now and write an engine capable of more than all of them. Just wouldn't be able to run it properly for 10/20 years heh damn

 

[Big Fat Duck]

you people need to realize that developing engines is limited to the current directx9 and opengl2 technologies available, you can only add new features to a certain point

 

[RogueShadow]

Didn't a thread almost exactly like this get closed not long ago?
Maybe not...

Anywho, I've seen some great vid's of the UT3 engine.
It's very cool. I want it.

 

[RogueShadow]

photorealism has been possible for years. There's been absolutely nothing stopping programmers writing engines than could simulate everything and do everything and have millions of polygons per object... There's only one problem.. Hardware needs to catch up before an engine like that could be run.

So when you look at it that way, the UE3 engine isn't all that impressive atall, neither is Source, or x-ray or any of the others. Any good coder could go out right now and write an engine capable of more than all of them. Just wouldn't be able to run it properly for 10/20 years heh damn

Which is why people don't make games on raytracing engines.
At least not very many, and with all the details very low.

 

[Crusader]

Quoted in Agreement:

photorealism has been possible for years. There's been absolutely nothing stopping programmers writing engines than could simulate everything and do everything and have millions of polygons per object... There's only one problem.. Hardware needs to catch up before an engine like that could be run.

So when you look at it that way, the UE3 engine isn't all that impressive atall, neither is Source, or x-ray or any of the others. Any good coder could go out right now and write an engine capable of more than all of them. Just wouldn't be able to run it properly for 10/20 years heh damn

At last someone else gets it, I have been saying this in every thread about graphics for ages. The fact is Doom 3/HL2 are not pushing the boundaries of graphics in the way some people are imagining, they're just making use of technology that our computers can now handle.

Which is why all these arguments about which game has got the best graphics annoy me. If they really wanted to, any developer could just go out and say "screw you, you moaning bar stewards" and make an insanely good looking game, nigh on hollywood quality, and laugh at us all when we have to wait 20 years before we can run it at more than 0.3 frames per second. It's OUR hardware that is the limitation on game graphics, not the artistic skills of game developers.

It's like people think that when Doom came out, the people at iD could only see and draw in 8 bit

 

[guise]

i thought its pretty obvious the quality of graphics that can be produces... we see them all the time in movies.

the challenge is getting the best possible graphics to run on the current technology. i think developers are doing pretty well . obviously as years go by theyre gna get better and better though :dork:

 

[Big Fat Duck]

HomeLAN - How do you think your next generation engine compares to other game engines such as id's Doom 3 and Valve's Half Life 2 (Source) engine?

Tim Sweeney - UnrealEngine3 is essentially one generation ahead of the hybrid DirectX7-9 engines shipping in 2004. We've designed UnrealEngine3 from the bottom up with 100% focus on DirectX9 and later hardware and the kind of graphics and CPU specs we're expecting from next-generation 2006 and later PC's and next-generation consoles. By throwing out compatibility with past hardware, we've been able to create a much purer and more general renderer that's not constrained by old fixed-function and 8-bit-per-pixel rendering architectures, and this fact really shows up when you look at the engine's visual fidelity, support for very detailed outdoor as well as indoors environments, and very open and unrestricted shading and lighting system.

Of course, we're targeting games shipping in the 2006 timeframe, so being a full generation ahead ought not to come as a surprise. But Epic stands somewhat alone in terms of shipping major engine releases and games coinciding with each major generation of GPU's and console game generations.

so basically, unreal3 is ditching support completely for older non dx9 cards and they are maxing out on what dx9 has made available

 

[urseus]

^^^^good.

Cheapasses with crap computers need a kick up the ass every few years....

 

[Devilphish]

UE3 is a couple generations beyond Source in reguards to rendering technology. Source is the last major engine using the old lightmap technology. The next step is a unified lighting solutuion like that in Doom3. Such a unified system is a revolution in real-time rendering for games. What UE3 does is the basic evolutionary steps taken after the major revolution if a unified system is introduced. It will take the revolutionary solution and begin to make incremental improvements on the new technology available, such as making the hard shadows soft and so on. Doom3 is capable of looking just like UE3(for example: fuzzy shadows has been added to the engine already), but the features cannot be put into the game and still run on the target system specs. Later games using Doom3 tech will make these evolutionary steps. Expect to see full fuzzy shadows and all the fireworks of UE3 in Quake4.

As for Source's role in the next generation, I'm not sure. They say they will update it, but so far all I have heard is higher res textures and new shaders added. Source cannot compete in 2006 with these superficial updates. The lightmap technology Source uses will be obsolete by late 2005/early 2006. In the early generation of unified lighting I can see some reasons why lightmap technology would be favoured by some developers, but by the later generations of the Doom3 engine, and when UE3 engine comes out, there really won't be a need for lightmaps anymore. In order to compete in this timeframe, the rendering engine of Source will have to be replaced by something more up-to-date. That could happen, but so far I have heard no mention of it by Valve.

The other major aspects of Source, such as physics and shaders, will only have to be modestly updated to compete with UE3 and later iterations of Doom3 tech in 2006. Source is well off in that reguard.

 

[Styloid]

Lightmaps are the best looking lighting systems right now while still allowing resources free for other features. When we have a real-time lighting/shadowing solution that looks as good and isn't a cpu killer should be when Valve adds it in. Valve has always been more about quality of the content than how 'advanced' its tech is.

 

[poo?]

You guys seem to be overlooking the fact that Epic is showing off the engine now, which says to me they are trying to shop it around to other developers to be used soon not 2 years from now.

They also never said the engine would be done in 2006, they said to expect the first games using the engine in 2006 which would mean the engine would need to be finished to license out sometime soon to give developers time to make their games.

I don't know about you guys but from what I saw the engine looked pretty much done to me, although they may not have generic AI routines or the netcode in yet (they could for all we know) as well as all the other bells and wistles, but from a licensing standpoint isn't the rendering and physic systems basically what the developers are looking for in an engine? From what I saw those looked complete to me.

So given all that whats to keep them from licensing the engine out to developers right now? And if you agree with me and am wondering what my point is in all this, it's that Valve is trying to get into the engine licensing market with Source and I think when developers consider competing engines and have a choice between the Unreal 3 tech and Source then Valve's engine won't even be considered because of Unreals superior tech.

Whew that was long, you could probably make a drinking game out of the many times I said the word "engine".

 

[Devilphish]

Good point poo, and thats one of the reasons I really don't think Source will be liscensed very much. Maybe four or five games. Source happened to come out when we are undergoing a paradigm shift in the way games are rendered, and it's on the backend. If you are developing a game with a projected release for 2006, I don't see any incentive to go with Source over UE3 or Doom3.

 

[Sparta]

If you want realism and scalable graphics for older hardware over unifed-lighting and a graphics engine that demands the latest and greatest hardware, i'd go with Source. Plus, i personally think HL2 looks better then Doom 3.

Also it really depends on the developers needs. Take Troika for example. Bloodlines wouldn't be going along the same lines it is now if it were on the Doom 3 engine. It would be limited to narrow spaces and small enviroments like Deus Ex 2 was, because the graphics are so demanding

 

[Devilphish]

If you want realism

I don't see anywhere where Source looks more realistic than UE3. Doom3 doesn't look as "real world" realistic, only because certain things couldn't be added to the engine due to target system requirements, Fuzzy shadows and higher resolution spec maps for example.

scalable graphics for older hardware over unifed-lighting and a graphics engine that demands the latest and greatest hardware

We are talking about games that will hit shelves in 2006/2007. For developers beginning to develope now, older hardware is not a concern. The recommended specs for hl2 should be the min specs for these games, at the least. The scaleability of Source is a plus for games coming out right now and in the very near future, but not so much for games which are starting development right now.

Also it really depends on the developers needs. Take Troika for example. Bloodlines wouldn't be going along the same lines it is now if it were on the Doom 3 engine. It would be limited to narrow spaces and small enviroments like Deus Ex 2 was, because the graphics are so demanding

There are a couple things here. For one Doom3 is not limited to anything. The cramped feel of most of the environments is designed to heighten the fear of your surroundings. There will be large environments in parts of the game. More importantly, this is a non-issue because as I said earlier we are talking about games coming out a couple years from now. Those games mentioned may be demanding now, and Source may be more suited for a game like bloodlines, but we are talking about games coming out in a couple years where these engines won't be demanding, they will be the norm. Bloodlines could easily be made with Doom3 engine or UE3 engine and run well in 2006/2007 hardware.

 

[Styloid]

So Source is missing real-time shadows and thus it will be dated within a year?
Valve has already said that they are looking into a more efficient real-time shadow system for the future. Source is made to be upgraded. Deus Ex 2 beat everyone to the real-time shadow market on the Ureal2 engine. Heck, there is a mod for Quake 1 that has real-time shadows and lighting added in. It's not a big deal and it is certainly not a huge leap in technology. As said before, developers only have to worry about the hardware available in the market and how much other stuff they want to pack into their game.

 

[Devilphish]

So Source is missing real-time shadows and thus it will be dated within a year?

It's boarderline dated now. Granted it still looks acceptable, but thats only because the alternative solution is only just now hitting the shelves. By 2006 I'm sure there won't be more than maybe 1 or 2 significant titles hitting shelves that doesn't feature a unified lighting solution, and no significant titles in production without a unified lighting system. Wether Source will feature a unified lighting solution in that timeframe is up for grabs.

Valve has already said that they are looking into a more efficient real-time shadow system for the future.

If you have a link I would like to get Valve's take on their plans for Source in more detail. Anyway, Source already has a dynamic shadow solution. Keep in mind when I'm talking about unified lighting system I don't just mean dynamic shadows. Source's current dynamic shadow system is very mundane, although it's the best solution Source has. It may produce a fuzzy shadow, but the shadow system it'self is limited and flawed. Objects won't cast shadows on themselves, nor will characters. Objects won't cast shadows on other dynamic objects, infact they cast shadows through other dynamics which is an eyesore(see the 2004 videos with the crane). Objects don't cast shadow for more than one light source. Shadows often remain the same shade even when the casting object passes through different lighting conditions(infact it's not clear wether the shade changes at all". Only dynamic objects will cast a dynamic shadow, which is a huge inconsistancy. I have seen people here say it is more effecient that way and therefore better. If by more effecient you mean calculates less shadows, you right. If you mean better as in looks better, your very wrong. Wether the inconsistancy is an acceptable tradeoff depends on what you are trying to do, and as I said for Source it is the best solution - but not the best solution overall.

So your proclimation that Valve is looking into more effeciant dynamic shadow solutions means that they could just be looking for a slightly more consistant way to perform their current solution, rather than adopting a unified lighting system. We have seen that Valve is somewhat reluctant to point out the major pitfalls of their solution in Source when compared to a unified lighting system, for example when they say "we do dynamic shadows like the Doom3 engine does, we just do it different.". This is the usual word game you get when developers dance around the truth - the truth being here that Source is worlds below the shadowing capabilities of the Doom3 engine. They ofcourse don't want to say that outright.

Deus Ex 2 beat everyone to the real-time shadow market on the Ureal2 engine. Heck, there is a mod for Quake 1 that has real-time shadows and lighting added in. It's not a big deal and it is certainly not a huge leap in technology.

It is a big deal and certainly is a huge leap in the way games are rendered. A unified lighting solution changes everything. Ever since Carmack set down and wrote the first lightmap based FPS engine, games have relied on hacks and tricks to make the lighting conditions of the world look realistic. These hacks and tricks have progressively become more and more convincing, but they are still hacks - illusions. It looks more and more real, but doesn't act any more real. In a FPS, it is very important that it acts real because a game isn't like looking at a picture. In a game the world is dynamic and changes, the lighting - which is by far the most important aspect of how we view the world - must be made to change along with it. At the present this is the biggest step we could possibly make to bring games to a higher level of realism. This paradigm shift is and will be viewed in the future as a revolution in gaming technology, and by the very definition of the word revolutionary, it fits.

As said before, developers only have to worry about the hardware available in the market and how much other stuff they want to pack into their game.

This only backs up my earlier point that virtually no one is going to be interested in liscensing Source at this time because they could do so much more with other more advanced engines and 2006/2007 hardware. That may change when and if Valve puts in a unified lighting system, but all we can do on that is speculate. It may not even happen until they go in and overhaul their engine in preperation for hl3.

It is a certainty though that lightmap technology is on it's last legs and by 2006 it will be as good as dead. We will have to wait and see what valve does before we can fortell how long Source will last. But even now, as I said earlier, I see no incentive for anyone to liscense it, even right now today, until valve does impliment a unified lihgting solution. Source is a good option for games coming out now, but games 2 or 3 years away? No way, not in Source's current state.

 

[Styloid]

Devilphish, I'm not sure but I'm PRETTY sure that we both know that this topic has been brought up over and over and over again. If you want to tell me that Deus Ex 2 graphics are revolutionary and that real-time shadows is an ability that only a select few developers possess that is your business. To me it seems more like devs have been waiting for hardware to catch up to use it so they can stick it in their 'features' section on the box.
Call me naive, but if a bunch of nobodies can stick real-time shadows into Quake 1 code and release it on the internet, I don't see why Valve can't patch it into Source over Steam when they feel that it looks/runs well enough.

 

[Devilphish]

Styloid, I don't know where your coming up with this stuff, or even if you have read my messages. You address me, but you don't seem to be replying to any of my points or even saying anything that relates to my points.

If you want to tell me that Deus Ex 2 graphics are revolutionary and that real-time shadows is an ability that only a select few developers possess that is your business.

I'm not trying to tell you anything of the sort. Infact, I have never even mentioned DE2. IMO, the game is rather ugly. But introducing a unified lighting system into games IS revolutionary. That is a fact. By definition, it is revolutionary. Developers can take a revolutionary technology and make something ugly with it, but that doesn't change the fact that introducing the technology is a revolution in games.

Nor did I suggest in any way that only a select few developers possess the ability to create a unified lighting system at any time. Where you got this from I don't know.

Call me naive, but if a bunch of nobodies can stick real-time shadows into Quake 1 code and release it on the internet, I don't see why Valve can't patch it into Source over Steam when they feel that it looks/runs well enough.

If you mean as an update for hl2, I don't see how this would even be possible. They would have to rework all map files and many other apsects of the game to get it looking right. You can't tack on a unified lighting system onto a game that was designed for lightmap technology and expect it to look right. Even the guys who did the quake mod had to rework everything, but reworking hl2 will be a much larger task. It would be almost like making the game all over again, something I don't think the developers will put time into. I'm sure they have other projects they will want to focus on rather than spend a year tweaking their finished projects.

As a Source update(not hl2 update), I never said or even suggested this isn't possible or that Valve wouldn't do it. We just don't know(unless you knwo something I don't)

 

[FISKER_Q]

There is absolutely no comparision. UE3 is a next generation engine. The Source engine will be improved and upgraded constantly and they'll use it to power Half-Life 3. When that happens you'll be able to compare them.

The Unreal Engine also gets improved and upgraded constantly. The Unreal Engine has been upgraded for several years now into what is now UT2004. And in 3 years what is Unreal Engine 3.

But yeah it shouldn't surprise anyone if they would look somewhat the same when it gets to a public release, but remember Epic has the technologies implented now. They just can't release it yet.

Valve are still a bit behind and need to work it up on their level.

 

[Crusader]

It's boarderline dated now

I love this stuff. Where the hell are you guys living? Where are all these insanely good games that I've never seen? I want them so bad, I am here sitting on my ass with a new graphics card waiting for HL2 to come out so I can run it at a decent 30 fps, but apparently I should be playing all these games that look way better than HL2 that I have never seen.

I want to come to the magical land of the fairies too

Edit: Btw, if you can get light to react realistically, and still use lightmaps in areas which are never going to be subject to dynamic lights *which you can* you'd be a bit stupid to waste a ton of processing power dynamically lighting stuff that never moves. This is the position valve take, too. Also real time lighting is nowhere near the quality of pre-calculated lighting at the moment, and I don't see it being able to calculate radiosity or anything for quite a few years, basically it'd be like recompiling the lights in your map for every frame. A frame an hour? No thanks.

 

[fido]

....But introducing a unified lighting system into games IS revolutionary. That is a fact. By definition, it is revolutionary. Developers can take a revolutionary technology and make something ugly with it, but that doesn't change the fact that introducing the technology is a revolution in games.

If you mean as an update for hl2, I don't see how this would even be possible. They would have to rework all map files and many other apsects of the game to get it looking right. You can't tack on a unified lighting system onto a game that was designed for lightmap technology and expect it to look right. Even the guys who did the quake mod had to rework everything, but reworking hl2 will be a much larger task. It would be almost like making the game all over again, something I don't think the developers will put time into. I'm sure they have other projects they will want to focus on rather than spend a year tweaking their finished projects.

As a Source update(not hl2 update), I never said or even suggested this isn't possible or that Valve wouldn't do it. We just don't know(unless you knwo something I don't)

Your confusing Half-Life2 with Valve's Source Engine that it runs on. Restrictions imposed on Half-Life2 so that it will run on low end machines, are not reflective of the Source Engine capabilities.

I remember Valve saying that Half-Life3 would have DirectX9 as a minimum. They have also said that Source is a very modular design and components can quite easily be added or removed.

I think Epic revealed the new Unreal Tech this early, partly because they feel threatened by so many other game companies that are now easily able to rival or beat their newest tech. Unreal3 looks good but will be no better than the competition.

If the current Source engine is cheaper to license than the Unreal3 engine, and more flexible, then that will weigh heavily on who developers choose to lease from. Games like Call of Duty are choosing 3 year old engines like the Quake3 Engine over the current Unreal Engine, and that was and is one the best selling games of 2003.

 

[EVIL]

Good point poo, and thats one of the reasons I really don't think Source will be liscensed very much. Maybe four or five games. Source happened to come out when we are undergoing a paradigm shift in the way games are rendered, and it's on the backend. If you are developing a game with a projected release for 2006, I don't see any incentive to go with Source over UE3 or Doom3.

I wouldn't be to sure about that.. source saves money for developers.. its a hybrid engine.. they can update the rendering engine with new modules and features instead of buying unreal engine 3 for 300.000 and having to buy unreal engine 4 for 300.000 if you want better graphics for the sequel.

Source will be used for Half Life³ because they will update it over the years, and the developers who licence the engine will have the same tools available to do the same or can get updates from valve.

 

[^Ben]

I think to much emphasis is put into the renderer. It's only about 15% of the final code base!

Valve has got UE3 beat in every other area(that we have seen) at the moment maybe apart from tools. Physics, AI, and other assorted good stuff.

 

[Varsity]

Whats the argument here? That Source's lighting is outdated and it won't be updated?

 

[EVIL]

"There is no argument"

 

[Kyorisu]

As I'm more of a gamer than a Techie I'll simply say this "HL2, U3 and Doom3 look good in their own regards and that'l be that"