Page 1 of 3

Author:  varsovie [ Sun Aug 14, 2011 11:24 pm ]
Post subject:  CQB on IDTECH 4

id Software: Doom 3 source code to be released this year:

Maybe time to switch to ID Tech 4 in CQB , instead of fighting with old Q3.

Author:  Dany0 [ Mon Aug 15, 2011 12:27 pm ]
Post subject: 

If a switch then to UDK. Or wait for RAGE.

iD Tech 4, or d3radiant to be more specific is one of the most user-unfriendly editors ever.

The development would be horribly slow and painful.

Switching to UDK wouldn't be a problem at all. I can even help with the level design, I'm a whole month into it. Also UDK allows you to make it fully stand-alone and no need for licensing.

Anyway, if you want a proof that switch of engine to UDK is a good idea, contact me, give me a week, the textures and top-down layout and I can re-make the cqb level in UDK. With some fancy stuff. But it'd still be UDKGame, and that's just slimmer UT3. But I can't promise you the level of polishing that TC:E or CQB offer.

As well as there won't be Linux support unless someone pays the license for UE3 source code. Linux support as well as OpenGL is there it just needs well, enabling. And manually going through everything checking it'll be compatible.

There is another engine that is available and that is Shiva3d. It hasn't got these game mechanics built-in but it's very strict in what you can do in it. I have 5 months experience with it but I left in rage because of the horrible documentation. In other words I've found myself trial&erroring some basic stuff over and over again, for days and weeks. Also importing models is pain in the ass. Almost impossible. But it has a very big range of supported platforms. But I don't recommend it. The stuff I did in 5 months I've done in a week in UDK.

I was thinking that another relatively rational switch would be to CryEngine 1. But then I realized how shitty the multiplayer support was. CryEngine 2 and 3 are way out of our league, and most of the stuff can be done in UDK, just by using brain.

BTW iDTech 5 is looking very very good. But I think if the switch happens today, the day RAGE is released True Combat Unreal: Elite would be waaay ahead.

Author:  affine [ Mon Aug 15, 2011 3:54 pm ]
Post subject: 

I have worked extensively with UDK and can promise you that is a really, really, really terrible idea. It is internally a complete train wreck, truthfully appealing only to artists and ignorant programmers. Radiant is more or less intuitive, but it was designed for practicality. The guys working on COD games use it. Practicality != easy, but it almost goes hand-in-hand with power.

You two have little insight as to what an engine switch implies. What makes Id Tech 4 or UDK better than ET? If this engine got realtime global illumination, like Crysis or the Frostbite engine, would you then realize that they're all the same? Just because the engine could use some improvements doesn't mean it is obsolete. Period. I highly doubt the player base would skyrocket overnight if coroner announced he was switching to a closed-source engine.

There are numerous papers on realtime GI out there. You should think about what you are suggesting - would it encourage a programmer more to suggest an engine switch or to implement new features one at a time, until you have a sick engine?

Author:  Alpha Red [ Tue Aug 16, 2011 1:50 am ]
Post subject: 

Dany0 wrote:
As well as there won't be Linux support unless someone pays the license for UE3 source code. Linux support as well as OpenGL is there it just needs well, enabling. And manually going through everything checking it'll be compatible.

That is exactly why UDK is not an option.

Author:  affine [ Tue Aug 16, 2011 4:13 am ]
Post subject: 

UDK just isn't a practical option. Plus, there is already Tactical Assault out there screwing around. I tried getting in on the project a while back. To be brief, I was quite a bit disappointed with their social skills (I didn't feel "wanted" at all, even though they have wanted ads on their site.)

UnrealScript is just so bad... It violates just about every rule of optimization in C/C++ that I can think of. Execution time is so stupidly slow compared to "native" code, and that's a sign that they are doing something seriously wrong.

Alpha Red, that signature is so true! Here, let's predict the rare probability of seals mating in one week by using a Poisson distribution...

Author:  sxy [ Tue Aug 16, 2011 9:27 am ]
Post subject: 

Yeah, RAGE should probably be a good cornerstone for future TC as long as it gains big enough player base and Coroner is interested in it.

Author:  Damien [ Tue Aug 16, 2011 4:52 pm ]
Post subject: 

Funny how no one is mentioning ETXreal. As soon as the mod support gets fixed, both TC games will be able to gain from instantly it and the disadvantages of using open source software are the same as they will be with iDtech4 and 5, eventually.

Author:  Dany0 [ Wed Aug 17, 2011 9:20 am ]
Post subject: 

I agree that UDK is not very practical. But it's very fast to develop in. I didn't say that ET sucks. That's what OP implied.
But radiant isn't practical either. GTKRadiant yes, but the original radiant is horrible. Also I am aware the radiant that infinity ward uses has been heavily modified for internal use lately.
Yes I am aware of that. I think Coroner is having a good time delaying the next release because of the source code release ^^
But I wonder why just OGL3.1 if you can go as high as OGL4.1 where all the new fancy stuff is 0.o' I mean tessellation, the texture optimization and so on.
iDTech 5 has a wonderful map editor. I'm looking forward to what will True Combat: Elit5 look like :D Hey that's a nice name, True Combat: Elit5. From gamers to gamers, polished with passion, developed with caution. Hehe.

By the way, when Coroner (hopefully) finally stops doing everything on his own, I'd like to propose CQB needs a new team.

Author:  Haraldzz [ Wed Aug 17, 2011 9:22 am ]
Post subject: 

Technically, I would say Xreal. ID Tech 4 has pretty horrible modding tools and stuff which would just slow down the progress (which sadly right now seems to be almost stopped) even if we would get our cool DooM 3 like graphics.

I wouldn't recommend using UDK too - somehow all games made with it look like "Oh, this game uses UDK engine, I can tell it by the looks." feeling. Also, never saw an actual overhaul on everything that is in UDK and looks atleast a bit of realistic. So far what I've seen is:

Xreal is pretty much what we need - it's the same ID Tech 3 just with better graphics and other pretty small nice tweaks. Importing/exporting maps and models would be no problem.

Author:  Dany0 [ Wed Aug 17, 2011 11:21 am ]
Post subject: 

I didn't suggest it. Personally I think there are reasons to rather make TC:E completely hardcoded rather then just switching engines. Switching engines is reasonable however I agree ET with source code available is powerful enough. Now TC:E or CQB doesn't need a graphics overhaul, what it needs is to become more competitive. Look at UrbanTerror, years without an update and it has much, much more players then TC:E and CQB together. The upcoming HD version is just so little better then it was before. It's still mostly BSP based, unlike any modern previously BSP based games.

I think CQB rather needs to start using static meshes even more often. I support shaders as well, but only a tight list of shaders that are welcome in 75FPS combat:
  • DOF, mainly while aiming
  • Very subtle bloom
  • Volumetric lighting, subtle too
  • Sophisticated particles, real-life looking, so not subtle but small and sharp.
  • Tessalation, yup there is a tessalation shader for opengl. Where I want it to see is to enhance normal maps to become real geometry on close-up, as well as getting damaged by bullets(temporarily)
  • Shutter based, weight-biased supersample subtle motionblur with limited range. This way it won't be disorienting so it won't change the combat for people with it off
  • (Ray-traced) pre-computed reflections. This would enable night combat(which would be AWESOME) but shouldn't be able to be turned off, just lower resolution.
  • Multi-resolution screen space ambient occlusion with wider FOV. That's the best one so far. Not only it's the fastest one (around 1ms on X360 I think, that's more then twice as fast) it completely removes the noise problem as well as it is actually more realistic and removes the problem when clipping through polygons and z-fighting(sometimes). And it's source code is available for free!
  • Volumetric particles. This includes: volumetric clouds(they look nice, and the same technique can be used on smoke), volumetric blood(enables low-cost sub-surface scattering)(though I highly doubt this would be implented) and volumetric explosions.
  • Sharpening, all these shaders will *i-need-a-broader-vocab* up a lot of pixels on the screen, you need to adjust this.
  • Screen space eye assimilation simulation. Basically simulating adjusting retina based on the environment, or in other words automatic color shader. There were some papers on this but I can't remember what's it's called like :(((
  • I won't even mention normal maps because that's so obvious, I will however mention detail maps as that's IMO the biggest pit between ET and newest engines in terms of, well, detail.

Also you need to remove all of the limitations in the engine. Like max polycount, max texture size, max sounds at once etc. The characters can live with some extra detail.

Also I suggest some advertisement. Because the word-to-word sharing didn't really work well did it. Though that was mainly fault of the pro-longed updates and "beta" status. Remove the "alpha" status, say it's "experimental". And go with revision numbering. So not CQB 0.1.1 0.1.2 alpha and so on but CQB r12 CQB r20 etc. It's much more appealing to newcomers and will lead to a greater community. And I think more support and more good, talented people willing to help is what you need at the top priority now.

What can I say... I don't want to sound like a smartass, I'm sure in the past a lot of people came by telling you what to do or not.

But I think I've listed just some rather obvious suggestions, and some worth a discussion. Though based on how often anyone from the team and Coroner visit the forum lately, I doubt anyone actually read this. On the topic, iD Tech 4 source code has the super fast stencil shadows and generally fancy lighting implementation. Why not some copypasta.

Author:  affine [ Wed Aug 17, 2011 3:38 pm ]
Post subject: 

Dany0, you're doing the same thing here you did with UrbanTerror many years ago. I used to play Urban Terror from 3.7 onwards, and I'm quite disappointed with the way development with it has gone. Hopefully you are not suggesting a similar route.

Here is what I suggest: stop. Why suggest an engine switch when you have no idea what kind of work is involved? It has been this many years and you still haven't drawn such a conclusion.
Programming a better renderer isn't difficult and wouldn't take too long. For this reason, we should all stop saying that all TC:E needs is a better renderer.

Author:  Baal [ Wed Aug 17, 2011 7:39 pm ]
Post subject: 

Why bother suggesting this? This project is very self serving and obviously won't go anywhere. The engine isn't the question. Look at TC:E. If you don't fix the problems, then the people will leave. No one wants to play some guy's wet dream project, and that's exactly what we have here. TC:E and CQB should have been two developments, to at least fix bugs and ensure the Mod's survivability. Besides, if there is no anti-cheat, where will your hardcore player support come from?

Author:  Alpha Red [ Wed Aug 17, 2011 8:12 pm ]
Post subject: 

Dany0 wrote:
Blah blah blah Gfx upgrade plz!!

Do you have any idea what this entails? This is insane, just utter insanity.

Author:  affine [ Wed Aug 17, 2011 10:47 pm ]
Post subject: 

Guys, understand that Dany0 is a at essence trolling here. If I recall correctly, he was banned from the Urban Terror forums a couple years ago for starting a flame war over an engine upgrade.

Is there anyone here that would mind telling coroner to check his private messages?

Author:  anon912 [ Wed Aug 17, 2011 11:46 pm ]
Post subject: 

It doesn't look or sound like he's trolling. His ideas don't reflect the current development status of Urban Terror. If anything, Coroner should consider his ideas, as it could bring in more people.
Besides, there's no guarantee it's the same Dany0 from the Urban Terror forums.

Page 1 of 3 All times are UTC
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group