Which one is the best (FPS) gaming engine? (discussion)

g00ey

Distinguished
Aug 15, 2009
470
0
18,790
I remember the good old days when I was a little kid playing around with the games Unreal and Unreal Tournament 1. Soon I started fiddling directly with the engine experimenting with more exotic things such as AI simulations or making movies. What fascinated me albeit it was really poorly documented, was the UnrealScript feature which basically is kind of like JavaScript but tailored for elements in the environment allowing for more advanced scripting of elements such as actors (monsters, NPCs, turrets, vehicles, etc), weapons, particle systems, light effects, advanced AI and so on. I looked upon this engine as the engine of the future but my friends didn't share my enthusiasm as much cooler games came along at the time. Now, many years later, the Unreal engine still lives and is thriving.

Then a few years later I came upon Doom 3. Something that struck me was that it was considerably more technically advanced than other contemporary games and much more stable. Firstly, it was visually stunning, you had a metallic shine that I hadn't seen in other games, you had heat effects beautifully visualizing distortions in the air from turbulent flows of hot air. Also, you didn't only have dead textures but you also had interactive panels directly in the textures in the game which I considered mind blowing. But the final aspect of this engine that really made it stand out was how stable it was. For example, switching between the game and other programs is still an issue today, the switching between the game and another program commonly takes around 10 seconds and it is common that when you switch back to the game, you either have a black screen where you cannot do anything but restart the game, or when you switch back, you end up in windowed mode unable to go back to the same old fullscreen mode without restarting the game. With the DOOM3 engine, switching between it and another program was a breeze and you could always rest assured that switching to or from the game will never lead to bad behaviour. Also something else that struck me was how well the engine worked with other applications. You see, on my laptop where I ran the game, when adjusting the volume or changing the contrast or any other hardware specific properties with one of the custom hardware keys of the laptop, a nice graphical HUD or OSD with a bar or other indicator of current volume pops up, beautifully overlayed on top of the desktop with a semitransparent alpha blend effect. Usually while running a game in fullscreen, these HUDs never show up. But in DOOM3 these features had no problems with blending in with the DOOM3 game engine and I was very surprised that it even was possible!

In more recent years I have partially lost touch with the computer games front but I read still with great interest about different technologies out there. I think ID software is or at least used to be one of the major pioneers in the field coming up with very interesting new technologies from time to time. I heard they used a revolutionary texture application method for the game of Rage that speeds things up considerably.

I really liked the half-life series but apart from the physics that were introduced with hl2, I was actually not that impressed with the engine. I like the TES games but to tell the truth, I find the engine those games are based upon to be buggy and bloated. For instance, a game like Skyrim when considering it runs on a PS3, should run very smoothly on a PC even with 2011 years of PC hardware. But the GUI is hopelessly laggy and sluggish.

So I was delighted to read that Bethesda were considering teaming up with ID Software for a new game engine to be used with future Fallout and TES series. But my delight was not shared which I experienced in the following now closed thread:

http://www.tomshardware.co.uk/forum/id-2753286/bethesda-fallout-presentation-gamescom-opinion.html#16442789

It doesn't look like Fallout 4 is based on any of the id Tech games but who knows what contributions have been made to their Creation engine.

I cannot say that the Wolfenstein or 'The Evil Within' games that supposedly are based off of the Rage engine are very impressive but I guess that is a matter of feature implementation. But the preview of the latest DOOM game looks very promising:

https://www.youtube.com/watch?v=QjzUFlk0eR8

The particle system implementation together with volumetric lighting and global illumination looks really nice so it still looks promising for ID Software.

Then I saw the following Youtube clips of potential future engines:

https://www.youtube.com/watch?v=5AvCxa9Y9NU

https://www.youtube.com/watch?v=vTF7wz0-AQs

Some people say that Euclideon is fake though.

I'm aware that there is much more to this than just the visuals such as the 'logistics' behind the development process and the availability of tools to handle such 'logistics', and other features such as storytelling or even AI which is still abysmal among games and hasn't improved significantly during the past 20 years!!!

So which engine has the most potential and where is the FPS technology going?
 
There is always lots of really impressive looking demos and things out there, but with the $2000 worth of GPU's you don't see it anyways. lol.

Eye candy doesn't sell games. CS:GO is one of the most popular, competitive games out there and it's old and nothing special. BF, COD, haven't changed much in years. In competitive gaming, that stuff actually gets in the way. I remember playing F.E.A.R. back in the day and was in some big tourney's online. Kept getting pecked off hiding in shadowed areas and couldn't figure out why. I shot out the lights, it was dark, crates casting shadows to hide in. On a whim in between rounds, I turned down my graphics and shut off dynamic lighting and shadow effects, could see everything.

I think the Doom engine never took off because Carmack put all his egg's in OpenGL, thinking it was going to take over, despite DirectX gaining and gaining. He never gave up on OpenGL but everyone else did for gaming. I think he become so arrogant that his engine would be so good that GPU companies would embrace OpenGL because of him and his engine, and, well, we see where that landed.

Some games were ahead of their time, like red faction. Destructible environments in red faction 14 years ago, in multi-player. You could tunnel half way across the map. The last game I saw about touting destructible environments is some new Xbox One game coming out, that has to use the power of the cloud to render the destructible environments since the XBox One can't do it on it's own. How did they do they 14 years ago on a single core PC with a few MB of RAM. lol.

I played in some big UT tourneys back in the day too. Few months ago, tried the Alpha UT newest super version. Really didn't much different than 20 years ago either. Didn't look much better either.

 

g00ey

Distinguished
Aug 15, 2009
470
0
18,790
Now, I think we should make a distinction between a game and a game engine. I'd consider the ability to produce more "eye candy" is a natural step in the evolution of game engines.

And sure, much of what we see more and more of in a real-time engine has already been implemented in ray-tracing engines. The evolution we then see here is the fact that features that used to be too "heavy" for real-time implementations in prior hardware can be implemented in current and/or future graphics hardware as it gets more and more powerful. This could also be a matter of newer, smarter and less computationally intensive algorithms for said feature, (perhaps at the expense of accuracy) that enables the use of it in real-time applications and not just hardware improvements per se.

I think engines should have features that are too heavy for current hardware as a way to stimulate demand for stronger hardware. But implementing such features involves development costs that a company might rather want to spend elsweyr.

Another not visible feature described in the prior post was the "logistics". With "logistics" I kind of meant content management. When developing a large game, large teams of developers, graphic designers, game level designers and so on must be coordinated efficiently so that the content from one team doesn't break the content from another team. I think in this regard a good content management system make the users focus on "content creation" without getting bogged down by menial administrative tasks.

I don't think going with OpenGL is such a bad idea. With OpenGL, you get platform independence which I think is more important than ever now that we have Android and perhaps even Steam in a not too distant future. The reason why Direct3D became so popular was a result of a FUD campaign that Microsoft spread about it; that it was buggy, Vista wouldn't support it, performance not as good as Direct3D (when in fact D3D was SLOWER than OpenGL), not as many features (when in fact OGL had MORE features than D3D and is still developing ...).

The companies know that the gaming market is big and lucrative, so the do what they can to hog a share into this market and force users to their software and hardware products through technological lock-in strategies. Now Apple is trying to do a similar thing releasing METAL that supposedly is "faster" than OGL and having more features than OGL while showing fake "benchmarks" comparing their performances in favor of METAL while real benchmarks show that OGL is better and more feature rich than METAL. So it remains to be seen whether developers have wisen up to these companies' scams or will fall for the FUDing once again. I'm quite sure that companies will spend HUGE loads of money to make that happen.

OpenGL is used extensively in the professional market, has been around for a long long time, is still actively developed and it is going to be around for another long long while. And I think there will be tools (such as MetalGL) available to make OpenGL applications interface efficiently with both METAL and D3D.

Also, more physics would be nicer too. I constantly see all those fluid simulations like water splashes, deformations, waving hair/textiles and so on but never see them reaching the games. I have not yet seen caustics in games either. I've seen real-time raytracing engines but I don't think we are there just yet but who knows, instead of going for multiscreen 4K, that extra GPU performance could be better spent on lower screen resolution but more "eye candy". And when one wants to go for competitive gaming one could strip off the eye candy and go for 4K instead.
 
I think CryEngine is the best. A game that come out in 2007, and still looks pretty awesome? Very demanding on faster hardware but designed to give a bit of slack to slower one? I'd say that'd be quite a feat.

For me, those engines are a turn-off which require high-end i5's and GTX 780's as minimum requirements. Just Cause 3 and Assassin's Creed Unity/Syndicate come to mind. Also ones which gobble up a lot of VRAM(Shadow of Mordor).