On Physics Engines

[cool_breeze]

http://www.edbordenblog.com/2009/04/future-games-will-battle-over-physics.html

Insomnia readers may be interested in the recent debate over future use of computational power in games. Do we need better physics more than we need better graphics? My preference for future development is biomechanics. I am sick of unsuitable use of rag doll physics, which is really only meant for unconscious people and animals not giant robots knocked about by explosions (worst example: "Transformers: Prelude to Energon"[PS2]).

[Worm]

Do we need better physics more than we need better graphics?

This question is rather disingenuous considering that "better physics" are used for cosmetic purposes 90% of the time (fluttering cloth, debris from explosions, water effects, ragdolls, etc.). It's not something you can answer, anyway--or do you think there's a single approach that can be applied to all games?

The article you linked should go in the lol thread:

...Far Cry 2, which I recently blogged about being at the absolute cutting edge of this trend: interactive flora and fauna, environmental destructibility, realistic environmental fire behavior, integration of real time and weather.

Anyone who's played the game would know that weather affects nothing other than tree branches blowing in the wind (cosmetic). Only small wooden structures are destructible, and never in a way that can kill enemies or open up new areas--it's certainly less robust than Crysis. The propagating fire can hardly be called "physics," since all it does is kill enemies, set off explosives, and turn structures black (cosmetic).

Because [hardware] technology is so far ahead of the software that game developers are putting out at this point, that no one needs high-end hardware to play games!

This is laughable on the face of it. Try playing Supreme Commander on that budget PC and tell me that we don't need more polygon-pushing power. Apparently, he thinks that 40 fps in Crysis at sub-maximum settings is more power than anyone will ever need.

Barely over a year ago, before NVIDIA bought Ageia and its PhysX engine, no one was really paying attention to physics functionality.

Yeah, because no one was talking about Half-Life 2 and Havok Physics in 2004, or Euphoria in 2006. It's been a talking point for over half a decade.

[icycalm]

After Worm's comments I don't think I'll bother reading the article, but I'll say this much: more complex physics engines are indeed the way forward -- as long as we have decided that we really want to go forward (which is indeed what the big companies will end up deciding eventually, because they have no other choice). So in other words, in the long run, more complex physics engines are the only way to keep increasing the complexity.

At the end of this road lies an almost perfect double of the real world. "Almost" because there is one characteristic of the world which is impossible to simulate: death.

[FallingUp]

Well, you could simulate it, but I don't think that's a game anyone would want to play.

[icycalm]

You don't understand the concept of simulation. If we set the game up so that if you die in the game you also die in reality, that is not simulation. That is still real death.

[FallingUp]

That's not what I was getting at.

Consider roguelikes. When your character dies, it's permanent. Of course, this isn't really simulated death because you can just roll another character. A simulation of death, however, would be a game that you can only play once. If you die, that's it. You don't get to play the game again for the rest of your life.

Actually enforcing this restriction would be a pain, and nobody would ever buy a game like this, but it's theoretically possible.

[icycalm]

That's not what I was getting at.

Then I guess I overestimated you. Because what you were getting at was extremely naive and ignorant.

Consider roguelikes. When your character dies, it's permanent.

In videogames there is no "your character". "Characters" exist in novels and in movies. In videogames there is "you". So when YOU die in a roguelike, or in Tekki, or in an arcade game, or in whatever other videogame you can think of, YOUR death is indeed not permament. It is in fact not even a death -- when you use a word like "death" in reference to what happens in a videogame, you are simply abusing language. Now for everyday conversation that might be acceptable, but when you are trying to get to the bottom of things, to understand what is really happening while you are playing the game, you must use words with precision.

Actually enforcing this restriction would be a pain, and nobody would ever buy a game like this, but it's theoretically possible.

More naiveness lol.

[FallingUp]

So when YOU die in a roguelike, or in Tekki, or in an arcade game, or in whatever other videogame you can think of, YOUR death is indeed not permanent.

Which would be why it is a simulation of death and not real death. You're confusing the two.

If the term 'death' is an abuse of language when referring to what happens in games, what is the correct term?

[Bradford]

If the term 'death' is an abuse of language when referring to what happens in games, what is the correct term?

lol

How about "Game Over"?

[cool_breeze]

Do we need better physics more than we need better graphics?

It's not something you can answer, anyway--or do you think there's a single approach that can be applied to all games?

The original context was first person shooters, and the question was considered from the point of view of game developers in an attempt to gain insight into the future trends of FPS game design. A broader discussion covering all games would have to ask the question for each genre or subgenre of game if not for the real or imagined sequel of every game ever made since "Doom". Not only that, the question would be answered differently by players, game designers, game producers, game publicists, game artists, etc. so who is asking the question? I am interested in the answer from the point of view of an expert FPS player, not anyone involved in game creation.

By "better physics" I mean "more complex rules in the visual medium (ignoring audio for now) that give rise to the need for ingame abilities or problem-solving acts during player interaction with the game world." A new rule that does not require a skill or problem-solving does not count.

By "better graphics" I mean "more complex rules in the video medium such as improved rendering, ray tracing, camera behaviour that facilitate the player learning skills but do not in themselves give rise to a skill."

By "more complex" graphics or physics, I mean, making more use of the computational power of the platform for the improvement of one to the exclusion of improvement in the other.

At this point, I am ready to clarify the question for expert players of FPS games:

"Given limited computational resources for games in the near future, would FPS players prefer more complex physics or more complex graphics in future games?"

[icycalm]

So when YOU die in a roguelike, or in Tekki, or in an arcade game, or in whatever other videogame you can think of, YOUR death is indeed not permanent.

Which would be why it is a simulation of death and not real death.

There is no such thing as a "simulation of death". There is, for example, a "simulation of a jump". We can trick our senses into believing that we are jumping even though we are not jumping. That would be a simulation of a jump. And similarly for the simulation of running, shooting, talking, etc. But to simulate death, to trick our senses to think that we are DEAD while we are not dead is impossible. WE DON'T EVEN KNOW WHAT DEATH FEELS LIKE FOR FUCK'S SAKE -- HOW WOULD WE EVER BE ABLE TO SIMULATE IT?

I cannot explain this in plainer words. If you are not getting it -- too bad. Just don't ask me to explain it to you further, because I can't.

You're confusing the two.

There's no "two" -- there is only one. And as for the implication that I am confusing anything -- that's rich coming from someone who can't tell the difference between reality and a fucking game.

If the term 'death' is an abuse of language when referring to what happens in games, what is the correct term?

Like for instance in Chess. No one talks about death in Chess -- the pieces do not die -- they are simply lost -- you lost your queen, for example -- she did not die. And when you lose a round of Chess you did not die -- you simply lost a round of Chess for fuck's sakes.

But when it comes to videogames, everyone immediately becomes a fucking buffoon, and their entire brains are tricked by a few fuckin' textures into hallucinating about "sex" and "death" and whatever the fuck.

[icycalm]

A broader discussion covering all games would have to ask the question for each genre or subgenre of game if not for the real or imagined sequel of every game ever made since "Doom".

You seem to be under the impression that Doom was either the first videogame, or the first FPS. It was neither. Swear to god: look it up.

Also, lol at your naive reasoning.

By "better physics" I mean "more complex rules in the visual medium

Stay of the definition game, because you suck at it.

By "better graphics" I mean yadda yadda yadda

Yes that is why I built this forum. To have random idiots come in here and spew out some bullshit about what BETTER GRAPHICS IS SUPPOSED TO MEAN.

At this point, I am ready to clarify the question for expert players of FPS games:

You can take your stoopid question and your "expert FPS players" and stick them up your ass.

"Given limited computational resources for games in the near future, would FPS players prefer more complex physics or more complex graphics in future games?"

And here is why your question is stupid, and more importantly uninteresting. On the one hand, it is plain that, in the long run, better graphics presuppose better physics and better physics better graphics. On the other, why should anyone have to choose? We never had to choose before -- why should we have to choose now? NEWSFLASH: This is how things have always worked, and how they always will: someone makes a neat engine which allows the players to perform a few extra tricks, then all the other companies copy this engine, then they fight amongst themselves for market share by adding bells and whistles (better graphics), then everyone gets bored with the old tricks and makes a new engine that does a few more tricks, and so on and so forth until we reach the doubling of the world I mentioned above.

End of story. It's really not that complicated.

[icycalm]

On the one hand, it is plain that, in the long run, better graphics presuppose better physics and better physics better graphics.

For those who don't understand this, consider: what does "photorealism" mean in a 3D environment?

Read also this:

If you are still wondering where I am going with all this let me make my position clear: I believe that the PS3 and the Xbox 360 are not powerful enough to handle true HDTV resolutions, and at the same time deliver the large variety of new effects necessary to approach photorealism. The quest for photorealistic graphics is, after all, the main reason for designing new consoles every four or five years -- once it is accomplished the console cycle will grow longer and everyone will start concentrating more on peripherals and human/machine interfaces (Nintendo simply jump-started this process with the touch screen of the DS and the remote of the Wii, because they didn't have the R&D budget to stay in the graphics game).

http://insomnia.ac/commentary/not_powerful_enough/

[walrusdawg]

At the end of this road lies an almost perfect double of the real world. "Almost" because there is one characteristic of the world which is impossible to simulate: death.

How do you simulate pain?

[infernovia]

Even though my example is not in video games, I do believe that they are related. I am speaking of the phantom limb syndrome, a simulation that tricks the body to (mostly) painful sensations that it cannot physically experience.

[Four]

How do you simulate pain?

Watch a football (soccer) match to find out!

I am speaking of the phantom limb syndrome, a simulation that tricks the body to (mostly) painful sensations that it cannot physically experience.

You cannot trick your senses in feeling pain because pain is a trick of the senses to start with. If you can feel it, it's as real as it'll ever get.

We all are already simulating reality with our brains and that simulation is the only thing we perceive of reality. If we cannot simulate something then we will not acknowledge its reality. (See magic tricks, and all the various ways our brains are shown to be incapable of processing reality.)

I don't understand what is impossible about simulating death. By definition, death happens to others, and it's something we can see, smell and whatever else we can do with someone else's dead body.

[icycalm]

You cannot trick your senses in feeling pain because pain is a trick of the senses to start with.

Pain is not a "trick of the senses" -- it is a sensation. There is nothing tricky about it. And exactly because it is a sensation, you cannot simulate it in the highest sense of the term "simulation". You can fake it (as in soccer), in order to fool others, but you cannot fool yourself. You either feel pain or you don't, and that's that. So you are correct in saying:

If you can feel it, it's as real as it'll ever get.

Then, of course, you proceed to abuse the term simulation and to muddle and obfuscate the entire issue:

We all are already simulating reality with our brains

We are not simulating anything with our brains. The brain accepts the data from the senses and constructs a model of reality from them. This model, which will of course be necessarily inaccurate, it then uses in order to decide its next move.

and that simulation is the only thing we perceive of reality.

No. The only thing we perceive of reality is the input of the senses. The MODEL of reality we do not perceive -- we CONSTRUCT it.

If we cannot simulate something then we will not acknowledge its reality.

I like the solemn and matter-of-fact tone some of you idiots take while blurting out fucking nonsense. It's charming.

(See magic tricks, and all the various ways our brains are shown to be incapable of processing reality.)

Our brains are perfectly capable of processing reality, thankyouverymuch. Or at least mine is -- I don't know about yours.

I don't understand what is impossible about simulating death.

Too bad. I've already done my best to explain it. Try the IGN forum next.

By definition, death happens to others

By the definition of a fuckin' retard, maybe.

He posted a bunch of other nonsense which I couldn't bear seeing on the page (let alone reply to) so I deleted it. Four, do me a favor and do not post in this thread again.

[icycalm]

Oh, and as for the "phantom limb" syndrome, I really know next to nothing about it. Right off the bat it sounds like charlatanism to me, but I'll look into it.

[infernovia]

Then you couldn't say that my example is a simulation then. I made the mistake of thinking that a tricked sense would be considered a simulation.

Phantom limb syndrome deals with the wiring of the brain. For example, a person with a limb cut off can "feel" pain in their arm if you touch their face. When you touch different part of the face, they feel like they are being touched in different part of the limb.

It's interesting that this pain decreases when you construct a virtual reality where the person can see themselves moving two limbs.

[Bradford]

we define death by these observations, and that can be simulated.

If you want answers and clarifications, please think about what you are asking for longer than two seconds. Of course we can simulate the observations we have about death. As in, we could simulate what it is like to be alive and present when someone else dies. That's not even close to a simulation of what it is like to die.

If Icy hasn't totally lost patience with this by now, perhaps he will explain the difference with more precision.

[icycalm]

I deleted his post that you replied to and banned him. If people do not want to listen to me, that's fine, but in that case they cannot post in my forum. I am not obliged to answer anyone's questions -- I will only answer the ones I want to. Whoever presumes to demand answers from me, or anything else really, will simply be banned.

Back on topic, I'd like to offer a couple of clarifications. First off concerning this:

But to simulate death, to trick our senses to think that we are DEAD while we are not dead is impossible. WE DON'T EVEN KNOW WHAT DEATH FEELS LIKE FOR FUCK'S SAKE -- HOW WOULD WE EVER BE ABLE TO SIMULATE IT?

In fact death, at least as we currently understand it (and we will never understand it better...), feels like nothing -- and that is the problem. Death is not only a cessation of the flow of data from the sensory organs (which in itself could, in fact, be simulated), but at the same time also the end of the functioning of the very instrument which accepts this data and interprets it (the brain). Death is a singularity -- it cannot be doubled, it cannot be exchanged -- it cannot be simulated. You are either alive and very well aware of this fact (or at least of some facts -- even unconscious people are aware in many respects) -- or you are dead and aware of nothing. There is no middle state here. It is impossible to think that you are dead while not being so, because dead people don't think. "I think, therefore I am", said Descartes, but he might as well have said "I feel, therefore I am", and in fact feeling and thinking are one and the same thing -- but that is another long story...

On to the subject of pain.

I mentioned earlier that the only aspect of reality which could not be simulated is death, and then walrusdawg brought up pain. And indeed pain cannot be simulated, but for a very different reason -- for the same reason, for example, that hunger or cold or warmth, etc. cannot be simulated -- because they are sensations, and as such constitute the very object of simulation. The sensation can never be simulated -- what CAN be simulated is the real event that causes the sensation. For example: I can pinch you in the arm with a needle, and this event will generate an electrical signal that will travel from the appropriate nerve(s) in your arm all the way up to the brain. OR, I can hook up that nerve to a computer, and send a signal indicating the exact same level of pain that you would have felt if I had pinched you in the arm with a needle. In the first case the event is real, in the second it is simulated. But pain exists in both cases, and it remains the same -- indeed it HAS to remain the same for the illusion to be perfect.

So the object of simulation is never to simulate pain -- it is to simulate the event which causes pain. It is never to simulate the sensation, it is to simulate the event which causes the sensation.

Recreational drugs and anti-depressants do almost the same thing: they simulate joy. "Almost", because with them you are actually messing with the brain's chemistry instead of merely with the senses. No doubt the day will come when drugs will be unnecessary and it will all be done through computers. And if for some reason this day doesn't come for the beings on THIS planet, it will undoubtedly have already come or will come for beings on other planets.