Gasoline in a barrel + a bullet

actually fenric:
Originally Posted by nw909
Its impossible to calculate fluid physics right now.

then you said it was possible.
you said calculating the physics of moving liquid was possible.

it aint, end of story, unless quantum computers become a reality well never be able to do those kind of calculations. you may aswell set up a big bang simulation lol
 
That's for truly realistic simulations. But shortcuts most certainly allow realistic real-time simulations to be possible in a few years time. And if all you care about is the surface of otherwise contained fluids, you could probably do it right now if you wanted to throw all your resources at it.
 
bigun said:
actually fenric:
Originally Posted by nw909
Its impossible to calculate fluid physics right now.

then you said it was possible.
you said calculating the physics of moving liquid was possible.

it aint, end of story, unless quantum computers become a reality well never be able to do those kind of calculations. you may aswell set up a big bang simulation lol
Calculating the physics of moving liquid IS possible, which is what I have maintained from the start, its even possible in realtime, you just can't currently do it in realtime with everything else going on at the same time

I never once said true liquid dynamics was possible. So your claims are baseless.
 
vegeta897 said:
I would just be happy with, and I don't think HL2 has this, atleast not in the E3 demos, but it doesn't "flow" or anything, but ripples and such caused by objects going into/out of it or whatever. Plus all the other little splash sprites and such.

I am looking at this demo here- http://cgi3.tky.3web.ne.jp/~tkano/tlwater.shtml

It's pretty cool, and would be plenty for HL2. (Also isn't pc intensive at all)

downloaded and watched that demo, its pretty damn impressive. if hl2 had that, thatd be freackin awesome. (or any game for that matter). Although at times the demo didnt look to realistic when i made big splashes, it didnt look right. but for the most part it looked incredibly realistic. i almost wanted to drink it, lol.

about it not being cpu intensive.. if there ever were to be a lot of action in 'that water' it just might bring any system to its knees.
 
be more specific - what do u mean by the physics of moving liquid? the individual collisions between molecules?
 
I'm not sure about ingame, but in real life it would be random. Just because a bullet hits something does not automatically = sparks
I shot my Uncles old car (which still had gas in it) with almost 40 M4 Carbine (7.62mm?) rounds in the gas tank and nothing happened. Most of the explosions IRL don't happen from the gas itself anyways, most of the time it is the fumes.
 
im not sure if the barrel would explode. a bullet is an inert object if it hit the barrel without sparking the barrel and just passed straight through then its possible it might just puncture and spil the gas everywhere

i call it gas but im from the UK so its Petrol. :)
 
If you shot a barrel full of gas it would never explode. For one thing, gasoline in liquid form is not explosive. If you fill an open top can with gasoline and light it, the gas does not explode. It just burns. Fumes explode. Secondly, there need to be 3 things to start combustion. 1 heat, 2 fuel, 3 oxygen. The bullet puncturing the barrel can produce 2 of those - the gas is the fuel obviously, and the bullet may produce a spark that is hot enough for ignition - but not all 3. There is simply not enough oxygen available in that senario for the gas to ignite and blow the barrel apart. The bullet hole would be too small to supply sufficient oxygen and any combustion that could take place would burn itself out before it ever got started.

Oh yeah, and you can simulate fluid dynamics in as true as sense as they will ever need to be. You don't need to simulate individual molecule collision. A fluid is simply a mass with a specific volume but no specific shape. It fills whatever container it is put in. All you have to do is simulate the boundries of that mass and its collision with any other object while maintaining a set total volume. You don't need a quantum computer, and you can definately do it on a computer that can simulate a nuclear explosion. If you'd like me to believe otherwise, please produce a PhD in either computer science or physics.
 
they don't have a liquid dynamics sytem in source.
 
yes but thats not fluid dynamics is it cyanide? i mean lets say you give a material all those properties you state. well now lets say you shoot a bullet into a barre; of this material. will the material realistically create concentric flow circles which reflect from the sides of the barrel and refract over each other?

and i dont have a phd in anything but i do have an A @ A Level Physics and a degree in Economics, and am currently studying for a masters at the LSE if that helps??
 
bigun said:
yes but thats not fluid dynamics is it cyanide? i mean lets say you give a material all those properties you state. well now lets say you shoot a bullet into a barre; of this material. will the material realistically create concentric flow circles which reflect from the sides of the barrel and refract over each other?

Yes it will, if the geometry is complex enough that is.

But saying that liquids are impossible because you can't simulate all the atoms, is like saying that you can never create a wall because it's impossible to simulate the atoms in it. There's no reason why you should.

But I don't think you can simulate all properties of water with geometry, for instance the splashing of water with droplets flying out and merging with the water again is impossible I think. That should be done with particles I think, who form bigger blobs when close enough to each other.
Fenric, when you talked about liquid physics being possible in realtime, did you mean just the displacement of the water geometry or also splashing and the formation of droplets and stuff?
 
PvtRyan said:
Yes it will, if the geometry is complex enough that is.

But saying that liquids are impossible because you can't simulate all the atoms, is like saying that you can never create a wall because it's impossible to simulate the atoms in it. There's no reason why you should.

But I don't think you can simulate all properties of water with geometry, for instance the splashing of water with droplets flying out and merging with the water again is impossible I think. That should be done with particles I think, who form bigger blobs when close enough to each other.
Fenric, when you talked about liquid physics being possible in realtime, did you mean just the displacement of the water geometry or also splashing and the formation of droplets and stuff?
displacement of water geometry, and formation of droplets and stuff

It's all quite possible, pre-rendered/caculated and in realtime

Those who say it can't be done atall google for realwave (does the displacement) and realflow (the particle effect of liquid and/or gas) and Lightwave ParticleFX2 and those with Softimage XSI|EXP can use particles in that and get the effect they want

NatureFX by Dynamic Realities could do displacement using an object

Maya can do it

Dynamic Realities also released Napalm which could do the partile effect of liquids etc.

There's a realtime shader demo out there that shows an object effecting the water surface

Morrowind had a basic (though very cool looking) DX9 version of its water displacement

Pirates of the Carribean, the scene where Barbosa drinks the wine, that was a mix of particles and footage and painting to get the effect

Abyss, very old now, but that did it too

Shrek did it

Monsters Inc did it I think

Toy Story did it

Bugs life, Antz etc. had it

It's quite possible, has been for years in the pre-rendered stuff, realtime can do it now, just waiting for computers to get a bit faster and you'll be able to have it inside a game with lots of other cool stuff going on at the same time
 
Not everything has to have realistic physics applied to it, we've seen that sort of thing before.. some sort of liquid spilling outta a barrel when shot. It could be done nicely without all the heavy mathematics.

:)
 
First off, as a few people have quite rightly pointed out, shooting a barrel full of petrol, or most other liquid fuels, will not make it explode. Ignition depends on the presence of flammable vapours, where the density and surface area of the fuel in contact with the air is sufficient to be combustible.

Secondly, I don't know what people mean when they talk about "true" fluid dynamics simulation (I've never in my life heard of such a term), but reasonably accurate simulations are more recently possible. Finite Element Analysis can be used to do this, with quite accurate results. This requires a huge amount of processing resources though, and there's rarely any reason to simulate all of the physical properties of water, such as surface tension or thermal convection. There is no purpose to spending the time and resources simulating elements down to individual molecules, because beyond a certain point there is little to no useful gain in accuracy when weighed up against the amount of processing done.

Whether simulations are "real time" or not merely depends on the amount of processing power you have available, and the simplicity of your assumptions. In terms of computer games, you're performing very simplistic calculations on aggregate volumes, or on a relatively small number of point mass objects, and involving only a handful of variables, such as volume and mass. Usually this involves just simple Nnewtonian physics and linear equations, but it works on-screen because that's all you really need for it to look convincing.

All too often I see gamers demand "realism", without fully understanding what it is and what the implications are. I know that I'd take fun over "realism" any day - the only point to having a game with more consistent physics or better drawn graphics is to make it more convincing, and you don't need blummin' fluid dynamics to do that.
 
to the person who said those ripple effects weren't in any game... look at UT2003 :p

Sorry if I read the post wrong... I'm tired and crazy
 
amneziac85 said:
LoL, weve had exploding barrels since DOOM.

We've had exploding barrels since the very first day flamable liquids were stored in barrels. Haha, my great wit.
 
Seppo said:
We've had exploding barrels since the very first day flamable liquids were stored in barrels. Haha, my great wit.

Nooooo, my post was great wit, yours was just overkill that killed the joke. Dont try so hard next time.
 
Brian Damage said:
I wonder if a barrel of oil would really catch fire upon being shot in real life... I'm pretty sure bullets don't actually make sparks... y'know... being made of soft lead and all...

fumes of oil are the flammable part no?

and lol to the guy who said we'd need a transparent barrel to see gas!!!!!lol10101010111!!
 
Neither of those posts were witty...

Anyways, I don't think you could make a liquid leak out of a barrel in the HL2. Well atleast in "physics". HL2 still uses decals for bullet holes, which means that there is no "physical" hole for the liquid to leak out of.

You don't have to have supremely detailed physics here. We don't need the grand unification theory for playing a video game. As it stands, I bet the physics engine in HL2 will give all of our computer a good workout.

I think in a couple of years it would be cool to have some semblance of fluid dynamics. Throwing in Pascal Principle or bernoulli's principle shouldn't be that difficult to incorpate and would create some cool puzzles. The problem I see is that 2 things will eventually happen. 1) You are going to have great physics all the way around, or else something will seem weird, which means computers will have to be really power ful. 2) Too much physics are no fun.
 
Too much physics are fun as hell. In 2010 I'll frag people by dynamically demolishing an entire building, or group of buildings and a water tower, or something...
 
"Gasoline in a barrel + a bullet"
boooom!
freeman:burn!
 
yes we can!

[Edit]: Okay, okay, so that was spam. :imu:
[Edit 2]: Why would you have a barrel of gasoline in the middle of a city to be shot at anyways? :bounce:
 
Why, no combine army is complete without hordes of poorly-aiming minions and the ocasional stockpile of gasoline!
 
With gas prices the way they are, why the HELL are you shooting barrels full of it!?
 
Maybe the objective in HL2 is to save as much of it as possible from the combine's clutches, preventing economic recession? Gordon could have a little physically simulated calculator, on which he can make note of how much change he's saved by killing combines instead of buying from exxon.
 
jonnyapps said:
fumes of oil are the flammable part no?

and lol to the guy who said we'd need a transparent barrel to see gas!!!!!lol10101010111!!

My point was, fumage notwithstanding, no sparky, no boom-boom.
 
Does HL2 support TL water? I mean, it is not really very dynamic. It cant splash and all, it just waves heavaly.


...time to email gabe...
 
blahblahblah said:
Combine = OPEC :O

You could have used the spoiler tags! Geez, where are all the manners? What the hell is wrong with you people?

By the way, does photorealism have anything to do with realistic physics, or does it just mean completely realistic graphics? 'Cause you know how Carmack stated that photorealism could be possible to achieve in like ten years or something. Would we then have realistically simulated liquids too?
 
from what I've gathered, photorealism is just it looking real in a still image. Motion realism, combined with photorealistic graphics should be called something like "video-realism", but people call that photorealism too.
 
Back
Top