Slashdot Mirror
Comparing PC Game Physics
John Callaham writes "On Wednesday we posted up comments from Havok about rival AGEIA's use of their physics processor in the PC version of Ghost Recon Advanced Warfighter. Today we have an expanded article with point-to-point comments from AGEIA that address Havok's statements." From the article: "How much interaction do you want in your PC games? It used to be that graphics were the number one factor in picking up a new game but now players are asking more and more about interactions in the environment. One company that has provided such interaction is Havok. They have developed a physics engine that has been used in a ton of games, including most famously in Valve's first person shooter Half-Life 2. Recently, Havok announced plans for a new physics engine, Havok FX, that would use Shader Model 3.0 graphics cards to further enhance game interactions and physics."
I remember going back to play Duke Nukem 3D many years ago (I stopped playing the game many, many years ago) and found it nearly impossible to play. Half Life is not unplayable, but boggy by todays standards. It is really remarkable how the physics rendering advances along with the graphics, and how important it is to game play.
"It used to be that graphics were the number one factor in picking up a new game but now players are asking more and more about interactions in the environment."
I find myself buying fewer and fewer games as time goes by, and I believe it's thinking like that that really shows why.
It's not graphics that are the number one factor, it's gameplay. There's no debate here. I want pretty visuals from movies, and I want great gameplay in my game. Don't get Blink 476 or whatever's popular for audio, either. Put your money towards making a non-buggy QUALITY gameplay experiance!
Fuckdamnit, that pisses me off.
The only people who say "How are the graphics" are going to be buying "EA *SPORT GAME* 20XX" every 9 months, anyway. So, they don't know what they're talking about.
Lets get another Fallout or a Starcraft. The graphics can be a generation or two behind as long as it's fun to play!
Just look at the Revolution and what it has to offer. Graphics aren't very improved, but the chance for gameplay being amazing is there, and that's what's important.
/rant
Not a Twitter sockpuppet... but I wish I was.
When Duke3D came out (seems like it was ages ago... forever, one might-- ouch, okay, sorry, sorry), it was right around the same time as Quake.
As a 13 year old, I figure I represented the "market" a lot more accurately than I do in my wiser (and more bitter / broke) years. It was Duke3D all the way for me, and I didn't think twice about it. Sure, Quake had better multiplayer (according to PC Gamer at least), but I was still netless at home. The novelty of shooting a wall and leaving actual bullet holes was thrilling. Getting to "play" pool, leaving footprints in the bloodstains left behind... all of this added up to a game that was fun way beyond the point where it should be. I don't mean to knock Duke3D, of course, but after the first episode the level design took a nosedive. Compare anything from the second episode to, oh, how about Healing Vats from DOOM. For me, it's a no brainer, at least when it comes to the simple question of "which of these levels is better, from a strictly looking-at-it-in-the-automap perspective". However, Duke3D's interactions had me playing, playing, playing, searching for the next deadpan line, or little "extra".
Also, this was the time when I became disillusioned with PC Gamer. I recall Duke3D edging Quake out in the ratings by about a percentage point or two. Heck, an issue or two later, Duke3D beat Quake in the "Best games of all time" list. Then, a year later, once the PC Gamer staff saw which game was completely dominating the online world, they scrambled to look "all knowing" by handing Quake the Best Game of the Year award. It'd be one thing if they alluded to some lasting value, but really it was your typical "press release" copy-paste. Fucking PC Gamer. I wipe my ass with that magazine now. Anyway...
One other thing. Is it just me, or does Capcom really have a finger on the pulse of the "heart" of physics? Every single game of theirs - well, since about the third Mega Man at least - has this perfect "feel" to it, that even makes games from genres I normally don't give a crap about (3D platformers) addictive and fun. I'm thinking of Maximo: Army of Zin here.
Anyway, I know that sounds like a lame attempt to make sure I avoid the -1, Offtopic mod, but it's the first thing that popped into my mind when seeing this. Midway's another company - for the most part, excluding the budget line, their games handle very, very nicely. Compare Blitz to Madden - and yes, I am quite aware that one of them is arcade football and the other attempts to be a simulation. Crank the Madden settings until the players are fast and whatever, bottom line is that Blitz feels nicer. Hitz beats the EA and SegaSports hockey titles hands down, largely for the same reason (even though the last version of Hitz had the worst "player editor" I've ever seen - major flaw in my book).
For a counterpoint, try comparing Bible Adventures or any of the Color Dreams games to, oh, geez, any of the major platformers. Compare a shooter from the Action 52 cartridge to Gun*Nac. Move up to SNES, compare The Combantants with Final Fight, or Kyle Petty's No Fear Racing with Mario Kart, or the second Ken Griffey game to the first. Which games "suck" by popular consensus? (PROTIP: The first games mentioned). What's a major uniting difference? The physics, the handling, the speed of play and the "oomph" behind a home-run / tight turn / nick-of-time-bullet-dodge / enemy stomp. In the first games, these are always an afterthought. I imagine the coders just kinda throwing darts at a wall, figuring "okay, player jumps, lands - now make sure all the platforms can be reached from the player's height (last step strictly optional - Active Enterprises, I'm looking at you)". In the case of the second games listed, I could easily see whole months being spent on nothing more than making incremental number-changes, in the 0.000000004 range of things. And that's why (IMO) the second games have always not only sold better, but been a better experience than the other, som
I confess I've never played the game much myself, but I do remember with a smile comments on the impressive physics engine Bungie developed for their "Myth" series of games.
One early player posted on a discussion forum that he wanted to incinerate a dwarf with the biggest explosion he could make just by surrounding it with grenades, and the resulting explosion dropped the dwarf's weapon back down out of the stratosphere several long seconds later. He did the math and calculated that the weapon was blasted straight up a couple of miles before coming back down.
Granted, that's not very realistic, but he was very impressed that the physics engine was willing and able to track a piece of debris for that long.
Physics engines are an essential component of any 3D game, and the more consistent they are with the real world the more believable the game is. You can throw everything else out the development window, I think, as long as objects bounce correctly under 9.8 meters per second per second of gravitational acceleration.
Interaction is great and all, but please give humanoid NPCs more rigid joints! It looks silly seeing them flopping around with elastic joints, or doing backflips after being shot in the face.
:-P ) - instead of this, how about we collectively lobby to get these things done right? An RTS with millions of genero-zergish units per side. A FPS with real-time rocket-based-dynamic-level-modification and death-physics so mind-jarringly violent (and bloody) that Jack Thompson and Joey "Senator" Lieberman simultaniously combust. A sports game where I can choose to attempt a brutal tackle tha
I disagree, to a certain extent.
When it comes to NPCs and enemies "reacting fo' realz" - I disagree. Sure, give them better AI (so long as "better" means "less predictable" and isn't a codeword for "can spawn other enemies to hate-rape you on sight, and requires so much processor power that there is only one enemy per stage"). But frankly, attempts have been made to make realistic physics, and without exception these games always feel muddy and unplayable. Give me Burnout Revenge over Flatout any day of the month of the week's year, kthx.
What works in the real world, with near instantaneous brain-body 3D real time control, and TOTAL SENSORY IMMERSION(TM) (note - I've patented that trademark, so now everyone has a damn good excuse to avoid the outside world) tends to take a bellyflop when you're interfacing via a mouse/keyboard/gamepad/John Romero's Magic Glowing Orb, looking at a monitor that, at best, does a good job at tricking your eyes into 2.5 dimensions.
It's been proven that people do not want "real physics" - they want "Hollywood physics". When they say "better physics", what they're saying is that they don't want paper-thin enemies who fly at 100 MPH from a shotgun blast. They want ragdoll dudes who will spin 1080 when you blast off an arm, then look at the stump, still gushing blood, and fall face-first, even though real people would scream in pain and probably not do much after the blast.
HOWEVER - when it comes to scenery physics, HELL YES. Nothing irritates me more than the Magic Unbreakable Door, found in virtually every 3D shooter. I've got rockets the size of a HUMAN BEING here. You mean to tell me a wooden door will take five of them? Other objects of note:
* - Telephone poles OF DOOM (found in most racing games, Grand Theft Auto)
* - Wooden Support Planks WITH ARMOR-ALL (found in a lot of shooters - okay, one or two shots isn't going to do much, but if I take a tommygun to a 2 by 4, the tommygun wins)
* - Ground of SOLID STEEL (almost every FPS - see my next point for more)
Dirt mound OF GOD (if I hit a dirt mound with an RPG, it should fly apart. I think that games should be REQUIRED to accurately simulate the effects of RPGs on scenery - and maybe this will keep the next five or six clone-developers from adding the damn thing. When I was in my formulative years I never imagined that I'd be saying this, but I am sick and tired of Rocket Propelled Grenades, Rocket Launchers, Giant Phallic Things Which Explode On Impact, and/or "Bazookas". They are done in every action game. They are always virtually the same. Once I play a game where a hit with a rocket will cause buildings to explode, key cards to become redundant, and mazes to be a thing of the past, I will buy back into the "Bigger and more explosive is BETTER" philosophy. And I'm not talking about Zombies Ate My Neighbors or Duke Nukem style "oh look, it's a suspicious crack in a wall, PERHAPS A ROCKET WOULD LOOK NICE HERE" linearity. I'm thinking more along the lines of the (criminally underrated) Future Tactics, except more brutal).
Anyway. Instead of worrying about the Next Big Thing, and bitching about how all games are the SAME, and becoming suckers for arm-deadening, fruitily-named attempts at brute-force "innovation" (like, uh... gee, nothing's coming to mind, so I guess this is strictly hypothetical
Never, really.
Not because its not feasible (it is, although not in the near future), but it just doesn't pay off. Pay attention to the bump mapping effects. Normal mapping was introduced - BIG impact (lighting really looks quite different, and the bumpmaps add a lot to the scene). But parallax mapping and the like? Their advantages are not as obvious, sometimes you actually have to look for it (watch the Unreal3 video, they really had to emphasize the use of virtual displacement mapping, which is just another parallax/relief mapping derivative). The point is, the cost/benefit ratio becomes unacceptable after a certain limit. Choose the techniques that have a big impact, like: the aforementioned bump mapping, cheap non-physics-based refraction (like HL2 uses), some good skies, GOOD character animation. You would be surprised just how far you can get with this. In fact, sometimes you do want cheaper visual quality, for example when you want to draw lots of entities, because better visual quality means more expensive pixel shaders, which in turn hit the fillrate limit quickly. So, if you want a space shooter with 5000 ships, you should stick to simple bumpmapping (which really makes a difference in space sims, since the hard light in space outlines surface structures quite well) and leave out the fancy parallax mapping stuff out. These kinds of stuff will become easier once batching & instancing becomes easier.
For physics, the same applies. Previously, the game world wasn't all that interactive, now I can throw around stuff. Great! Has a huge impact, changes a lot. But now, as physics advance, the advances become less relevant. At some point, it just doesn't matter if I can collide 15000 boxes in realtime.
This sig does not contain any SCO code.
As much as you can give. Physics provides depth and quality. Show the way and demonstrate this; be a legend.
Or not.
Your call.
Lurking at the bottom of the gravity well, getting old
The real problem with game physics engines is that nobody is making much money. One by one, the physics engine companies have gone out of business or merged. Havok is the last one standing, and they're smaller than they were at peak. Game middleware just isn't very profitable. Havok charged about $60,000 per game title a few years ago, and you can multiply that by the number of games they're in and figure out their revenue. The numbers just aren't that big. Their user base expects lots of support and handholding, too, so the margins aren't all that great. It's not just Havok. Middleware vendors generally are at a poor point in the food chain.
But look at Ageia. They sell to end users. That has growth potential. This is Ageia's real breakthrough. We'll have to see where this goes.
Yeah the grandparent is asking for unreasonable possibilities. I am though looking forward to games where many objects can break apart. One game a year or two ago already had destructible walls. The next step is for objects to break into smaller pieces when required to. Meaning castle walls should have their physics calculated most of the time as a solid mass. No point doing the physics for 1000 pieces of stone all the time if a wall section isn't under attack. But when a cannonball is about to hit it, replace the bump-mapped object with 1000 polygonal stones, each with its own physics.
good old AC coward to the rescue..
you do know humans ARE WALKING BAGS OF BLOOD right?
have you seen someones arm get shot off? it makes like 25 gallons of blood...
and as someone else pointed out hollywood has very rigid dead scenes and acting, as comapared to real life
and now my keyboard isnt working right fggdsg help
Better yet, no crates at all. Every bloody first person shooter has crates, more often than not because it's about the only thing you can move. They also happen to be cuboids which is rather convenient for simplifying any gravity calculations. Some games like HL2 & Far Cry try to be a bit more imaginative and you'll also see barrels and some more complex objects, but it has a way to go yet.
I'd like to see destructable landscapes & buildings where you can demolish and blow shit up and the leave a smoking ruin behind you. And scenery that interacts with you and other plays in decent ways. Then we'd be getting somewhere. The upcoming game Crysis (by the makers of Far Cry) seems to be doing some of these things and it will be interesting to see if they're the first to make a good reason for getting a physics engine. The physics engine is something that all future graphics cards should have.
If you give the player rockets, then a simple way to encourage them to use them properly is to ensure that they don't have enough to waste taking out scenery.
Or you can go the Crusader: No Remorse route. A rocket will take out any door that isn't made of Sci Fi Future Alloys (from the Future). A significant portion of keycard-doors are regular old doors. Should you choose to blast them, however, alarms are going to go apeshit, and you may not find an off-switch for quite some time.
The game worked well. It truly was "isometric action from a different perspective". Now tell me why 3D game designers can't possibly bear to part with their pacifiers and security blankets - er, I mean, their keycard hunts and mazes?
(sidenote: the one LOL moment I had during Serious Sam came when he said "sure beats finding keycards" upon blowing up a cavewall with a suspiciously-keycard-like-in-execution detonater. The irony, plus the 4th wall crashing down combined for an uproarious giggle)