Sony @ GDC: Phil Harrison's Keynote

[scooby_dooby]

For me, the difference between Outrun (or more Lotus Esprit Turbo Challenge in my case), GT2 and GT3 shows how improved physics makes for a better game that's more challenging and more fun. Having written code using physics libraries, I can assure you that just being able to interract openly with a virtual physical world is extremely engaging. The possibilities for new and very fun gameplay from advanced physics is very real. Whether devs manage to use it is a different matter, but thinking there's no point researching and developing better physics is passing up a great opportunity.

I think the point is we can do all of these things now so why do we need so much more power? How much of a difference will it reallly make, and will that difference be worth the cost?

Oblivion fully physics based world, Forza & GT4 are both extremely realistic car handling. And honestly, interacting with a physics based world is cool at first, but on a 40+ hour game like oblivion the novelty wears off reeeal quickly.

 

[!eVo!-X Ant UK]

I think the point is we can do all of these things now. Oblivion fully persistent world, Forza & GT4 are both extremely realistic.

And honestly, interacting with a physics based world is cool at first, but on a 40+ hour game like oblivion the novelty wears off reeeal quickly.

That is true, i think the best physic's will the be the one's that are'nt so obvious.... MGS4 dust, wind, paper flying in the wind. It's subtle stuff like that they will make the difference. Maybe not help the gameplay but it might depending on the genre type and developer imagination.

 

[London Geezer]

Well my 2 cents is this, I think that physics are already very good in games, and that there are many much more important advancements in gameplay that we should focus on first before physics, games need better storylines, better Ai, they need more cinematic moments like COD2(which 99% of games can't pull off), they need more non-linear storylines and more control over the story and more replayability.

All these things I think are much more deficient in most games than physics.

If you look at Oblivion, it has a whole physics based world, everything is realistically rendered and you can interact with it. Yesterday I fought a stone beast made of hunbdreds of stones, all whirling and swirling in cyclones as I fought him, the physics were great, and this is a basic PC game from early 2006.

How much better do we really need? The thing about physics is, the effects aren't always noticeable, for 90% of a given game you may not notice any physics, unlike other graphical upgrades which are present at all times. The other thing is, improvements in physics are very hard to notice, like would I really have nopticed if the stone monster had 300 stones instead of 100? Would I really care?

Physics seems to have huge diminishing returns, and we already have good physics processing capabilities, I don't see what more we need.

As far as motorstorm, I don't see why the ruts/track affecting handling gameplay elements couldn't have been done in previous generations, the splatter and eye candy seems to be all that's really 'enabled' by more processing power.

You're saying we don't need any more power for physics? We don't need proper fluid dynamics? proper [insert name here] dynamics?
There's still so much room for improvement, i recommend you watch some of the fluid dynamics videos from AGEIA.

 

[scooby_dooby]

I'd disagree totally. Looking at some of the stuff coming from AGEIA etc. it's pretty obvious what more power is bringing. With more power you can either do more of the same as before, or do new more complex kinds of simulation. "More of the same" is arguably very noticeable depending on the context, while the latter will be where the more exciting stuff still will probably happen.

The problem is when game developers use these cool techniques, they almost always become nothing more than gimmicks, that are cool the first 3 or 4 times you see them and then quickly lose their novelty.

I think it's because in real life you don't see alot of examples of cool physics going on, so developers have to explicitly create opportunites to 'show off' the physics engine, this always ends feeling tacked on.

I think all physics need to do is stay realistic enough that you aren't taken out of the game, and that's done very well already in games like Oblivion or Fight Night 3.

Sure it could be a little better, but there are SO many things I think are more important for developers to focus on other than marginal improvements in already good physics.

 

[Edge]

Oblivion fully physics based world, Forza & GT4 are both extremely realistic car handling. And honestly, interacting with a physics based world is cool at first, but on a 40+ hour game like oblivion the novelty wears off reeeal quickly.

Physics like High Dynamic Range rendering or anti-aliasing is not the meat of a game, but the icing. For the most part, you could live without it, but it's nice to have.

To me having water that behaves like real water, in flowing over the landscape, and pooling, adds a lot of immersion to a game, and that requires a strong physics engine, that CELL with all it's SPUs will excell at. Having realistic water is not marginal.

 

[expletive]

For me, the difference between Outrun (or more Lotus Esprit Turbo Challenge in my case), GT2 and GT3 shows how improved physics makes for a better game that's more challenging and more fun. Having written code using physics libraries, I can assure you that just being able to interract openly with a virtual physical world is extremely engaging. The possibilities for new and very fun gameplay from advanced physics is very real. Whether devs manage to use it is a different matter, but thinking there's no point researching and developing better physics is passing up a great opportunity.

Yes i agree, and to scooby's point, we've already had these things for years now. What is this new, highly computational, physics world giving us that we arent already getting that will actually enhance gameplay? Is it somethign i really want to spend $300 for an add-in on? How about an SPE or 2? Or some Xenon core space? If youre telling me ive gotta give up some horsepower (or in the case of the PC, money) to do this stuff, it better count for somethign during the game!

I understand your point though, and i've already conceded this generation may be the puberty for physcics where we need to go through it but wont really come of age until next gen (i.e. the 'global illumination' version of physics).

We'll have to agree to disagree then. I really disagree most wholeheartedly. The scope for some interesting scenarios is definitely widened by including something like this. Gamespot mentions one..imagine also mixing in a weather system, whereby the track can change or not with passing rain showers (i.e. after rain, you can manipulate the track more easily to your convenience, and perhaps not to that of your opponents, also forcing you to rethink the best racing lines each time the track changes etc., or the the potential for a well timed nudge into an opponent as their negotiating trickier 'ruts' that could leave them with in more trouble). Asides obviously from just the aesthetic contribution to the experience...

We couldnt do all that stuff already? Says who? There hasnt been varying 'slickness' on driving games? I know I've been nudging people for years now!

I dont see why i need to spend $300 or any additional horsepower for a certain area of a dirt racetrack to be more 'slick' then another, making my competition more prone to 'nudges'. Just sounds like lots of marketing to me.

 

[scooby_dooby]

You're saying we don't need any more power for physics? We don't need proper fluid dynamics? proper [insert name here] dynamics?
There's still so much room for improvement, i recommend you watch some of the fluid dynamics videos from AGEIA.

I'm saying that I think that it should be way down on the priority list when we have games who can't even put together a decent storyline, have completely linear missons with 10hours of gameplay, total lack of quality cinimetic moments in-gam, and AI that can't hear a gunshot 10 feet from their head.

Proper ____ dynamics is just not nearly as important as fixing the other gaping holes in the majority of game design today IMO.

 

[wco81]

They just started scratching the surface as far as physics are concerned. How long has HAVOK and similar technologies been around in gaming?

People would rather have the status quo? Don't bother getting next-gen then.

There's no physics when a supposedly 190-pound CB can take down a 240-pound RB in a football game like he was flinging around a frisbee. Unfortunately, EA probably won't use physics in Madden any time soon.

 

[Titanio]

We couldnt do all that stuff already? Says who? There hasnt been varying 'slickness' on driving games? I know I've been nudging people for years now!

I dont see why i need to spend $300 or any additional horsepower for a certain area of a dirt racetrack to be more 'slick' then another, making my competition more prone to 'nudges'. Just sounds like lots of marketing to me.

Nudging a bike into the side of a rut I previously created with my tires, isn't something I've done before..

I've said it before, but I cannot believe people on a technology forum are coming over all technology-sceptical! Arguing that something like physics is pointless if it can't affect gameplay is akin to arguing that every graphical advance we've had in the last x years has been pointless - but physics has actually the additional potential to open up new gameplay possibilities. Honestly, you may as well just pick up an emulator and go away and play your games. You're going to see a little lost on a forum like this!

Well, actually, I can believe some are arguing these points, and I could take a couple of stabs as to why the people who are talking it down are doing so, but I'll save my breath!

 

[scooby_dooby]

They just started scratching the surface as far as physics are concerned. How long has HAVOK and similar technologies been around in gaming?

6 years.

 

[mckmas8808]

Yes i agree, and to scooby's point, we've already had these things for years now. What is this new, highly computational, physics world giving us that we arent already getting that will actually enhance gameplay? Is it somethign i really want to spend $300 for an add-in on? How about an SPE or 2? Or some Xenon core space? If youre telling me ive gotta give up some horsepower (or in the case of the PC, money) to do this stuff, it better count for somethign during the game!

I understand your point though, and i've already conceded this generation may be the puberty for physcics where we need to go through it but wont really come of age until next gen (i.e. the 'global illumination' version of physics).



We couldnt do all that stuff already? Says who? There hasnt been varying 'slickness' on driving games? I know I've been nudging people for years now!

I dont see why i need to spend $300 or any additional horsepower for a certain area of a dirt racetrack to be more 'slick' then another, making my competition more prone to 'nudges'. Just sounds like lots of marketing to me.

I don't get some of you guys. Why don't some of you argue so hard against HDR lighting, 4X AA, blahX AF? Why is it that when it comes to physics all of a sudden it's a marketing ploy?

 

[Titanio]

6 years.

This tells us what about the maturity of physics in games, exactly? We've been rendering in games for 20-odd years now, but I'd be strung up if I suggested that realtime graphics was a solved problem!

 

[Edge]

You need a powerful processor to have physics. You might say CELL will allow the imagination of developers to really explore what to do with physics.

CELL is the beginning of the next generation when it comes to physics. That's why Sony is hyping this up, as they know with CELL they have an advantage in this area.

CELL is a physics monster.

 

[scooby_dooby]

Physics like High Dynamic Range rendering or anti-aliasing is not the meat of a game, but the icing. For the most part, you could live without it, but it's nice to have.

But there's a big difference though. HDR and AA affect the game 100% of the time in all scenes. Improvements in physics are only evident in specific moments throughout the games (usually) making them seem much more gimmicky.

Also, lighting and aliasing last gen was horrible and desperately needed improvement, physics on the other hand were already pretty good.

 

[scooby_dooby]

This tells us what about the maturity of physics in games, exactly? We've been rendering in games for 20-odd years now, but I'd be strung up if I suggested that realtime graphics was a solved problem!

Just answering a question!

 

[Edge]

I don't get some of you guys. Why don't some of you argue so hard against HDR lighting, 4X AA, blahX AF? Why is it that when it comes to physics all of a sudden it's a marketing ploy?

Because they know Sony is hyping it, and CELL is far more powerful than the 360 at physics. Any advantage that Sony has is to be downplayed like crazy.

To me, it's just a sign of insecurity on someone's part. Let them say what they want, it's not going to diminish physics on CELL this generation.

 

[mckmas8808]

Well, actually, I can believe some are arguing these points, and I could take a couple of stabs as to why the people who are talking it down are doing so, but I'll save my breath!

I think we both know why "all of a sudden" technological advances are being deemed not useful. People cry over 4X AA, but all of a sudden high amounts of physics that actually change gameplay is not needed.

 

[Titanio]

But there's a big difference though. HDR and AA affect the game 100% of the time in all scenes. Improvements in physics are only evident in specific moments throughout the games (usually) making them seem much more gimmicky.

Also, lighting and aliasing last gen was horrible and desperately needed improvement, physics on the other hand were already pretty good.

Physics to date has been quite basic. HL2 was really the first to make a really compelling use of physics, at least outside some racers perhaps, and as a "proof of concept" for physics in games it was really a very recent one. If I thought the extent of dynamics in games would be limited to what we've seen in the PS2/Xbox/GC or heck, even the likes of HL2, I think I'd probably weep for the future! We're just at the tip of the iceberg. It's not about throwing more boxes around

 

[wco81]

6 years.

Havok has been around that long?

How long has HW assisted physics been around?

Why did both MS and Sony try to design a floating-point monster?

 

[scooby_dooby]

Arguing that something like physics is pointless if it can't affect gameplay is akin to arguing that every graphical advance we've had in the last x years has been pointless - but physics has actually the additional potential to open up new gameplay possibilities.

Well obviously if you completely misunderstand what people are saying it might seems like a strange argument.

It's not that physics are useless, it's that they are already quite impressive and have a very high degree of diminishing returns, in addition developers have simply not shown that they have the skills/imagination to use physics in any real meaningful way to impact gameplay. Thridly, there are large much more obvious holes in game 'immersion' that I think need to be addressed first.