Sony @ GDC: Phil Harrison's Keynote

[expletive]

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?

If i had to choose between HDR lighting or Titanio's muddy, rut-nudge, in a game i would choose HDR. What would you choose?

I think you guys are missing the point.

Developers arent saying "Here take our uber physics, theyre free!"

Theyre saying "New uber physics are awesome and you need them! BTW, its going to cost you $300, or 25% of your SPEs, or 50% of one of your Xenon cores"

I'm just asking for a justification for that cost.

Not sure why we cant just pleasantly disagree here. Instead, we're here alluding to alterior motives...

 

[scooby_dooby]

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.

Sorry but that's just weak. MS has also built a 100gflop CPU and a GPU with features like memexport that could be used to accelerate physics.

This has nothing to do with system comparisons, sometime people aren't playing devils advocate, and simply stating their opinion. Funny eh?

 

[Titanio]

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

It's like...writing off heaven before you get there, just because you haven't seen how good it can be Crazy analogy of course, but saying that because physics hasn't been terribly compelling to date implies it will not be with a broader pallette to work from is nonsense. You might aswell have said that graphics were pretty crappy on X generation of machines, thus we shouldn't try any harder, we're at the point of diminishing returns etc. It's almost contradictory!

The visual potential alone from a broader physics capability alone is worthwhile, IMO, whatever about gameplay impact.

 

[trinibwoy]

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.

Interactivity with the environment is laughable in current games - physics processing is no where near adequate. HL2's gravity gun is not the holy grail of physics interaction, think of it along the lines of hardware T&L - it's just the beginning.

Physics are gimmicky now, just like hardware T&L was gimmicky then. Once developers fully integrate physics interaction and make it a core part of the gameplay experience, those people lacking imagination will understand why it's such a big deal.

 

[wco81]

Just vote with your dollars.

Don't buy the PPU, don't buy any games which makes anything more than a cursory nod to physics, go for the eye candy games.

Stop trying to talk physics down for people who are interested in them. They don't have to justify anything to you.

 

[trinibwoy]

It's like...writing off heaven before you get there, just because you haven't seen how good it can be

Hehe, pretty good summary of what I was trying to say

 

[Sis]

It's like...writing off heaven before you get there, just because you haven't seen how good it can be Crazy analogy of course, but saying that because physics hasn't been terribly compelling to date implies it will not be with a broader pallette to work from is nonsense. You might aswell have said that graphics were pretty crappy on X generation of machines, thus we shouldn't try any harder, we're at the point of diminishing returns etc. It's almost contradictory!

The visual potential alone from a broader physics capability alone is worthwhile, IMO, whatever about gameplay impact.

To be honest, this whole debate is really strange. Let me state a few things:

1) Having tires dig ruts into the ground is a feature of the designer's vision, not a feature of the system
2) Having more power to do better physics is a good thing, period.
3) Having more power to do better AI is a good thing, period.

What, exactly, is there to debate?

 

[mckmas8808]

If i had to choose between HDR lighting or Titanio's muddy, rut-nudge, in a game i would choose HDR. What would you choose?

I think you guys are missing the point.

Developers arent saying "Here take our uber physics, theyre free!"

Theyre saying "New uber physics are awesome and you need them! BTW, its going to cost you $300, or 25% of your SPEs, or 50% of one of your Xenon cores"

I'm just asking for a justification for that cost.

Not sure why we cant just pleasantly disagree here. Instead, we're here alluding to alterior motives...

Can you please give the developers more time to program their games with great physics in mind first.

You say ----> HL2, Forza 2, and GT4 nice physics no need to upgrade
I say -------> HL 2 , Forza 2, and GT4 nice physics, but lets see better and more.

Also

You say ----->HL2, Forza 2 and GT4 nice graphics, but lets see better and more.
I say -------->You say ----->HL2, Forza 2 and GT4 nice graphics, but lets see better and more.

I again I'll ask you, why all of a sudden should we and devs stop caring about physics and care only about storylines and AI? Why can't they tackle them all at the sametime?

 

[wco81]

This has nothing to do with system comparisons, sometime people aren't playing devils advocate, and simply stating their opinion. Funny eh?

Right.

It's just a coincidence that you talk down any perceived advantage of systems competing against your favorite one.

If it's not physics, it's Blu-Ray. Or HDMI. Or probably built-in Wifi. Or lack of an online service.

Ad infinitium.

 

[Titanio]

What, exactly, is there to debate?

I didn't think much, but heh..

 

[mckmas8808]

To be honest, this whole debate is really strange. Let me state a few things:

1) Having tires dig ruts into the ground is a feature of the designer's vision, not a feature of the system
2) Having more power to do better physics is a good thing, period.
3) Having more power to do better AI is a good thing, period.

What, exactly, is there to debate?

Nothing really. Some people just don't agree with your number 2 statement.

 

[Shifty Geezer]

I think the point is we can do all of these things now so why do we need so much more power?

Physically animated characters witrh behavioural synthesis, who move over terrain, up ladders, crawl around, without any mocap limitations, who's abilities and motions are affected by the weights they carry. This 'Holy Grail' elliminates a lot of production workload in designing and scripting games and creating the visual assets for each feature. Take a game like that Island Survival game on DS. Someone had to program that when you get two sticks and rub them, it creates fire. In a physics-modelled world you could have friction generating heat, and materials burning as a result. You wouldn't then need to script a fire situation as it would happen in response to the physics engine. The developers would only need provide resources like stone and wood, and the player could find uses for it themselves. To cross a river, one player might dig at the roots of a tree and make it fall across, another might axe the tree with a sharp stone, a third use a log as a raft, a fourth find a spot to throw stones and create a dam. All these options naturally appear in a fully physically designed world. Without that you're constrained to whatever features and solutions the designers add. They will need to individual program multiple solutions, some of which may not fit in with how you would tackle the problem.

ICO is a good example of limits. There were points where you had to cut ropes to drop bridges. At first I tried and failed. Standing right next to the rope and swinging my sword did nothing. I had to attack in the right way to cut the rope in the scripted position. I actually wasted 30 minutes getting frustrated how to solve that problem the first time because the obvious solution wasn't working and I hadn't found the right attack method. In a physics simulated world all the solutions that the player thinks of, relating to their own real uderstanding of the world, become useable. From throwing a stone to make a noise to divert a guards attention, to pushing a bookshelf down a flight of stairs to squash some skeletons, to building a boiler to skirt a jet of compressed water out a nozzle to hit a small button that opens the secret door, are things that need either robust and complete physics engines. Without that you're spending human effort on designing limited interactivity with constricted gameplay.

 

[Titanio]

Possibly old, but if anyone missed it IGN has some new videos of the tech demos on the showfloor:

http://ps3.ign.com/articles/698/698086p1.html

Also, a little more commentary on the behind-closed-doors Unreal demonstrations, with some interesting tidbits:

http://ps3.ign.com/articles/698/698213p1.html

Even more intriguing, he alluded to the possibility that users creating content on the PC side with the Unreal Engine might be able to bring their invented levels and mods over to the PS3.

That'd be sweet, if it happens.

 

[j^aws]

Physics is a broad subject. It can encompass projectiles/ particles, collision detection, fluid dynamics, radiosity, ray-tracing etc. etc. i.e. elements that describe the real "physical" world.

Now these elements basically boil down to solving large matrices on CPUs/ GPUs and architectures that encourage more "physics" is-a-good-thing (TM). Giving developers more options is-a-good-thing (TM) for gamers.

 

[Titanio]

IGN also appears to have been stealth-updating some of their game demo impressions. There are new Warhawk impressions up, with some detail on..clouds!

http://ps3.ign.com/articles/697/697737p1.html

Perhaps the two nicest looking elements were the natural elements, the clouds and water. The clouds were rendered via software and mixed with the hardware rendered elements from the rest of the scene, which is apparently a very new trick in videogames (movie effect houses have been doing this for years to great effect). The clouds featured volumetric shading and looked thick, not like some placed sprites that are there to fool the eye. The demo took the ship through the clouds over and over again, proving that the clouds contained depth and are rendered both inside and out. With the excellent shadow effects flowing over the sides of the clouds, they truly looked like 3D objects in the scenery, solid or not.

Ooh-err. Such a tease, not being able to see this for ourselves, though

There are newer impressions up for others too..at least they were new to me, not the same as their very first reports.

 

[wco81]

Will say one thing about clouds.

Rather than just be cosmetic, if they could be used for tactical advantage in a dogfight or flight combat game like Warhawk may be, that would be cool.

But Warhaws as I recalled on the PS2 ws shooting at stationary targets, not too many flying things. Or maybe I didn't get far enough in the game.

 

[Titanio]

I wonder if they're extending the technique for use for other things..like smoke or explosions and the like. Or would the often fast-paced dynamism of such effects introduce more challenges? If they've done a good job with this, you could have some really nicely rendered stuff in that regard, or not just in this game, but many others. Some of the nicest aspects of..certain demos..were the quality and very voluminous nature of things like smoke rendering etc.

 

[mrdarko]

Developed by Ninja Theory, the PS3 title was one of the most amazing in the E3 2005 presentation reel, with internet talk on the game hot and heavy after its showing. The demo then was not realtime, but this one was.

I thought apart from the close-up of the face,it was realtime?

Ign got there wires crossed?or is it me?

 

[DigitalSoul]

I believe the difference this time was that it was demonstrated running on actual hardware and not simply a video....

 

[Titanio]

I thought apart from the close-up of the face,it was realtime?

In-engine, "indoor" bits were realtime, but outdoor bits were not at the time. Everything was sped up to a nice consistent framerate for the E3 trailer.