Sony @ GDC: Phil Harrison's Keynote

[mrdarko]

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.

Thanks for the explanation.:smile:

 

[expletive]

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

I dont think anyone said that.

 

[overclocked]

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?

Yep quite strange.

I said this before that its absurd that people are taking positive things and turn them into something negative or just negliate the whole *thing* to somehow make them feel better. Why even bother to post/read anything relative to Sony/PS3 if you not gonna participate in a constructive way?

And if i hear about the ps2/3 tech-demos or toystory gfx/killzone2 *trailer* anymore im gonna puke, so tired of hearing the constant nagging.

 

[function]

Physics: the new gigaflops.

 

[Mintmaster]

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 at a PS3 event, it is a marketing ploy.

All these new effect have everything to do with software and almost nothing to do with hardware, but Sony is trying to make you believe it's the other way around. The best physics (and A.I. too) I have seen in a racer is Viper Racing, which I used to play on my P166. Unfortunately people bitched about how "light" bumps damaged the car, how skidding was too easy, how turning was too slow, etc., and thus arcade-style NFS type games take off. Even with NFS, the first one had somewhat decent physics but they sucked thereafter. This comment about Motorstorm where deformed mud affects other racers is very ambiguous, and could simply mean that if a tire goes over the same spot it slows down. It's all about coding the support for it.

If I saw a completely interacting world, where everything was destructible, that would impress me a lot, but I haven't seen anything near that. Such an undertaking has almost nothing to do with FLOPS, either. The biggest challenges are keeping track of everything, and figuring out how efficiently and realistically affect the rendering.

Having said all this, it is good that developers are finally putting effort into upgrading the physics. Hope to see more of it.

 

[scooby_dooby]

What, exactly, is there to debate?

The debate is: Are improved physics really as important as we're being told by the hardware manufactures (who are trying to sell their PPU's), as well as Sony who is using it for a marketing tool as a distunguishing factor between consoles.

 

[Mmmkay]

The debate is: Are improved physics really as important as we're being told by the hardware manufactures (who are trying to sell their PPU's), as well as Sony who is using it for a marketing tool as a distunguishing factor between consoles.

So you're asking whether PR hype is going to be met? Does a genuine answer to this have any relevance to the importance of enhanced physics modelling? It sounds like you're trying to straw man, blaming PR hype. How about "Do accurate physics models increase game immersion and expand gameplay options?". What exactly are we looking for in a generational leap? More immersion and a greater variety of gameplay? Hmm.

But yeah, PR hype tends to cloud reality, I mean Microsoft's rhetoric about anti aliasing...

 

[scooby_dooby]

As far as I can tell phsyics is being positioned as the 'next big thing' by many different companies, I guess my question is...will it be?

I don't believe it's meaningful to discuss whether or not something will "increase game immersion and expand gameplay options" without discussing the degree of impact it will have. Sure it will have some impact...but how much? The biggest benefit I think we'll see is realistic hair and cloth, that's cool and all, but it's certainly not going to improve your gameplay experience all that much.

IMO, since we already have pretty cool physics in games, I don't see it giving any huge benefits or really doing much to increase immersion beyond what we've already seen.

 

[dukmahsik]

i wonder if next gen we'll see a ppu instead of super massively cored cpus

 

[Shifty Geezer]

I don't believe it's meaningful to discuss whether or not something will "increase game immersion and expand gameplay options" without discussing the degree of impact it will have. Sure it will have some impact...but how much? The biggest benefit I think we'll see is realistic hair and cloth, that's cool and all, but it's certainly not going to improve your gameplay experience all that much..

Start from Endorphoin/Euphoria and work you're way up from there, IMO. The potential for a change in gameplay is, as I've said, like the difference between Outrun and GT3. Moreso. That might not be realized this gen, but to stop working towards that because we can't perceive much difference in one generation strikes me as short sighted.

Talking of hair and cloth, that allows things like trying to catch creatures in nets, throwing ropes around giant's limbs and trying to tether them while getting dragged along, pulling the rug from under a zombie to trip it up, making a slingshot from some string and cloth, playing a Marble Madness type game with a flexible, deformable game world, and designing and building a kite to try and signal for your rescue. Off the top of my head. And without anyone needing to script when they're possible. And without gamers having to guess when a particular object has been enabled to be pushed or pulled or lifted or thrown but knowing that if it's in the game, they can use it. And with detailed character synthesis, without any artists needing to animate those things.

 

[BlueTsunami]

IMO, since we already have pretty cool physics in games, I don't see it giving any huge benefits or really doing much to increase immersion beyond what we've already seen.

I'm peronally all for lots of physics use in games. Also, heavy physics in consoles seems like its going to be realized this generation. With that said, I would give developers time to work with it and you WILL start seeing phsyics used in gameplay instead of visuals (one case being Motorstorm).

Also, you seem to listen to what companies say more than the actual relevance to physics being used in games instead of basing it off common sense. You should really be listening to what developers have been saying on this issue (instead of Sony, Microsoft or Nintendo). All in all in the end, physics should benefit gaming.

*for some reason, the Physics = Minimal Benefits or no Benefits at all mindset seems to be the same thing as Nintendo saying Online and High Def gaming is a fad. I see Physics and games being Online as the natural evolution of games. Physics in games shouldn't be getting the bad overcast people seem to be giving it

 

[Mmmkay]

And as others have already said, this is a technology forum and you're arguing against the progression of technology. It's akin to asking why bother moving on from normal mapping with things like displacement or occlusion maps when normal maps do the job just fine?

Physics have a long way to go yet just like non-simulated geometry in games. Most latter-gen games employed pre-canned scripted 'physics' elements tied with rudimentary physics models to provide their solutions. It's exactly the same for all aspects of game design, you fake until you can do it for real. Sometimes fake solutions can be adequate for the specific task at hand, but the overall model can become useless when applied in a different context.

The only way to address this is by providing the resources necessary to develop a full physics model. It may make very little difference to <specific title A> but its application over a variety of game genres improves greatly. It's a simple touch I know, but I would love to see a damn fighting game with proper collision detection and without clipping. Where in the game, your combat outfit affects the performance of your character due to its weight and rigidity. Soul Calibur 4, where if you equip a weapon that is unsuited to your character, you can lose balance when you swing it. Reintroducing damaged weapons in the Soul Calibur series too, where the type of outift your opponent wears can directly affect the degredation of your weapon along the impact zone.

This is just one example here of how physics can develop genres in new directions, and I think really we seem to be arguing that because we can't yet visualise how improved physics models could be applied in games, we should dismiss them as unnecessary.

 

[EPe9686518]

And as others have already said, this is a technology forum and you're arguing against the progression of technology. It's akin to asking why bother moving on from normal mapping with things like displacement or occlusion maps when normal maps do the job just fine?

Physics have a long way to go yet just like non-simulated geometry in games. Most latter-gen games employed pre-canned scripted 'physics' elements tied with rudimentary physics models to provide their solutions. It's exactly the same for all aspects of game design, you fake until you can do it for real. Sometimes fake solutions can be adequate for the specific task at hand, but the overall model can become useless when applied in a different context.

The only way to address this is by providing the resources necessary to develop a full physics model. It may make very little difference to <specific title A> but its application over a variety of game genres improves greatly. It's a simple touch I know, but I would love to see a damn fighting game with proper collision detection and without clipping. Where in the game, your combat outfit affects the performance of your character due to its weight and rigidity. Soul Calibur 4, where if you equip a weapon that is unsuited to your character, you can lose balance when you swing it. Reintroducing damaged weapons in the Soul Calibur series too, where the type of outift your opponent wears can directly affect the degredation of your weapon along the impact zone.

This is just one example here of how physics can develop genres in new directions, and I think really we seem to be arguing that because we can't yet visualise how improved physics models could be applied in games, we should dismiss them as unnecessary.

This is exactly why I don't think we need an extemely high level of physics realism in games. What you are describing here sounds way over complex for games. Games are supose to be fun, entertaining and accessable to most people. If you start pumping that type of detail into games I see them getting more and more complex to a level where only the hardcore gamers can play them. This is not where the industry needs to go, it needs to grow and start bringing in more casual gamers.

Physics like that would go over most casual gamers head and they wouldent really care about them that much IMO. Sure nerds like us would get all warm and fuzzy inside if we saw games with that level of physical realism, but do you honestly think thats going to make any differece to the casual gamer?

I think the state of physics right now is perfect, and I like the way it's scaling. With each new hardware release you can do a little bit more with physics than was possable before. I really don't think we need a physics cpu, and I really don't think developers should spend an insane amount of time making perfectly realistic physics for games. Now in some game aspects I can see this as importan, but I really think developers could better spend their time making better AI, more creative level design and other features that directly effect the gameplay in a much more noticeable way.


That is just my take on it.

 

[function]

The debate is: Are improved physics really as important as we're being told by the hardware manufactures (who are trying to sell their PPU's), as well as Sony who is using it for a marketing tool as a distunguishing factor between consoles.

I disagree tbh. I think the debate is just how advanced are the physics being talked about in these PS3 demos, and are they representative of something significantly beyond what we're currently seeing on the 360 or are likely to see in a similar timescale.

We don't even have any remotely meaningful comparisons between the two systems' CPUs for this kind of work, let alone any idea just how advanced the physics in any of these games is. What we do know is that Sony are pretty desperate to find some powerful yet intangible marketing angle for the PS3 technology, and this falls in neatly while being hard to pin them down on.

There are other questions too, such as at what point does an increase in physics power allow you to create an experience significantly better than on another platform, and how complex can you make interactions and gameplay before they become unmanageable from a game design POV or no longer provide an improved gameplay experience to the end user.

These are complex questions, particularly the second one, and developing answers isn't aided by seemingly misguided/mismatched comparisons to the benefits of improvements in rendering technology.

 

[Mmmkay]

This is exactly why I don't think we need an extemely high level of physics realism in games. What you are describing here sounds way over complex for games. Games are supose to be fun, entertaining and accessable to most people. If you start pumping that type of detail into games I see them getting more and more complex to a level where only the hardcore gamers can play them. This is not where the industry needs to go, it needs to grow and start bringing in more casual gamers.

Physics like that would go over most casual gamers head and they wouldent really care about them that much IMO. Sure nerds like us would get all warm and fuzzy inside if we saw games with that level of physical realism, but do you honestly think thats going to make any differece to the casual gamer?

It was an extreme example of a bunch of things I could think of that could be implemented in fighting games, not a design pitch. But thanks for using it to dismiss the progression of physics entirely.

 

[mckmas8808]

Physics: the new gigaflops.

What is that supposed to mean? Did you ever say anti analising the new gigaflops? Did you ever say HDR lighting the new gigaflops? I still don't understand why is it that when ATI and MS were yelling from the rooftops about "free" 4x AA you guys were cheering and clapping with them, but now all of a sudden we have to question heavy physics in games. Why?

The sad thing is games in the future will use heavy amounts of physics and you guys may not even change your mind about this by then. I guess BLACK is the best physics money should buy until the PS4 and Xbox 3 comes out right?

 

[mckmas8808]

I think the state of physics right now is perfect, and I like the way it's scaling. With each new hardware release you can do a little bit more with physics than was possable before. I really don't think we need a physics cpu, and I really don't think developers should spend an insane amount of time making perfectly realistic physics for games.

Okay fine but what about for the next-generation systems (i.e. PS3 and Xbox 360)? What do you think about games and physics on these two systems?

 

[Titanio]

We don't even have any remotely meaningful comparisons between the two systems' CPUs for this kind of work, let alone any idea just how advanced the physics in any of these games is. What we do know is that Sony are pretty desperate to find some powerful yet intangible marketing angle for the PS3 technology, and this falls in neatly while being hard to pin them down on.

Cell was designed with physics & simulation as one target workload in mind. We even have seen implicit and explicit comparisons of the systems from this POV from AGEIA, if you really want something more concrete. General simulation would appear to be a genuine strength for PS3, not because they simply wish it to be so.

These are complex questions, particularly the second one, and developing answers isn't aided by seemingly misguided/mismatched comparisons to the benefits of improvements in rendering technology.

When people dismiss the value of simulation for the advance of the visual, it is an eminently relevant comparison to make. Simulation and rendering indeed will only become more tightly bound going forward.

As an area, physics offers potential for multiple aspects of a game, perhaps more than any other technical area in a game. You can use it purely to visual ends, which is fairly low-hanging fruit and should be relatively accessible, or you can use it to heighten immersion or realism or to impact directly on your game mechanic. All of these things contribute positively to the experience. Having more power broadens the pallette for the developer, which is always a good thing.

The apparently selective 'ludditism' on display here is quite astonishing. By all means argue against the value of greater simulation capability going forward, but don't dare draw the line at physics if you are to do so.

 

[Tap In]

[FONT=&quot] [/FONT]

I disagree tbh. I think the debate is just how advanced are the physics being talked about in these PS3 demos, and are they representative of something significantly beyond what we're currently seeing on the 360 or are likely to see in a similar timescale.

We don't even have any remotely meaningful comparisons between the two systems' CPUs for this kind of work, let alone any idea just how advanced the physics in any of these games is. What we do know is that Sony are pretty desperate to find some powerful yet intangible marketing angle for the PS3 technology, and this falls in neatly while being hard to pin them down on....

I agree.

I think this is the debate and to presume that this demonstration proves that this type of experience is only possible because of Cell is premature.

Until further proof I'm of the opinion that this is not something out of the realm for either of the Next Gen systems and is a software decision (just as Full Auto was a game decision to create fully destructible environments).[FONT=&quot][/FONT]

 

[scooby_dooby]

And as others have already said, this is a technology forum and you're arguing against the progression of technology.

I don't see how I'm arguing against the progression of technology at all. What I'm speaking about is the progression of games as a medium, and what is most important in taking the medium to the next level in realism and immersion.

When I ask myself what will be the thing that takes gaming to that next level, and really create a great movie-like experience the answer is not improved physics, it's improved storylines, impoved AI, improved cinematics, and more realistic large, non linear worlds, and much improved GFX so as to not destroy the illusion of a real worl. I think these are all the key areas that gaming needs to progress that have a MUCH greater impact on ruining the immersiveness than having 'inaccurate physics.' Physics are a part of the eye candy, but they are a very small portion.

I'm not arguing against better technology, just questioning how impact this will actually have on making games better.