Killzone PS3 shown at GDC 07?

Status
Not open for further replies.
Do i understand this correctly:

"I need to kind of couch this first. This is to show you the technology, not to show you the game. So please don't be thinking about the game too much, I just want you to be looking at the imagery and how we've constructed some of the scenes and how complex some of the scenes are."

They showed "EDGE" or something developed with "EDGE" using Killzone assests? And if they did, can we expect that Killzone is developed using "EDGE" ?
 
Do i understand this correctly:



They showed "EDGE" or something developed with "EDGE" using Killzone assests? And if they did, can we expect that Killzone is developed using "EDGE" ?

EDGE was jointly developed by many first party PS3 developers, its a possibility that GG contributed some of there game code into EDGE for other developers to use.
 
somebody should examine every frame of that HOME presentation and trailer and search for hints and secrets... :devilish: . Maybe we can find the date of second coming of Christ
 
Maybe you should clarify: Are you talking about the technical quality of the assets and render quality or the artistry?

Because the former is a technical issue and your opinion is irrelevant; the later is subjective and no one can tell you your opinion is wrong (although they can state why they disagree).

The Killzone CGI vehicles, honestly, far exceed the UT2007 vehicles on a technical basis. The tires alone have more polys than the entire UT2007 vehicle. Art is purely an issue of taste and preference.

Technically it's called brute force versus elegance. If a realtime solution is found to be visually indistinguishable from a non-realtime solution, for all practical purposes it is virtually technically as good... The non realtime solution is simply exponentially less efficient at providing for all practical visual purposes the desired visual effects, which is what matters in the end, it's more brute force and thus inelegant and something which should not be used in realtime applications(maybe precomputed or approximated somehow.), but an alternate efficient solution(there are always alternate solutions, even if they're a compromise) should be used instead.

In technical terms, the alternate solutions that are able to run in realtime with the hardware of the time are always the most efficient and appropriate ones to use.

If somehow the dev.s manage to use an extremely advanced LOD algorithm with next gen h/w, that is finding a way to make objects at certain distances nigh computationally free*(say sub psone lvs, pixel blobs or something.), they'd be able to do the stuff we saw in ps2 16k armies or more with a few shader effects only applied to those characters close enough for it to matter, and such shader effects scaling with the distance the camera is from the characters.(aka, high detail skin shaders are used or setup to appear only when a face is extremely likely to be close to the camera, the rest of the time simple textures and the least possible number of shaders are used in as many objects as possible without compromising visual fidelity.... an automated camera algo that adjusts camera distance in relationship to computational complexity to stabilize framerate not just to get the most pleasing visual angle might help here. ).

Nigh photoreal skin shaders have already been shown on nextgen consoles, so we know that cutscene wise at least photoreal looking characters are possible, having an army of destructible psone/ps2 lvl detail soldiers running in the distance during such cutscenes should be relatively trivial(even last gen consoles had no trouble running hundreds or thousands of moving destructible objects, and after a certain distance a ps2/psone object is indistinguishable from a ps3/360 lvl of detail object.). The question is if there's a way to somehow pull this off ingame, at least at 30fps and 720p.

July 11th 2007 ;-)

That's gonna be difficult for people like me that like helping to fulfill all prophecies(I'm a fan of the Lain anime, just some quirky behavior I took from there which I do from time to time.). I'll help in whatever way I can, which is basically doing nothing, but I don't think it'll accomplish anything at all.
 
Technically it's called brute force versus elegance. If a realtime solution is found to be visually indistinguishable from a non-realtime solution, for all practical purposes it is virtually technically as good...

The subtlety of the English language, but my emphasis in the very sentances you quote was on quality in regards to the image. It is pretty obvious to most of us that the PS3 will NOT be using offline rendering techniques as they take hours to render a single frame a GBs of data to store the shadows, texture assets, mesh assets, etc

It is granted they will use different techniques, so the question is will they attain the same technical quality. As you say visually indistinguishable results on screen.

So back to my specific example: Apply your arguement to the UT2007 media, specifically the vehicles, with the KZ media. A glaring difference can be immediately spotted in the tires where the KZ media has extremely high poly assets whereas the UT media's source media is equally high poly but the on screen image is composed of normal maps mimicking the source data yet show a significant visual differences and are by far technically inferior.

I am all far using render tricks to get the same results but with faster realtime methods. It doesn't matter if the mesh is completely loaded, or if you are tesselating HOS, or if you are using displacement maps to get the same result as long as it is the same visual quality. There are even cases where you can use parallax maps to make visually indistinguishable "poly data" (e.g. ATI has a nice demo where they put a hedge stone on the corner of building walls which prevents users from looking flush with the poly frame, preventing the illusion to be broken from the parallax map which indeed looked identical to the raw poly source the media was taken from... this is good use of art to prevent the illusion to be broken).

Obviously we all want to play games -- not slideshows -- so it is a given that the realtime techniques will win out over exponentially slower offline ones. We do, afterall, want to play games, not watch screens render.

But in regards to the point Ben was making (that UT2007's vehicles are as good as the KZ CGI vehicles) there is an absolute difference in the on screen visual quality.

And that was the point: Posters need to distinguish between the artistry and the visual quality when they say something is "as good, if not better, than CGI".
 
Technically it's called brute force versus elegance. If a realtime solution is found to be visually indistinguishable from a non-realtime solution, for all practical purposes it is virtually technically as good...

I don't really like what you're doing here - trying to mix objective analysis and subjective opinions together, going against Joshua's arguments.

VFX studios are still using more complex and computation intensive approaches because the difference is there. Noone would invest millions of dollars into renderfarm hardware and all kinds of software R&D if it wouldn't provide a significant advantage. And believe me, we avoid brute force as much as possible, because efficiency is key in offline work as well. Our deadline may be longer than the next few miliseconds, but it is there.

Differences can be analyzed and pointed out. Less aliasing is better, dynamic lighting is better, actual geometry instead of normal mapped impostors is better, and so on. Their significance may be subject to another discussion - but anyone's personal opinion about how realtime graphics and offline CG compare and wether the difference is important to him is irrelevant in a technical discussion.
 
I haven't been able to check myself first-hand, but apparently the next issue of OPM UK (released March 30), will have the world-first preview of Killzone 2 with 'first in-game shots and hard info'. If I'm passing a newsagent I'll try and pick up an issue to confirm that, but that's the word out on the interwebs at the moment.
 
having played the original killzone why is anyone excited about this ?

Probably because this is a graphics heavy forum, and Killzone was pretty decent in the graphics and sound department, I think. Now that we know more about it, it will also be a showcase title for Playstation Edge, apparently, and it will be interesting to see how things stack up against the Unreal Engines of this world.
 
Probably because this is a graphics heavy forum, and Killzone was pretty decent in the graphics and sound department, I think.
Yeah, right! You and I know the reason everyone's keen to see this game is because of the E3 '05 video! Everyone's waiting to either say 'Told you so! Sony lied and PS3 can't handle the CG trailer' or 'Told you so! They have managed to pull KZ off as well as the trailer.'

:D
 
I had always thought everyone was keen on it because it was supposed to be the Halo Killer....


I am sure the Teaser trailer didn't hurt... ;)
 
I don't really like what you're doing here - trying to mix objective analysis and subjective opinions together, going against Joshua's arguments.

VFX studios are still using more complex and computation intensive approaches because the difference is there. Noone would invest millions of dollars into renderfarm hardware and all kinds of software R&D if it wouldn't provide a significant advantage. And believe me, we avoid brute force as much as possible, because efficiency is key in offline work as well. Our deadline may be longer than the next few miliseconds, but it is there.

Not a fair comparison either.

If they wanted they could have rendered it much better than what was shown at E3'05. But they didn't because it shall be an equivalent of what's to expected for the real game (regardless if it can be achieved or not).
 
Killzone was pretty decent in the graphics and sound department, I think.

Are the enemy voices covered by the "sound department"? I mean... yeah...

That aside, it'd be nice if this info is indeed true, and we finally get a look at it. Hopefully some info on non-graphics stuff, being one among several that has played the original Killzone all the way through.
 
Yeah, right! You and I know the reason everyone's keen to see this game is because of the E3 '05 video! Everyone's waiting to either say 'Told you so! Sony lied and PS3 can't handle the CG trailer' or 'Told you so! They have managed to pull KZ off as well as the trailer.'

:D

Oh yeah, I suppose that too ... :D

Seriously though, when Killzone came out, it made a splash and everyone was stunned by what they managed to squeeze out of the PS2 back then.
 
I had always thought everyone was keen on it because it was supposed to be the Halo Killer....


I am sure the Teaser trailer didn't hurt... ;)

IMO there have been many Halo killers, Halo wasn't that great. I want to see a Half-Life 2 killer, I think Resistance came close.
 
I think Timesplitters2 and 3 deserved as much praise as Halo. It was a wonderfull but underrated FPS
 
Halo was successful through a combination of a decent FPS and great Live support, combined with being on a fixed platform (i.e. everyone has the same specs). In that sense, I think Resistance and Socom are the closest games on the Playstation platforms ... ? I certainly thought Halo 2 was a polished online experience, even if it didn't keep me from going back to GT4 within 2 weeks ... that has more to do with GT4, the Driving Force Pro, and my passion for racing games than anything else.
 
Status
Not open for further replies.
Back
Top