marconelly!
Veteran
Again, those very same guys claim their PS2 racing game pushes up to 18M polys, so I don't know how exactly to interpret their numbers.
New Killzone screen grabs
- Thread starter whipper snapper
- Start date
"Engine" versus game - what is this referring to? All polygons, including those resent versus geometric complexity?
It was above 80,000 polygons per frame at peak with the older 60fps DC build of the game, before they instead chose 30fps for an increased frame budget to allow display for a whole field of cars.
It was above 80,000 polygons per frame at peak with the older 60fps DC build of the game, before they instead chose 30fps for an increased frame budget to allow display for a whole field of cars.
Check your math again, Fox5. They were able to use more polygons per frame when they weren't updating at the old rate anymore.
There should be some data on this stuff around here:
http://www.beyond3d.com/forum/viewtopic.php?t=9374&start=0
There should be some data on this stuff around here:
http://www.beyond3d.com/forum/viewtopic.php?t=9374&start=0
Ok, lets assume their engine could render 4.8mpps, which is more than possible on Dreamcast.
That doesn't change the fact that it's still a peak number of what the engine was capable of, just like the one they gave for their F1 game (18 mpps, multitextured), or the ones given by PD and others. We also know that, as measured by the PA2, PS2 version is pushing around 2.3 mpps, which if sustained would qualify their game as a low end PS2 title when it comes to geometric detail.
You can't both ask for people to use sustained performance when it comes to PS2 poly counts, and then use a peak number yourself to show them how close DC rendering capabilities were compared to Sony's system. If they did the same, they'd be using numbers as high as 22mpps, a peak number given by a developer in this very board.
We know that most DC games were pushing 500k-1mpps (sustained), and, based on that report, most PS2 games were pushing 2-5mpps last year.
Both systems have games that are above and below that average, and both systems have games that look good enough for me.
That doesn't change the fact that it's still a peak number of what the engine was capable of, just like the one they gave for their F1 game (18 mpps, multitextured), or the ones given by PD and others. We also know that, as measured by the PA2, PS2 version is pushing around 2.3 mpps, which if sustained would qualify their game as a low end PS2 title when it comes to geometric detail.
You can't both ask for people to use sustained performance when it comes to PS2 poly counts, and then use a peak number yourself to show them how close DC rendering capabilities were compared to Sony's system. If they did the same, they'd be using numbers as high as 22mpps, a peak number given by a developer in this very board.
We know that most DC games were pushing 500k-1mpps (sustained), and, based on that report, most PS2 games were pushing 2-5mpps last year.
Both systems have games that are above and below that average, and both systems have games that look good enough for me.
Lazy8s said:
Sorry I wasn't quite clear enough.
On the DC your figures tend to be engine figures - ie: All triangles passing through the rendering pipeline, even though many will be backface culled or clipped. The PS2 figures from the analyser only show the triangles that generate pixels ie those that make it through the entire pipeline.
This is one of the reasons why the PA results are really mainly of use to engine programmers, who can see the results in context, rather than just throwing numbers around.
As an example I have a GS bound engine with no bf clipping that pushes 20M polys/sec to the screen - I add BF clipping, so only 10M polys/sec are drawn, then change the engine to add a hair layer ( seperate poly - not texture ) so 20M polys/sec are drawn - I could claim that my engine processes 40M polys/sec before culling, as that is the number of inputs.
( In most cases this wouldn't occur - as calculation time might become the bottleneck.. but it's only an example
)Right... So.... Back to the issue here... Hint: Killzone thread... :?
Is it me or the interiors are quite similar to Red Faction? Mind you, i guess the settings are slightly similar, in some ways... I mean there's only so much u can do to depict a grey-ish building from the inside.... Nice to see they're focusing on nice lighting, which is unusual for this kind of games, until lately, and is the best way (IMO) to add atmosphere to a game...
Is it me or the interiors are quite similar to Red Faction? Mind you, i guess the settings are slightly similar, in some ways... I mean there's only so much u can do to depict a grey-ish building from the inside.... Nice to see they're focusing on nice lighting, which is unusual for this kind of games, until lately, and is the best way (IMO) to add atmosphere to a game...
Well, E3 is "soon"; only almost two full months away... Bleh! :?
Deepak said:
Maybe because KZ is not anywhere as near to completion as RE4 is... 2 pennies. Anyway, i've seen those RE4 and K7 videos, and they are migthy impressive. Seriously.
The one thing that the video almost shows ( and the covermount dvd movies here in the UK definately show ) is that the AI is ropey to say the least. Watch the trooper when you get to the top of the stairs - he starts looking around like he's trying to locate you. The AI seems to get very confused when you get within a certain distance of a character ( again, not so obvious in this video )
Lets hope they can fix it for the final release, and that Guerilla havent sucked up too much cpu / gpu time with the nice gfx and sound to allow decent AI....
Lets hope they can fix it for the final release, and that Guerilla havent sucked up too much cpu / gpu time with the nice gfx and sound to allow decent AI....
_phil_ said:yes , but we can seeit that way : it's allready on par with the AI in far cry in normal mode.One is released ,the other not....
Care to elaborate? Played the game and the enemies are pretty smart, considering it's a game and not an AI simulation on Blue Gene...
Farcry is a frustratingly difficult game, and although a lot is due to scripting, the AI is pretty decent to say the least.
But yeah, one is released, the other won't be for months, so no need for comparisons...
whipper snapper
Newcomer
http://www.killzoneps2.com/
Official website and don't forget to click the little music symbol top left.
Official website and don't forget to click the little music symbol top left.
MoodyB said:
I believe it's been said more than once these videos are from development builds where the AI isn't fully enabled. Jesus, some of the soldiers don't even react when shot, you'd think the game would actually have enemies in it that act like that?
Similar threads
