New Killzone screen grabs

[Guden Oden]

Sampling of the full PS2 library without bias to scene selection in a game - therefore a measurement of the performance typical to PS2

What kind of silly talk is that?

If you have a big building FILLED with games that do 1M polys/s, and then you have one that does 25M, what does that tell you about the hardware's capabilities?

Is 1M or 25M closer to the true performance?

Think about that for a while before you answer.

 

[Almasy]

Sampling of the full PS2 library without bias to scene selection in a game - therefore a measurement of the performance typical to PS2 - puts the highest performers near 7.5M and most titles in the range of 2-5M polygons/sec. Naturally, a title can peak higher at times under certain game conditions as is evident from the developer claims.

And I suppose that using your own rationale, a sampling of the full DC library without bias to scene selection in a game, would put DC´s numbers around 250k polys right?

Seriously, I know you like Sega more than sex, cars and food combined, but would you please at least try to stop the BS?

 

[Lazy8s]

Vince:

BS mechanic

PA mechanic

No PS2 game's sustained performance exceeds about 7.5M polygons/sec - a game is characterized by its typical performance, not its peak.

 

[Fox5]

Vince:

BS mechanic

PA mechanic

No PS2 game's sustained performance exceeds about 7.5M polygons/sec - a game is characterized by its typical performance, not its peak.

Yeah, and that's why they said rogue leader was 12M-15M polys.

 

[Almasy]

Vince:

BS mechanic

PA mechanic

No PS2 game's sustained performance exceeds about 7.5M polygons/sec - a game is characterized by its typical performance, not its peak.

So, if a game can keep a stable 60fps framerate while runring a very heavy scene in terms of geometry, we just dismiss that because it´s not the average poly count, right?

 

[Lazy8s]

Almasy:

So, if a game can keep a stable 60fps framerate while runring a very heavy scene in terms of geometry, we just dismiss that because it´s not the average poly count, right?

If it were sustainable, any scene could potentially use that kind of budget. If the majority of the game is made up of scenes like that, then it'd be typical performance.

By the way, average polygon count is a different concept, is not something that is easily measured (would be an exhausting and rather purposeless endeavor, and what you encounter in a game varies between playthroughs anyway), and was not being discussed here.

 

[Almasy]

Almasy:

If it were sustainable, any scene could potentially use that kind of budget. If the majority of the game is made up of scenes like that, then it'd be typical performance.

By the way, average polygon count is a different concept, is not something that is easily measured (would be an exhausting and rather purposeless endeavor, and what you encounter in a game varies between playthroughs anyway), and was not being discussed here.

But what if the scene does not need that kind of geometric detail? There are different kinds of environments you know, and some may want to portray a different atmosphere...not to mention that you cannot establish an average so easily because the number of objects on screen changes depending on the camera.

Anyway, as I said before, stop your Sega branded nonsense. Who cares if there´s a library of 100 average games that are in the 3M range when there is a number of games that reach much higher numbers? Are they dismissed just because the majority doesn´t reach it? If a machine manages to sustain at a good framerate a very high polygon count and is indicative of what a machine can do, you just cannot dismiss that because "it´s not the average".

What would that say about your beloved DC, anyway? Wouldn´t its average polygon rate would be 150k because it´s filled to the brim with enhanced PSOne games then?

 

[Lazy8s]

A system can't sustain its peak numbers typically because of the restrictions (stalls) and requirements of maintaining traffic flow. A lower, more applicable number arises that better describes the performance realistically: the typical throughput.

This has nothing to do with low-end games; nothing is being averaged together. The performance of the very best PS2 games are characterized with a geometry rate of about 7.5M polygons/sec rendered - this is the performance typical of them (not their peaks), that which best describes what they're doing.

Almasy:

But what if the scene does not need that kind of geometric detail? There are different kinds of environments you know, and some may want to portray a different atmosphere...

So if a scene of the same concept were produced for a PS3 game, they wouldn't bother raising the polygon rate above PS2 levels?

 

[marconelly!]

Why is this crazy polygon pushing discussion even going on? It's very obvious that the guy who wondered if KZ would be pushing some 12-15M numbers was thinking about peak numbers, not some average sustained performance sampled from random part of the game. Peak numbers, after all, is what developers always feed into their interviews, PR, and whatnot, and those are the numbers everyone has in mind when asking questions of that kind.

Having said that, no, I don't think the game will peaking anywhere near 12-15M, because it would have to have insane amounts of geometry to achieve that running at 30FPS. They are going for much more balanced poly/texture/effects look here.

 

[Lazy8s]

I sure hope that people are not looking for a misrepresentation when they're inquiring about the performance! I'd guess what they're looking for is the best description of the game's rates - not a peak, some number which is about as revealing as knowing that the trough is 200k polygons/sec or the like.

And developers are most definitely not all spin doctors. Thank goodness. For years, sustained figures instead have been given for games like Daytona USA, Virtua Fighter 3 and Soul Calibur, and real-world performance specs for hardware like Model 2/3 and Naomi2.

 

[marconelly!]

I sure hope that people are not looking for a misrepresentation when they're inquiring about the performance!

What most people are looking for with that question, is point of reference, and unless you can provide that, there's no need quoting some numbers with no game names attached to them. If someone knows that game XYZ has peak performance of ZYX (which is what developers are giving us sometimes), he might be curious to know how another game compares to something he has already seen.

Frankly, I care about developer's peak numbers probably even more than the sustained performance from that SCEE programmer, as sustained performance measured the way he did, tells me that on one level of said game, during the course of several seconds, that performance was sustained at number he gave us. That's all it's telling me, and considering how game levels can often be so completely different from one to another, or how much can they change from one moment to another (for several seconds you might be going through an empty corridor, then emerge into a bustling city) it's not telling me much at all. At best, it's as much of a misrepresentation as those peak numbers we are sometimes given. But, with those numbers, I at least know what the most impressive scene in the game might bring, not some random event I may or may not care about.

 

[Ty]

This has nothing to do with low-end games; nothing is being averaged together. The performance of the very best PS2 games are characterized with a geometry rate of about 7.5M polygons/sec rendered - this is the performance typical of them (not their peaks), that which best describes what they're doing.

Where are you getting your information from?

 

[Bowie]

Lazy8s,

Let's take Daytona as an example. It sustains less than 10 cars on screen, but it peaks at 40 cars on screen. So what you're saying is that we should just disregard the fact that 40 cars can be on screen.

What about Shenmue 2? It only sustains about 3 or 4 characters on screen, but peaks at over 20 characters. I guess we can disregard the 20 characters on screen because that's a misrepresentation.

 

[_phil_]

and what % of these polys are static ? skinned ? stripped ? textured with love ?

 

[whipper snapper]

What about Grand Prix Challenge?

- 22 cars on track
- 18,000 polygons per car
- 60fps
- 2km draw distance

 

[Phil]

cthellis42:

From what I recall, Guerilla has stated that Killzone is "stable 30fps"--or at least that's what's being aimed at.

Unless they changed it, it's most definately 60 fps they are aiming:

http://www.beyond3d.com/forum/viewtopic.php?t=6209&highlight=killzone[/url]]ere's everything we think we know on the title thus far... Kin is expected to release in Fall of 2004. It is the first PS2 project for Lost Boys, which was established in 2000. Inside sources who claim to have viewed the game in action have been quoted as saying that the game will in fact include online multiplayer co-op, in addition to various other 16 player modes with voice communication via the PS2 headset. Running at 60 FPS the entire time, it is said to be the best-looking PS2 game in existence which is odd, considering the creators of the recently unveiled Gran Turismo 4 (also a very impressive looking title) have already stated that GT4 maxes the PS2 out completely which is why you won't see voice chat in their game. Definitely some mixed comments happening there. Expect a control scheme similar to Halo, as well as the use of streaming techonology ala Jak & Daxter which will ensure that the game will feature no loading times whatsoever.

There was more steady info on the framerate than this, but can't seem to find it. When the game was first known as "kin" and referenced as 'halo-beater', framerate was touted as being one of the advantages compared to Halo. There are more sites that I found that reference to the 60 number:

http://boards.ign.com/message.asp?topic=53841213&replies=15&ui=cb_post_02

Of course, one can expect them to drop to 30 if they can't maintain it, but reading up on recent impressions of the game, I still have my hopes high they can make the 60 fps.


Lazy8s & Co.:

Please quit the polygon nonsense. This has been debated enough and at the end of the day, the games overal appeal is what makes a game, not its number of polygons. If you so greatly feel the need to pursue this argument, please start a new thread before this one gets locked. Thanks.

 

[VNZ]

All footage shown to this point suggest 30fps with occasional dips. I derive this from the fact that the vids are capped at 30fps but still show some slowdown. I'd be very surprised if they pull off the magic 60.

 

[PC-Engine]

I think initially the goal was 60 fps, however, seeing how they said the AI was slowing down the engine, it'll probalby be 30 fps.

 

[whipper snapper]

All footage shown to this point suggest 30fps with slowdown. I derive this from the fact that the vids are capped at 30fps but still show some slow-down. I'd be very surprised if they pull off the magic 60.

Aren't the videos encoded at 25fps. And I haven't noticed any slowdown in the videos. :? That being said the footage of the game on the big tv is alot smoother.

 

[Phil]

I suggest checking out the video posted on page one (the second one, it's a MOV file and is a capture off a german preview of the game). The footage shown being played in the background during the interview runs very smooth - even the other in-game captures (the same ones as in the first video) are definately more smooth which suggest that there are quite a few bad captures around.

PC-Engine has a valid point though. :? 30 fps would come as a huge letdown for me at least. Anyway, time will show I guess... hasn't the game been pushed back to 2005 - or is still on for release fall this year?


whipper snapper:

Aren't the videos encoded at 25fps. And I haven't noticed any slowdown in the videos. That being said the footage of the game on the big tv is alot smoother.

Heavily depends on how they captured the video, but if you film a footage off a tv with a video-cam, thanks to motion-blur, you can still tell if a game is more 60 or definately 30 fps despite the video running at 25 or 30fps. This effect can be seen in the Quicktime capture of the whole german preview in which one can see a scene being played on the big screen during the interview. That of course doesn't make it fact, but still leaves some room for hope at least.