Tech Analysis: Crysis 2 Demo
Crytek analysis!
Can consoles run Crysis? It's a question Digital Foundry has considered on a number of occasions in the past, based on tech demo footage Crytek released showing its state-of-the-art CryEngine 3 running on Xbox 360 and PlayStation 3. The recent Xbox Live Crysis 2 multiplayer demo allows us to revisit this topic once more, this time with the benefit of being hands-on with live code.
Right now, it's safe to say that our findings are far from complete. The one level players get to sample is a multiplayer specific stage and doesn't seem to bear much relation to the campaign action we've seen of the game thus far. Crytek has also gone on the record to say that the demo is based on a build of Crysis 2 that is some weeks old, suggesting that a range of smaller changes will be made to the final game.
So while the sampler doesn't give us the kind of expansive overview of the tech that we'd like, it's still well worth further analysis and should give us some idea of progress since Crytek released its last tech demo just after GDC almost a year ago.
One of the key concerns on previous clips showing Crysis 1 or Crysis 2 assets within the CE3 framework has been frame-rate - as you might expect from tech where the underpinnings come from a top-end PC engine. On console, we've seen v-sync engaged and disengaged in various videos across a range of technology demos and not surprisingly, we've also noticed a large variance in the performance level. Reports on the recent Xbox 360 closed beta also complained about the fluidity of the gameplay, noting that the game operated with an unlocked frame-rate, adding further to a sense of fluctuating performance.
"We're looking at 30 [frames per second]," Crytek CEO Cevat Yerli said to VG247. "There's nothing below, nothing... I'm very happy the game works at all. I think the amount of push we have and the fact the game works like that puts me in a very happy situation right now."
While the old tech demos from the time Yerli made his comments clearly don't live up to his claims, this latest code gets very, very close indeed. The overall consistency of the gameplay in the demo is excellent, which is especially impressive when bearing in mind the advanced CE3 feature-set. The game is effectively v-synced, operating in a similar manner to Halo: Reach. Any torn frames are limited to the very top of the screen and are therefore virtually unnoticeable, although it's interesting to note that across the run of play Crysis 2 actually seems to be smoother than Bungie's last game based on what we've seen so far.
A montage of clips of the Crysis 2 demo, revealing a very solid performance level - something we expect to translate across to the full game.
There are other similarities with Reach, however. Native resolution on Xbox 360 is confirmed at 1152x720 - presumably chosen to make more effective use of the super-fast 10MB of eDRAM directly connected to the Xenos GPU. Again, just like Reach, high dynamic range (HDR) rendering is supported, so the chances are that Crytek opted for this framebuffer set-up for much the same reasons that Bungie explained in the recent Digital Foundry tech interview.
The parallels with Bungie's latest epic continue into other areas of image quality: Crytek has also employed a very similar anti-aliasing technique. Temporal AA is the process of merging data from the last frame into the current one, usually with a half-pixel offset on every other frame. The idea is that all the rendering time spent on the previous calculations isn't binned off in favour of fresh processing for the new frame. Since much of the new image will be very similar, it can be combined with the previous one in order to reduce edge issues and other forms of aliasing.
We discussed Crytek's innovative approach a while back, since it offered to bring something new to the table. Temporal AA has an issue whereby ghosting creeps into the image, especially on objects moving very close to the camera, or indeed any large, moving object. A CryEngine 3 tech demo showed an intriguing approach in using temporal AA for far-off detail, with an edge detect and blur on close-up objects.
The demo itself showed some excellent results:
In this technology demonstration we get to see Crytek's 'hybrid' technique for dealing with edge-smootihng - a combination of temporal anti-aliasing and edge detect/blur.
The implementation in the Crysis 2 multiplayer demo on Xbox 360 doesn't look quite so strong, however. Were it not for the edge AA being enabled in the demo's config file, it would be difficult to acknowledge it was actually in use at all, and it is clear that the temporal AA can incur some very obvious ghosting artifacts at a surprisingly close range.
Unlike Halo: Reach, we don't tend to see any kind of ghosting on the view weapon, but there are still issues with environmental detail - probably the most noticeable example is with the wind turbines in the Skyline demo, which stand in contrast in front of the skybox, emphasising the ghosting.
Mitigating the effect in places is Crytek's excellent implementation of object and camera-based motion blur. This may well be a 30FPS shooter, but many people have noted that the overall feeling of the gameplay is similar to Call of Duty. While all the COD titles top out at 60 frames per second, Crysis 2 runs at half the frame-rate but thanks to some swift response from the controller combined with the motion blur effect, the feeling is of a game that stands out from the 30FPS pack.
The temporal anti-aliasing in the Crysis 2 demo does a great job of edge-smoothing in static scenes but moving objects can produce noticeable ghosting artifacts.
That it should look so impressive shouldn't really be that surprising. The original Crysis on PC sapped resources on even the most powerful PCs at its loftier settings, but the implementation of the motion blur (especially in DX10 mode) helped in producing a look significantly smoother than the raw frame-rates would suggest.
You may also like...
-
Gravity Rush Review 48
-
Sony patents method to interrupt your gaming with an ad 107
-
Wii U Aliens: Colonial Marines is best-looking version because of console's "more modern tech" 74
-
Kingdoms of Amalur: Reckoning needed to sell 3 million to break even 65
-
Activision vs. Vince Zampella and Jason West: Inside the game industry trial of the decade 70
-
Skyrim gets mounted combat in new update 54
-
Arma 3 in-engine footage shows off lighting tech 18
-
App of the Day: Go Robo! 2
-
Ghost Recon: Future Soldier Review 132
-
Minecraft overtakes Black Ops on XBL activity chart 25
-
Dirt Showdown Review 89
-
Minecraft total sales hit 9.2 million 8
-
FIFA has a new world champion 8
-
How the Darksiders 2 delay benefits you 9
-
Anarchy Reigns delayed in the West, Platinum says 10


Comments (142) Latest comment 1 year ago
Comments for this article are now closed, but please feel free to continue chatting on the forum!
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
IIRC a Crytek developer mentioned that the demo was a relatively old build (1-2 months). Considering Crytek will probably be trying to do as much polishing as possible before release, it'll be interesting to see how the final game shapes up, especially in SP
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
No PC demo yet - although one has been promised. Interviews have stated that the performance should be better than the first Crysis. If you could run that game at High settings, you'll be able to do the same here with a nice little performance boost.
Here's hoping that's not PR BS as I'm scraping by at 1920x1080 High settings and would kill for an extra 10/20% performance boost in the sequel.
Comment below viewing threshold Show
Comment below viewing threshold Show
The controller feedback is anything but swift.
A major complaint of this demo on many websites, features and feedback have been the loose controls. The game doesn't feel precise and crucial controller issues such as dead-zone, acceleration and translation are not as tight as those found in titles developed specifically for console. I like the "heavy" feel of it and it's not near Killzone 2's level of "lag" (which I also got used to after a while) but I don't understand why you call this swift?
Am I missing something? Difference between wireless and wired?
I too heard they plan to address the lag and if they can get it near Reach's, *Anything by Valve*, Alan Wake or Uncharted's level of precision and performance, count me in! I did try various controller settings (usually need to "up" them a bit).
Having said that, I did have a lot of fun with the demo/beta. It is another engine that makes 6 year old hardware shine. I am happy I can spend my money buying games instead of having to upgrade hardware. And for PC gaming this focus on *also* consoles means that on top of best possible experience, it will also run great on the PC.
Hearing about a PC demo; if it performs better (more responsive, which is likely) the feedback could be that the console port of Crysis is broken even if it was just this build that that underperformed. Wonder how they fix that little marketing/pr nightmare if true.
edit: typos
Comment below viewing threshold Show
Comment below viewing threshold Show
A little bit sad really, in that for both my consoles 360 and PS3 I thought this would bring super performance to 'average' developers who chose cryengine....
Maybe I was being too optimistic expecting to blow Creed and UC2 out of the water, and maybe I was a litt;le too critical of UE III.
Need to play the demo, I read the EG Forum notes and many say lots of pop up, lag and floaty feel for movement, usually readers give good feedback so my hopes are 'more realistic' now...
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
I'm a bit sceptical about the PS3 version though, considering how old the GPU is. I don't think any amount of multi-threading through the SPUs is going to overcome that.
Comment below viewing threshold Show
It must read: Can Consoles run Crysis 0.5! Yes, but they even can run Gears of War3, Mirrors Edge, GTA4
So what IMPRESSIVE visible new thing brings C2 on the table? I don't see anything.
That isnt clearly enough for the self-proclaimed-tech-leader.
Comment below viewing threshold Show
Does it? Are you basing this off one map from the MP demo, or...?
Comment below viewing threshold Show
Comment below viewing threshold Show
We saw wonderfull demos which suggested it would blast everything out of the water, and I thought it would.
Not saying teh engine is poor, its just when you perceive it would be MUCH better than everything else, when in fact it looks to be on a par with other leading engines.
Neg all you want, all I am saying is I expected or hoped for more....?
Abd yes, roll back 12-18 months, the tech demos and claims of Crytek would be that it would surpass the likes of Kill Zone.
On the plus side, if a generic multiplatform engine can get close and make devs work easier and outperforms UEIII, then thats good.
Comment below viewing threshold Show
Comment below viewing threshold Show
then again i remember RL with carefully wording that a older engine (ME2) outperformed a new engine (ME3) just to have even the minor ups of ps3 (irrelavant to me really) in reasonable doubt even though the comparison had little interest since Xbox owners had the luck of having the game way before and it was more in line for ps3 to keep up gamers for upcoming entries.
i m waiting for an expertly worded ps3 analysis.
seeking for negs obviously
breathe in , breathe out ...there
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
- high render distance
- high resolution (textures, shadows, rendering resolution)
- realistic and complex physics ( almost every flower, bush, tree or house reacts if there's movement or an explosion in Crysis 1, while in the Crysis 2 demo almost nothing reacts like this)
- Detailed effects ( for example, when you shoot a wall, some debree will fall off. PC's will be able to process more of these things).
The game will probably be superb on a console, but for the best graphics you need a PC. If you don't want to spend much money for beautiful graphics, just don't.
Comment below viewing threshold Show
Also, lol at the PR talk about the 3D mode. Well, he didn't lie, framerate is mostly still 30 fps, he just "forgot" that there's a ridiculous amount of tearing. Magic indeed.
Comment below viewing threshold Show
Comment below viewing threshold Show
If the 360/PS3 versions live up to what we saw in the latest gameplay trailer (see these shots), then I'd say that it's pretty close to Killzone 3. It's pretty subjective, though.
@carlosfdn: Wow, such hypocrisy.
Comment below viewing threshold Show
Also still shots should be material for some executives to be accused of and be crucified in public for misinformation
Comment below viewing threshold Show
So sub HD then. Surprised Crytek couldn't manage an HD game on consoles.
Comment below viewing threshold Show
Done so they could use the EDRAM of the 360 to get free HDR
Comment below viewing threshold Show
From what I see concerning Crysis 2 console it runs exceptionally well, it looks great and it's coming to a console near you in a few weeks time. What more do you want?
Comment below viewing threshold Show
05/02/11 @ 11:53
ignore poster | #27
0
It's a demo. On a rooftop. You've basically only seen a stripped down version of what the engine is capagle of thus far. Crytek have kept the majority of Crysis 2's gameplay under wraps. You'll be holding your arse when the actual campaign falls in your lap I practically guarantee it.
Sure the campaign will look better, but so does the campaign in kz3 too. two very different art styles in these games imo, and whilst i think crysis 2 looks very good, it just doesnt sit well with me at all, it literally made my eyes hurt looking at screen(maybe it was the lag making me feel disorientated). Kz3 has a more cartoon style and looks beautiful in its won way too.
I wasnt really that excited about campaign for either game, and with crysis 2 it seemed like they had some fresh ideas, but the demo felt like a total let down, the controls were woeful and lagged horribly. Kz3 rann solidly.
based on the demos, id not think twice about choosing kz3 over crysis 2.
Comment below viewing threshold Show
The first one (and Far Cry) had really cool locations which lent themselves beautifully to an open ended approach to the campaign, but I just can't get excited about this new one.
Maybe in a couple years we'll get a nice Bioshock-style game with this engine to see what it can really do...
Comment below viewing threshold Show
Comment below viewing threshold Show
Rage deleted game, and by time id calmed down enough to have another go, the game beta had finished
Comment below viewing threshold Show
Killzone 3:
[link url=http://killzone.dl.playstation.net/killzone/kz3_screenshots/kz3_ss_2011-02-02_retropack_19.jpg
]http://killzone.dl.playstation.net/killz...[/link]
[link url=http://killzone.dl.playstation.net/killzone/kz3_screenshots/kz3_ss_2011-02-02_retropack_16.jpg
]http://killzone.dl.playstation.net/killz...[/link]
[link url=http://images.eurogamer.net/assets/articles//a/1/3/1/0/7/9/4/KZ3_preview_12.jpg.jpg
]http://images.eurogamer.net/assets/artic...[/link]
Crysis 2;
[link url=http://www.mycrysis.com/sites/default/files/gallery/image/statue.png
]http://www.mycrysis.com/sites/default/fi...[/link]
[link url=http://www.mycrysis.com/gallery/wallpapers/seeking_cover_grenade_blast
]http://www.mycrysis.com/gallery/wallpape...[/link]
[link url=http://www.mycrysis.com/sites/default/files/gallery/image/tracker_perk_0.jpg
]http://www.mycrysis.com/sites/default/fi...[/link]
Images say it all.
Comment below viewing threshold Show
doesnt say much really, as most of those are cut scene shots. I want to see the in game engines before i call which looks best, and even then it boils down to artistic choice and is totally subjective.
have u played the demo of crysis 2? doesnt look anything like that, secondly, they are probably shots from the pc, as most people have confirmed that these shots are running at settings on medium on pc, which the 360 cant replicate. (not huge differences mind you , but still higher quality) Based on mp demos alone, id choose kz3 in looks dept, i found crysis 2 very shiny, but also very bland. Kz3 had a more cartoon feel to it, but was smoother, more artistically appealing.
Kz3 campaign is stunning looking too, and im fairly certain the campaign in crysis 2 will be technically amazing too, but i wouldnt expect what u seeing in these shots or videos im afraid on ur 360. Further more i hope its more imaginative than mp artwork
[link url=http://imagequalitymatters.blogspot.com/2011/01/update-schedule-and-crysis-2.html
]http://imagequalitymatters.blogspot.com/...[/link]
[link url=http://imagequalitymatters.blogspot.com/2011/02/tech-analysis-crysis-2-multiplayer-demo.html
]http://imagequalitymatters.blogspot.com/...[/link]
Comment below viewing threshold Show
Comment below viewing threshold Show
Probably not the best images to be showing. Pitting press shots against actual in-game shots isn't a great idea. These would probably be more accurate:
http://dl.dropbox.com/u/12527604/crysis/pictures/BeStrongTrailer/Shot1.jpg
http://dl.dropbox.com/u/12527604/crysis/pictures/BeStrongTrailer/Shot2.jpg
http://dl.dropbox.com/u/12527604/crysis/pictures/BeStrongTrailer/Shot3.jpg
http://dl.dropbox.com/u/12527604/crysis/pictures/BeStrongTrailer/Shot5.jpg
http://dl.dropbox.com/u/12527604/crysis/pictures/BeStrongTrailer/Shot9.jpg
http://dl.dropbox.com/u/12527604/crysis/pictures/BeStrongTrailer/Shot10.jpg
Both games look pretty great regardless. Besides, a slight difference in graphics doesn't hold a feather to the quality of the game itself
Comment below viewing threshold Show
Comment below viewing threshold Show
[link url=http://imagequalitymatters.blogspot.com/...
]http://imagequalitymatters.blogspot.com/...[/link]
read these articles along with eg one and get better idea about mp and campiagn trailers etc
Comment below viewing threshold Show
05/02/11
are those confirmed 360 shots??
they look more like 360 shots, even the cut scene shots seem lower res and less graphical effects laden . Still looks nice tbf, but not quite as stunning as trailers and the majority of screen shots we see, which are apparently pc shots in reality
Also so far im not convinced on Crysis 2 gameplay, but if they manage to tighten up controls and get rid of lag, i could really like it, as there was enough about it to see that underneath was some good ideas etc, but it just played horribly online when i played it. lag was worst ive ever seen in a game , and having previously owned black ops on ps3, thats pretty frickin bad lol
Comment below viewing threshold Show
What did people expect from the engine on consoles? There are just limits to hardware.
I personally don't care that much for the prettiest engine, I prefer good artwork instead of best lightning and highest polygon count.
And isn't it great to have more engines than UE3, so a lot of games don't look so damn familiar in a certain way?
Comment below viewing threshold Show
They're captures from the 'Be Strong' trailer. No one really knows what platform it's from, but previous Crysis 2 trailers have generally held up to the console versions, so we'll see.
Honestly though, it almost seems too good to be at 360 settings, though one preview did say it was the best looking 360 game they'd ever seen.
*shrugs*
Comment below viewing threshold Show
I thought all the given screenshots were x360, but I guess I'm wrong (quality of the pictures is way too high). It that's really medium, I can't wait to see very high
I am absolutely sure that the game isn't rendered at 1152x720 or whatever it was, it must be stretched or there's something wrong with my xbox or TV. The pop-ups are horrible and the quality is low. It's still the best I've seen on a console, and those shots from the trailer look very good IMO.
But Crysis 2 is still more beautiful then Killzone 3. You can play it on a PC, and with maxed settings it will definitely look better than Killzone 3. You'll probably say that it isn't fair to compare a PC to a PS3 (and you are kind of right, it isn't fair), but we're not comparing the PS3 with the X360, but we're comparing Crysis 2 with Killzone 3, and Crysis 2 looks better, period.
Comment below viewing threshold Show
comparing console,Killzone 3 blows it away.
Comment below viewing threshold Show
I would also like to remind every one, Crytek said that Crysis 2 on console will raise the graphical bar over anything we have seen so far, SUB-HD!! surely they were talking last Gen. lol
I hope that Crytek can deliver a true HD Crysis 2 experience on the consoles & I really hope they manage to set the new graphical bar for this Gen of consoles, just to back up the hype if anything.
Thanks DF
Comment below viewing threshold Show
You should at least compare two cars of similar specs in a race, similarly, comparing crysis 2 on 360 and kz3 on ferrari is a much fairer test. If kz3 ran on a pc then fair enough, you could compare the two engines, but otherwise it just is a laughable comparison.
The reason the resolution looks worse is explained in that article too, as there are various graphical effects being added , which are probably applied to hide rough edges or lower resolution textures.
I personally dont see whats so special about the demo apart from the lighting, which is fantastic, otherwise it all seems kind of blah and soul less .i cant honestly say i think its the best looking console game, even if technically its great(not artistically from demo). kz3 seems to have more soul to it, its own identity, its own style, crysis just seems to be like a tech showpiece, without the power to show it off properly. if theyd put more creativity in terms of art style, they could have done so much more
Comment below viewing threshold Show
Hmm, OK. I wasn't trying to convince you that it was better than KZ3 or anything, but since early screenshots and videos are enough to convince you that one game destroys another (even before the games have been released, mind you), I think I'll just leave it here.
Just wondering, though, how is this 'just unimpressive'? I'm genuinely curious, since even if you look at it from an objective point of view, it's still pumping out far more than most other games do.
Comment below viewing threshold Show
Comment below viewing threshold Show
I'm still not convinced those pics are from consoles though , it all looks to crisp and clear, the only thing that suggests it maybe on console is the foliage on trees which looks poor , but that maybe just a low-med setting on a pc .
Comment below viewing threshold Show
However, it doesn't seem that ps3 version is disaster, and what crytek said in reality( at least later on down the line) was that each system runs various parts of engine better.
"Development of the cross platform tech wasn?t easy though, as Atkinson says. ?It?s been tricky, but we?ve got a strong PS3 engine with all the major systems running on SPUs.* With middleware, what you want is for someone else to do all that so the developer can just concentrate on making games.* We have parity between the platforms now: both run at the same speed.?
Although this ?same speed? clearly depends on what the engine is doing at that particular time.* ?If the game?s shader-heavy it runs a bit faster on 360; if it?s compute-heavy with physics and particles, then the SPUs take over and it?s a bit quicker on PS3.?"
According to techies the crysis 2 demo contains info indicting that ps3 version will be fine .
"Perhaps all that’s left, is to wax lyrical about the possibilities surrounding the currently MIA PS3 version of the game. However, if you’re looking for talk of the system’s bandwidth limitations potentially resulting yet further downgrades, loss of IQ, then you’ll be somewhat disappointed. If a recent, internal config file of the game is to be believed - the info contained within specifically puts the two versions on a par with each other, with the implementation of v-sync in relation to performance separating them – then this shouldn’t be the case."
[link url=http://imagequalitymatters.blogspot.com/2011/02/tech-analysis-crysis-2-multiplayer-demo.html
]http://imagequalitymatters.blogspot.com/...[/link]
Comment below viewing threshold Show
Msaa has it's own issues as well, and standard as is still preferable if resources available, but unrealistic on this gen of consoles.
I prefer kz3 , but saying it looks much better is just fanboy-ism
Comment below viewing threshold Show
Comment below viewing threshold Show
KZ3 is specifically optimized for the PS3 a hardware that is when properly used arguably a bit more powerful than the xbox360.
So it's okay to compare an engine that is multiplatform with an engine that runs only on PS3 but only if you don't compare it with a PC because that's just unfair because it's more powerful hardware, right?
And that's just besides the point that resolution and AA is by far no absolute measure for the power of an engine.
If people prefer KZ3 visuals it's fine but please don't try to argue it's objectively clear, because it isn't. Especially not after one multiplayer level.
Comment below viewing threshold Show
Killzone 3: 1280x720
Crysis 2: temporal AA with ghoting artifacts sometimes
Killzone 3: Morphological antialiasing
trying to say these things will tell us which game will be better is completely retarded thinking.
Even from a pure graphical standpoint factors such as draw distance, quantity of objects, effects and particles on screen are as / more important than pure resolution and the flavour of AA that is used.
As to the quality of the game it does appear that Crytek know that the open approach to achieving your goals was the main draw to Crysis and they're doing their best to include it in Crysis 2. Will be interesting to see how they manage with limits of consoles to achieve this.
Comment below viewing threshold Show
Comment below viewing threshold Show
But why would Crytek say "The PS3 version runs slightly better" if it was a disaster?"
Why would Bioware say that the PS3 version of ME2 is the definitive version? It's PR BS.
Don't you think that if they had something amazing in their hands they would show it to increase hype? Especially considering the fact that they'll be facing some huge competition on that platform?
Comment below viewing threshold Show
Still interested in the campaign though, their take on post apocalyptic new york could be good.
Comment below viewing threshold Show
Comment below viewing threshold Show
"Development of the cross platform tech wasn?t easy though, as Atkinson says. ?It?s been tricky, but we?ve got a strong PS3 engine with all the major systems running on SPUs.* With middleware, what you want is for someone else to do all that so the developer can just concentrate on making games.* We have parity between the platforms now: both run at the same speed.? "
Secondly whilst I also pointed out that they are unlikely to say the ps3 version is substandard, the evidence seems to suggest otherwise.
"Perhaps all that’s left, is to wax lyrical about the possibilities surrounding the currently MIA PS3 version of the game. However, if you’re looking for talk of the system’s bandwidth limitations potentially resulting yet further downgrades, loss of IQ, then you’ll be somewhat disappointed. If a recent, internal config file of the game is to be believed - the info contained within specifically puts the two versions on a par with each other, with the implementation of v-sync in relation to performance separating them – then this shouldn’t be the case."
[link url=http://imagequalitymatters.blogspot.com/...
]http://imagequalitymatters.blogspot.com/...[/link]
Do ppl bother to read posts or just spout their own views regardless if what's been said rather than in context to a conversation ???
Comment below viewing threshold Show
Comment below viewing threshold Show
Comment below viewing threshold Show
I did not say Crysis 2 is superior to KZ3, you must have forgotten*, it was you who claimed superiority of KZ3 "in consoles".
So why should i prove something that i did not claim. That would be mad, would it not?
*No wonder, it's convenient to forget such things. You should definitely think about some political career, they look for people of your caliber all the time.
The thing is i don't watch politicians debating and i also don't like to participate.
Comment below viewing threshold Show
Comment below viewing threshold Show
i think its better to say nothing ..and wait until final release...dont forget that Crysis 2 is a huge game and build by CRYTEC ..! so the final game will be awesome..im sure
Comment below viewing threshold Show
Comment below viewing threshold Show
"just a tiny mention of the terrible pop in and nothing on input lag. "
RL's probably saving those to emphasize on the PS3 version
Comment below viewing threshold Show
[link url=http://www.youtube.com/watch?v=0SDBiUjbZ4M
]http://www.youtube.com/watch?v=0SDBiUjbZ4M
[/link]
Wish they'd released that as the main playable map, it looks wicked.
Comment below viewing threshold Show
by the looks of it, it does.
Comment below viewing threshold Show
]http://www.lensoftruth.com/?p=27725
[/link]
Anslysis of kz3 vs crysis 2 betas
Comment below viewing threshold Show
Are you sure of that? I need proof.
Imo Crysis 2 is the first game but i am not sure.
Comment below viewing threshold Show
KZ2 and KZ3 doesnt use realtime GI with indirect lighting like CE3. Neither does KZ2 nor KZ3 have HDR, they uses LDR (Low Dynamic Range). Just browse Guerilla Games own presentations and publicaly released documents about their rendering tech. People forget games needs to be fit to console hardware and there are tradeoffs. Same for PC but the bar is higher. KZ2/3 also makes tradeoffs in particle buffer resolution why effects look very lowres due to 640x360 res buffer. On the other side it seems KZ2/3 uses selective AF filtering on some surfaces and rest trilienar or bilinear and Crysis 2 on console only trilinear.
Then you can see Crysis 2 has proper SSAO while KZ3 has simple foot contact shadows, C2 has soft vegetation with SSS effect which KZ3 lacks, C2 has proper specular shading and more advanced material shading, 3D deformable water with FFT and caustics, 100% dynamic lighting and shadowing which KZ2/KZ3 has limited amount of. On the other side KZ2/KZ3 has more particles onscreen, more stylised look, higher mp player count, less visible pop-ups. List can go on.
Shots from latest C2 trailer.
[link url=http://www.abload.de/img/c2bsttups3.jpg
]http://www.abload.de/img/c2bsttups3.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt12zsa6.jpg
]http://www.abload.de/img/c2bstt12zsa6.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt15dtfv.jpg
]http://www.abload.de/img/c2bstt15dtfv.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt4lr7j.jpg
]http://www.abload.de/img/c2bstt4lr7j.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt9vstr.jpg
]http://www.abload.de/img/c2bstt9vstr.jpg
[/link]
[link url=http://h-4.abload.de/img/c2bstt5cr7s.jpg
]http://h-4.abload.de/img/c2bstt5cr7s.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt10iti4.jpg
]http://www.abload.de/img/c2bstt10iti4.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt5cr7s.jpg
]http://www.abload.de/img/c2bstt5cr7s.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt11hqbo.jpg
]http://www.abload.de/img/c2bstt11hqbo.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt13uqox.jpg
]http://www.abload.de/img/c2bstt13uqox.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt142pxi.jpg
]http://www.abload.de/img/c2bstt142pxi.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt35tu6.jpg
]http://www.abload.de/img/c2bstt35tu6.jpg
[/link]
[link url=http://www.abload.de/img/c2bstt6kryu.jpg
]http://www.abload.de/img/c2bstt6kryu.jpg
[/link]
http://www.abload.de/img/c2bstt77rsp.jpg
Comment below viewing threshold Show
You may be right, I cant find anything for definite either way, but no articles i can find mention global illumination, i think kz3 just employs an advanced way of mimicking this effect, but its not all from one source in reality
Comment below viewing threshold Show
@Nebula
THx guys for explanation
Multiplayer in c2 is done by Free radical and we all know what they done....Haze
Single player will look much much better.
THis Global illumination stuff looks and sounds great
Similar to many nvidia demos like Luna or this program that shows HDR "rthdribl" etc?
Comment below viewing threshold Show
I know i was massively disappointed , going off the demo cryengine 3 on consoles is behind the likes of dice, COD, and many unreal 3 games in the visual department.
All the fancy tech terms on DF and the so -called achievements by crytec dont make any difference, crysis2 looks quite average.
Comment below viewing threshold Show
, I just think their artstyle was bland and felt soul less, whereas kz3 had it's own distinct look and feel
Comment below viewing threshold Show
I think without the crysis/crytec name this demo would have been laughed at ,if it was being claimed as some new peak in console engine development. Killzone 2/3 is far better and many 3rd party and 1st are superior on 360. I dont even like killzone much or most of Sony's supposedly technically , xbox360, superior exclusives but their far better looking than crytecs efforts.
Seems without unlimited hardware power to aim at, CRYTEC are not the wizards many , including themselves claim.
Comment below viewing threshold Show
Comment below viewing threshold Show
BTW, why did this thread turn into a KZ3 fanboy wankfest? KZ3 is a bit irrelevant here.
Comment below viewing threshold Show
You remember Gears Of War 2? that took a visual hit in MP vs the SP, there are a few games that have done the MP graphics downgrade on console, I just hope that it is only the MP that got a downgrade & not the SP as well.
Also remember, this is Crytek's first outing on console & whilst they have made many bold claims about their game on console, we all have to remember that this is just Demo & probably an old one at that.
Comment below viewing threshold Show
Was mainly logical conversation apart from 1 or 2 individuals, so wasnt totally irrelevant
Comment below viewing threshold Show
This siggraph article about Crytek's global illumination technique was published 5 months ago (from what I found) and from reading the article(the powerpoint document) and looking at some very minor and infrequent lighting bleeding errors in Killzone 3 open beta (similar to those described on page/slide 23) , it does look like they've used the same technique.
If I'm understanding page/slide 39 correctly, they seem to be suggesting that the global illumination algorithm is “Extremely fast (takes ~1 ms/frame on PlayStation 3)” at the 1280x720 resolution listed on page/slide 37; so I'm wondering if that means the GI on Crysis 2 on PS3 will be updated every frame rather than every 5 frames implied for all systems, especially given how small 1ms of processing is from a 33ms frame budget for games running at 30fps.
Comment below viewing threshold Show
I have been playing the Killzone 3 beta and I absolutely love it. I am getting the game based on the multiplayer alone, and I know I will probably really enjoy the campaign as well. Needless to say the graphics are beautiful and it runs quite well.
But Crysis 2 also looks very good and has its own strengths.
Comment below viewing threshold Show
Go with what you fancy, both will deliver.
Comment below viewing threshold Show
All these excuses about being crytec first time on console, im sorry but these devs are held up as being the best tech wise in the industry and have made many claims that their new engine blows everything out of the water and that they are "maxing out both consoles".
If crysis2 turns out to be a visual dud, which imo its looks like it may well be, then crytec should be held to account.
Why would they release the worst looking demo map if it wasn't indicative of the campaigns visuals, thier on record as saying its a month or two old, hardly going to be big changes in the time left.
Comment below viewing threshold Show
Comment below viewing threshold Show
Guess it's cause that until recently it was only exclusives that really gave a good representation of the console , as it was cheaper and easier to knock out bad ports on ps3 from 360 than produce good ports, but then the ps3 actually started selling well, and publishers figured that they better start making half decent ports at least, but most these days have optimised their engines and software to produce almost identical /identical ports. Anyone with half a brain Knows by now that each machine has strengths and weaknesses, and when utilised correctly offer similar experiences, with pros and cons for each version.
I find fan boys on either side obnoxious tbh , these are either ignorant about technical issues altogether, or just stupid and choose to ignore facts about issues. 360 fan boys are equally retarded as ps3 ones.
Comment below viewing threshold Show
Comment below viewing threshold Show
Because PS3 fanboys are special-needs, small-bus types who need constant validation and reassurance about their platform?
Comment below viewing threshold Show
What a pointless , deliberately aggravating comment , that has no other purpose than to piss off ps3 users.
Ps3 users have had years of substandard multiplatform titles , so it makes perfect sense for devs to reassure ps3 users about the performance of the game on the console , as ps3 users are naturally wary about non exclusives , and devs want ps3 to be excited to drum up sales. Finally , I'd say that the reality is that at that time it is likely that the ps3 version was performing better, but it wasn't very specific about what stage they were at and what aspect they were working on , they later stated that the engine runs better on ps3 in certain regards , but better on 360 in regards to different aspects, which is pretty unsurprising with each console having their own specific strengths and weaknesses
Comment below viewing threshold Show
but wasn't too impressed but mainly because I am not too keen on the art direction and style
the game itself felt very solid and controls where great... so technically good but game-wise probably not my thing
Will wait to see what reviews this gets
Comment below viewing threshold Show
Not saying you don't make some valid points, but reading some of the self-proclaimed "non-fanboy" posts on the Dead Space 2 face-off thread makes for a compelling argument in womble's favour.
Comment below viewing threshold Show
I'm sure there's complete and utter ps3 fan boy tits, I came across a few on ps3 forum within black ops site, who even when you explained to them , and backed up with evidence these points about why the ps3 version was inferior , due to the the engine not being optimised for ps3 and being gpu heavy, thus resulting in need to drop resolution and frame rates suffering, they still insisted that it was down to other non sensical reasons .
However, I'm sure that there's a lot of 360 fanboys too, who are equally retarded . I apologise if womble is not one of these , but the statement was deliberately inciting a fan boy war, and thought that was bit childish.
I own both consoles but try to remain balanced in my opinions on them, and don't feel need to put down other console owners down, and feel the best approach is to be rational with fan boys on either side, rather than taunting them .
Maybe I'm hoping for too much lol
Comment below viewing threshold Show
Put it in some perspective:
Face-off after Face-off, we see the same routine: a denigration of Richard by PS3 fanboys, who accusing him of bias, selling himself to Microsoft, yada yada. All over the technical merits of a game.
Of course there are 360 fanboys, and god knows there are PC zealots, but the PS3 seems to attract MORE than its fair share of disgruntled types. For some reason, the console is a magnet for malcontents and right bastards.
The local contingent can't come up with convincing counter-arguments regarding the technical aspects -- it is a technical blog after all -- so instead they attack Richard's credibility. Creeps.
I've never been attacked by a 360 fanboy for saying that Uncharted 2 is GOTY. Or that Dead Space 2 on the PS3 is the one to get. But mention that RDR is a better purchase on the 360 and suddenly the fanboys come out of the woodwork. And god help you if you mention Halo or think that Killzone 1/2/3 isn't the best thing to happen since choccy donuts...
PS3: good platform, terrible fans.
Comment below viewing threshold Show
Whilst I have always said to accuse rl of bias is ridiculous and that usually it's pretty obvious that multiplatform games ( especially running on ue3 are inferior, but improved drastically lately ) , this is often due to no optimisation of engines etc, although the ps3 clearly has some bandwidth issues,although this can be worked around with clever devs, but I have to say that there's no doubt been instances where rl doesn't highlight certain issues on 360 games in face offs or highlights issues in ps3 game in one face off but brushes them aside on 360 in another , and there have also been instances where a 360 loss equals poor port and ps3 loss equals bad hardware. Those are facts, but I always said I'll look at other sites too and make my own decisions based upon what I see overall:
I can understand the backlash in some ways, kz3 is from demo a very good shooter and in no way any less deserving of criticism than say blops, which escapes criticism for certain faults that are highlighted in kz3, ie poor story, lack of innovation blah blah etc. This to me is more indicative of the nature of industry and media though where certain titles especially ones where ms and a lot of money at stake , they seem to escape criticism .call me a cynic or whatever, but that's way I'm calling it. Halo games are fantastically polished and like any game may not be to everyones tastes, same with kz I guess. I can't honestly say that it's story is particularly great and it's mp isn't exactly revolutionary anymore.
Red dead, oh dear, I personally didn't like the game, but the ps3 version was an atrocity, I actually bought the ps3 version and wasn't even bothered by the graphics, I just got bored of game, and should have known I would really having disliked gta4, but great reviews swayed me. The ps3 version suffered at hands of an old engine designed around gpu that wasn't optimised and bandwidth issues again, there's a good chance that even if it had been optimised that ps3 version would still in some ways look inferior due to high use of alpha transparencies etc, but immense some clever design would have offered a reasonable alternative.
I own both consoles and accept that both consoles have strength and weaknesses. but overall the 360 is better designed for devs to exploit it's power easily, whilst ps3 is awkward. As crytek said about cryengine 3 both run different aspects of engine better at times depending whats going on at time in it.
People need to just accept that there's little in the machines themselves to seperAte them, but that Sony kinda made life difficult and ps3 owners still paying for it, although I think eventually there will be parity more or less as new engines are optimised to avoid weaknesses of each platform . The success of ps3 over last few Years was in reality a nightmare for devs and publishers, as they could no longer do ports as afterthought, it required spending money and figuring out how to get best out machine, as it was no longer a small market, but a major player in game. ( activision clearly didn't care still with black ops though)
Ps3 fan boys always claim that ps3 more powerful etc, and in some ways this is true, but in other ways it's hardware doesn't allow that to be taken advantage of, and the 360 is brilliantly designed to produce games, and more importantly at time, there was software that did all work for them available( ms master stroke IMO), whilst ps3 required them to invest heavily in time and money to make use of what is essentially a very clever piece if tech . Sony were extremely arrogant in their design though, and should have approached devs to see what they wanted, or thrown in some extra ram to allow them to get used to the approach to programming required. I actually think we will see this type if set up in future consoles too , as the ideas behind it are brilliant, but it will be in more dev friendly set up.
Overall both powerful in own ways and capable of very similar results, just require different approaches.
Fan boys on both sides need shooting , but there's times where they can also raise interesting points about conspiracy etc, as although I'm not really a conspiracy fan, immnot naieve either, and like any aspect of world I'm certain that politics, money and power play a part in it.
Comment below viewing threshold Show
Untrue.
I specifically said "PS3 fanboys".
I'm a PS3 owner, I have plenty of mates with PS3s, and they don't need to carry on like arseholes.
For some reason, the PS3 attracts zealots in disproportionate numbers. They can't enjoy a game unless they're enjoying it more than someone else.
Comment below viewing threshold Show
Actually Scientists/electrical engineers (OpenCl/DirectCompute) programmers will all tell you how much more powerful the PS3 hardware is than the 360.
All too often, people innocently regurgitate the same mantra “that they are roughly equal”. It just shows that it is cheaper to convince semi intelligent customers with propaganda & rhetoric than it is to produce amazing hardware. It only took 200 PS3 Cells (~Ł0.5M of computing power) to create a fake SSL(secure sockets layer) certificate for peat's sake, which should give you some insight that they are certainly not equal.
Name any recent AAA game(last 2years) on the 360 that you definitely know uses more than a fast/less accurate 24bit depth buffer with 24bit colour(with x8 HDR halo:reach), and then try and find any recent game on PS3 that uses less than a 32bit depth buffer with less than x128 HDR.
One or two bytes per pixel might seem like nothing; but it is about 4MB extra data (in a double buffer @720p), and even then many 360 games get a further saving by using less than the full 720p resolution to work inside the 10MB Edram.
If people are supposedly going to do proper analysis of game SKUs, then they need to have all the facts and report on them in a way that could stand up under scientific scrutiny. Something which DF articles haven't for years, and just pay lip service to the larger install base in the US.
Comment below viewing threshold Show
Oh god. It's 2006 all over again.
We'd better make sure the Iranians don't get the Sony super computers!
For a "much more powerful" machine, PS3 games look pretty damned similar to 360 games...
Comment below viewing threshold Show
And so do certain Wii games, Sega Sonic All Stars Racing, Tiger woods, etc look similar on all. But most DF thread readers will still acknowledge they are only similar and not equal, and would never call the Wii version as being technically superior.
Comment below viewing threshold Show
Comment below viewing threshold Show
There's very little separating the PS3 from the 360 overall.
Your problem is that you want there to be a difference, for some kind of ideological reason. But that's YOUR problem.
For actual GAMERS, the biggest differences are in terms of PSN vs XBL, Trophies vs Cheevos, and the DS3 vs 360 controller.
Comment below viewing threshold Show
Yes I know it has all this theoretical power, and also huge amounts of real power too, but it's gpu and diminished bandwidth seem to limit what it can do regarding games with even the most amazing looking ps3 exclusives like uc2/3 , kz3 , infamous 2 etc having to use clever methods or lower res effects to achieve certain effects like water, fire, smoke, explosions etc .
This power is limited like I said by aspects of it's hardware design, it's effectively a super computer, but for making games the 360 seems more balanced overall. If ps3 had more ram or that 10mb edram 360 has I think we would be looking at two very different results, but we are not. The ps3 has amazing ability to deal with physics etc , but struggles due to bandwidth with alpha transparencies etc.
Thus we end up with very similar results from both . Please feel free to add to this in tech talk though, as i respect ur opinion.
Comment below viewing threshold Show
Comment below viewing threshold Show
Is this down to the talent at ND or the ps3s supposed power?, as its the only series to be visual superior id have to say its more down to ND.
Likes alreadys been said , the ps3 is literally crippled in far to many areas to go pass or sometimes even match the 360 in graphics, the gpu, ram, blue ray drive, and awkward design, means it will never happen.
Comment below viewing threshold Show
Bluray drive isnt really a huge hindrance at all, and can be worked around as others have shown ,and can offer benefits too, hence devs keep hinting at the need for it on future consoles, and the gpu itself itself isnt bad, in some ways is better than that of 360, its just that the 360 gpu has that 10mb edram and unified shaders, plus the unified ram as well. The gpu isnt really meant to be used to process effects in same way on ps3 as on 360, thats what the spu's are there for, they can relieve the gpu of these jobs. The architecture itself, whilst different, once understood, and the software optimised, is extremely versatile. The lack of bandwidth available seems to be biggest issue on ps3.
As Mark atkinson at crytek said about Cryengine 3
"Although this ‘same speed’ clearly depends on what the engine is doing at that particular time. “It’s been tricky, but we’ve got a strong PS3 engine with all the major systems running on SPUs. With middleware, what you want is for someone else to do all that so the developer can just concentrate on making games. We have parity between the platforms now: both run at the same speed.” Although this ‘same speed’ clearly depends on what the engine is doing at that particular time. “If the game’s shader-heavy it runs a bit faster on 360; if it’s compute-heavy with physics and particles, then the SPUs take over and it’s a bit quicker on PS3.”
Furthermore, God of war 3 was stunning title yet to really be matched on 360 as a spectacle,and Infamous 2 looks phenomena l(Nd are working on it with the devs, so no big surprise). To say that Nd are only company capable of getting best out ps3 is not correct, it's simply that ND have worked very closely with Sony and learnt a lot about the ps3 doing so, they also now work with other companies to explain how to exploit ps3 too on behalf of Sony.
Again I would say, each has it's strengths, but the 360's biggest strengths are it's edram and bandwidth, and also the fact that most devs were familiar with its architecture and able to simply throw stuff at gpu to do, whilst the ps3 with the technically more clever design 9that extra ram really would have made a difference to the ps3 though and put it out ahead of 360 comfortably, but heck it was expensive enough already), required a new way of thinking and programming that wasnt predominant at the time, although it's now accepted as a better way of programming, and has led to a new approach to designing pc's, with multiple cell processors within multiple cell processors seen as the future.(i know cell itself if dead, but the idea has been refined and continued , heck the basic prenise itself formed basis for 360)
Comment below viewing threshold Show
That is not true; the Cell part is actually the SPUs cores and highspeed interconnects, the PowerPc cores that it shares with the Xenon are really just feeding the SPUs and could have been replaced by another BigEndian symmetrical 2-way CPU core like an intel Xeon.
All this talk about comparative bandwidth limits in PS3 are very odd with regards to true HD(720p and above); The Crysis 2 Demo engine works below HD according to this article, to get a bandwidth boost from EDRAM without tiling, while already having a native API boost from DirectX FX anyway, but it still fails to deliver the true HD resolution(similar to other sub-HD 360 titles DF has downplayed, like Reach).
When it comes to feeding symmetrical(Xenon) or asymmetric CPU cores(Cell) with data, the Cell has a bandwidth advantage; even without utilising deterministic real-time compression/decompression (only really suited to steam cores) because of the faster RAM speeds and zero caching inside the SPUs.
In late 2008 IBM completed work on the Cell B.E OpenCL 1.0 driver, which essentially allows programmers to utilise all of the 6xSPU cores above the 50% CPU efficiency limits of general purpose CPU cores(eg Xenon, Core i3,i5,i7, Xeon) running OpenCL/DirectCompute.
First party PS3 titles are adopting this new efficient (shader like)software development model much quicker. Discounting DF's article skewing, multi-platform developers have also been changing. But this year will be very interesting as Bungie will no doubt be using OpenCL on PS3 to make similar looking games to 360 versions.
Comment below viewing threshold Show
Please explain why other sites ( iqgamer reference the bioshock 2 http://imagequalitymatters.blogspot.com/... ) " The reason behind this seems to stem from the fact that the PS3 game is streaming textures directly from the Hard Drive, in which there is a 5GB mandatory install, whereas the 360 is having to load them in directly from DVD. Essentially the PS3 has greater available bandwidth to do this via the HDD compared to 360’s DVD drive, which allows it to push through more higher quality textures at faster speeds, though not necessarily displaying more texture detail, as this is still limited by the system’s internal RAM.
Earlier we mentioned that there was a noticeable difference on how each version renders its transparency and alpha effects. Basically on PS3 all effects are rendered in as little as a quarter of HD resolution, whilst they are of full resolution on the 360. As we have pointed out before in our Dante’s Inferno comparison, this is done on the PS3 to save bandwidth as there is much less available than on Microsoft’s console. The PS3 only has around 21.6GBs per-second worth of bandwidth available for framebuffer effects compared to a huge 250GBs that the 360 can draw upon. This means that in order to render all the same visual details they have to be displayed at a lower resolution in order to fit into the bandwidth requirements of the PS3.", and devs have spoke about such huge bandwidth limitations , and why uc3 , kz3 still use low res textures and sprite based techniques as opposed o traditional methods for explosions and water etc, and more importantly how what you described above Would circumvent these issues.
Yeah I knew I'd explain the xbox/ ps3 similarities wrong , I couldn't find site at time to red and explain from and I couldn't remember exactly what their shared history was, hence it's vagueness lol
Always interesting to hear ur opinions , even if I dont fully understand them all time
Comment below viewing threshold Show
Give it a rest already.
This "sub-HD" stuff is utter rubbish. You can't HONESTLY tell the difference between Halo Reach running at 1152x720p vs 1280x720p.
You could just as easily point out that many elements of Killzone 2 are running "sub-HD" at any given moment in the game. Or that 1280x720x30 is maintained because of tightly constrained environments and excellent art and production techniques.
You are cherry-picking your examples, ignoring the fact that there is not a SINGLE game available on the PS3 that does what Halo Reach does. Not one. Name me a single game that comes close to what Forge World does: truly massive INTERACTIVE environments, at a stable v-synched frame-rate, complete with sandbox physics and AI, and scads of enemies. The proxy rendering technology is a wonder in itself.
There is nothing as stellar as Uncharted 2 on the 360. And there's nothing as stellar as Halo Reach on the PS3. Each game is specifically tailored for the limitations and merits of the hardware.
Frankly, if you can't see the technical merits in games like Halo Reach, minor resolution differences aside, then there's simply something wrong with how you're looking at games. Try looking at these things from the position of a GAMER, instead of a platform warrior or fanboy.
Comment below viewing threshold Show
Ur right though about kz3, it is stunning looking, but the explosions and smoke are clearly sub hd and it does detract from the experience slightly. Dunno about arguing about UC2/3 and Halo reach though. Just because the ps3 doesnt have anything like it, doesnt necessarily conclude that it couldnt achieve it, whilst , it seems to be the general consensus that uc2 does do things 360 couldnt, albeit , its clearly tailored to ps3 strengths and avoids its weaknesses. However, there is almost certainly, based upon software upto date and current titles and even including near future titles like uc3 in this, effects that would have to be edited in line with ps3 architecture/limitations etc, if Halo reach was ported to ps3, based on things that ive seen in halo reach and things that tends to be altered or mimicked/edited out in ps3 ports and exclusives.
I think you and Vizzini could both learn a lot from each other if you would both stop bickering and discuss things in sensible fashion, and maybe you could both do with dropping slightly fan boy esque attitudes(yeah i know u own ps3 too, but it does come across that you rate 360 higher)
Now all this talk about uc2 makes me want to play it againn lol.
ps kz3 demo, wow its pretty. I think its artistically up there with UC2, although Uc2 noses it. Lighting still not as impressive as in crysis 2, but damn impressive nonetheless, and particle effects are stunning too.
Comment below viewing threshold Show
Split/Second(1280x672) only stole 48lines (48/720) or 6-7% and putting both games side by side you can clearly see a difference.
The different viewport ratio affects the rendering view frustum configuration, which either needs artistically checked/adjusted(for lighting attenuation, fog, geometry/shadow mapping depth bias) even forgetting a likely difference in z-buffer resolution to start with.
The loss of 128lines (128/1280) or 10% less width in Reach (letterboxing/stretching) will be more noticeable on bigger screens(46” & above) in this genre;
As a HD 16:9 TV(720 or 1080):
has a [inv]tan( 9/16) ~ 29.36deg angle across the diagonal
46” * cos(29.36) = 40” screen width
The 40” width that equates to 100% native resolution, which means the image (and visual fidelity) is either smaller (40*0.9) 36” or stretched by those 4” inches horizontally.
I'm not saying that this diminishes Crytek's or Bungie's game development in any way(as their job is making games), but just that in a sub-discussion about which HD hardware is computationally superior for games, screen resolution is a deal breaker as it taxes the system hardest.
The plethora of (technically speaking) sub-HD games on 360 since Bungie were allowed to bypass Microsoft's own TRCs and produce a Halo 3 @ 640p can't be discounted so trivially.
And I don't believe there is such a thing as sub-HD FX, just ones that look amazing(Modnation Racers' level 3 hydra missile attack) in real-time, and ones that don't. FX were never meant to be viewed statically by the gamer.
Comment below viewing threshold Show
Even exclusives like kz3 and uc3 use alternative solutions to alpha transparency due to bandwidth issues. Uc3 looks stunning , but the fire effects look poor when comPared to other titles on 360. The explosions and smoke in kz3 look too soft and don't fit with rest of the visuals, due to being lower res / particle based.
Water effects in uc3 seem more organic due to the solution they used but still , it's an issue they had to resolve using what is essentially an inferior effect.
Bioshock 2 suffers greatly due to the implementation of water effects on ps3 too.
Whilst I agree the stretching of picture is a technical mark against halo reach, you have to wonder whether ps3 would deal with it any better , or whether the use of alpha transparency amongst other things would result in it running worse, or having to be altered so that lower res effects etc would be used to ensure it ran even at that sub hd resolution.
I'd appreciate your responses, especially to my previous questions in earlier post on bandwidth etc
Comment below viewing threshold Show
All newer AMD GPUs have moved further towards superscalar design since OpenCL/OpenGL3.0/DX10 to copy the way SPUs hide latency, but at the cost of circuitry intelligence to reduce redundant processing that was their go-to move.
A V-synched, double/triple buffered, Deep Colour(proper HDR lighting), 32bit zbuffer at a full 720p 30fps/60fps isn't possible on the 360 without serious tiling(if it is at all possible in more than fighting games like VF5).
Tiling is an enemy of memory latency hiding, which is why even with a 5:1 bandwidth advantage for alpha blending(which is a form of redundant overdraw processing), very few 360 games match the PS3/PC for framebuffer setup. The size of alpha blending FX on PS3 I would assume are the result of being done on SPUs without tiling(with less than 256KB for data).
Kirankara: but so does the low res effects and textures that seem to dominate in ps3 games.
Your confusing the use of full exponent frustum fog(at 32bit & deep colour HDR) that reduces harshness/sharpness, as a loss in texture detail, in comparison to crushed blacks.
I took the first image from the bioshock 2 link, visual inspected for crushed blacks, full frustum exponent fog, and then used a paint package to process A-B(gamma increase 4.0), B-A(gi 4.0), difference of A & B(gi 4.0). Which certainly suggests the opposite scenario for texture detail, as all the high frequency detail is in the PS3 image, and all the low band frequency differences were from the 360 image.
Comment below viewing threshold Show
iqgamer state clearly that these effects are lower resolutions, and theres no doubting smoke, flames and water always look inferior on ps3 games. Why is that?
"Earlier we mentioned that there was a noticeable difference on how each version renders its transparency and alpha effects. Basically on PS3 all effects are rendered in as little as a quarter of HD resolution, whilst they are of full resolution on the 360. As we have pointed out before in our Dante’s Inferno comparison, this is done on the PS3 to save bandwidth as there is much less available than on Microsoft’s console. The PS3 only has around 21.6GBs per-second worth of bandwidth available for framebuffer effects compared to a huge 250GBs that the 360 can draw upon. This means that in order to render all the same visual details they have to be displayed at a lower resolution in order to fit into the bandwidth requirements of the PS3.
The effects of this can be seen above. Notice how the water running down the stairs is much blurrier than the surrounding stairwell and the stairs themselves. The same thing can be seen with almost all water, fire and particle effects in the game. It does mean that although textures are almost the same in both versions, the lower resolution effects tend to blur out those very same textures on the PS3. Basically the high res bump mapping and texture detail is effectively being displayed at a lower resolution and upscaled every time a transparency or alpha-based effect is rendered on top of them. With this happening frequently - as Rapture is an underwater city, leaking and slowing decaying with age - you find that the entire scene has a tendency to blur when all these visual effects are present, thus negating any advantage the PS3 version might have had with its use of better filtering and superior LOD system.
These lower res effects also feature less animation than those of the 360 game, with most of the water effects being affected, along with some rather strange errors when it came to rendering certain flame effects, and seemingly random objects in Rapture’s various rooms. Some pixallation occurs when viewing these at various angles and at long distances, and although this isn’t as apparent up close, you can still see that something doesn’t look quite right. In addition it seems that there is less, or more subtle use of bump mapping on the PS3 when compared to the 360. Sometimes it appears that the levels used are the same, at other times it seems like the PS3 is lacking in that department. Perhaps the reduced resolution alpha effects are to blame, as in areas in which there is very few of them, the bump mapping appears to be much better and can reach parity with the 360.
However there are many times when the use of lower res buffers hardly impacts upon image quality at all, looking nigh on indistinguishable from the 360 version. From what I’ve observed, this mainly applies to pools of water located on the floor in small dark corridors, or areas with low light levels. In these cases texture detail, bump mapping and IQ of the effects looked only slightly worse, and sometimes pretty much identical, showing that you don’t always need the technical advantage to produce similar results. Unfortunately this is the exception rather than the rule when talking about Bioshock 2."
whilst Eg states
"The big visual differentiator between the two games comes down to the handling of transparent "alpha" textures. These eat up bandwidth and fill-rate on the consoles, and as regular Digital Foundry readers will know, the 10MB dedicated RAM attached directly to the Xbox 360's Xenos GPU can give the Microsoft console a very real advantage here.
A very common solution on PS3 is to reduce the resolution of these textures: Killzone 2 for example scales them up from a quarter-resolution buffer, but adds multi-sampling anti-aliasing to smooth off the edges. For effects that are on-screen for a split second (for example, explosions) it's very hard for the human eye to notice much difference: it's a massive bandwidth-saver, with little impact on overall image quality.
BioShock 2 employs the same trick with its transparencies (without the MSAA). A massive amount of the game's alpha textures are rendered with a quarter-resolution buffer, which is fine, except for one problem: these aren't on-screen for a split second, they are there a lot of the time. All of the water, particles and fire effects in BioShock 2 are rendered in this way, meaning that depending on the scene, some or even all of the screen is being generated at quarter-HD resolutions.
Even some of the neon decals have a quarter-res effect on them that stays constant no matter how far you are away from them - the upshot being that, weirdly, the further away you move from the texture, the more obvious it becomes. To some extent or another, the sub-HD elements are with you for much of the game. After all, Rapture is an undersea city springing a hell of a lot of leaks: water is everywhere. It's as much a part of the BioShock 2's signature look as the art deco architectural style. Even transparent items such as EVE hypos exhibit the effect."
So my question again is, why are we not seeing the fruits of these ps3 advantages in reality with the addition of Cell B.E OpenCL 1.0 driver etc? I can accept that the bandwidth explanation , as it fits with other stuff ive read on this, and yet we still see lack of alpha transparency used in ps3 games, and fire, water, smoke etc, all looking poor, even the most talented of first party devs such as guerilla and naughty dog have been unable to match the 360 in this dept.
http://www.lensoftruth.com/?p=27376 fight night champion demo head to head, again we see less sharp textures and alpha transparency reduced on ps3 version
fight night champion
Comment below viewing threshold Show
Comment below viewing threshold Show
I think we've strayed too far and too long away from talking about Crytek's game engine and the graphics and performance we expect on each system.
So I will quickly just say the the Fight night demo is looking like Unreal 3.5 with multi-pass lighting implemented on PS3 cut-sequences like comparing ME2 again(ie LoT is wrong).
And quickly on your bioshock 2 stuff iqgamer bs2: The effects of this can be seen above. Notice how the water running down the stairs is much blurrier than the surrounding stairwell and the stairs themselves.
Try putting those two pictures side-by-side, and you'll see the 360 version is slightly further away and at a 20% degree angle to the stairs(which will help the water FX look nicer because of the specular reflection equation on square geometry), but when you match the gamma of the two images (increasing gamma to 1.4 on 360; at the expense of saturating whites) you start to see the 360 water exhibit the same problems.
iqgamer bs2:In addition it seems that there is less, or more subtle use of bump mapping on the PS3 when compared to the 360
There are storage and performance advantages to bump maps(8bit integer per component) especially on ATI cards and with EDRAM being 10MB, but I'd be quite surprised if bump mapping gets any use at all in modern PS3/PC games in favour of normal mapping(16bit floats per component).
OpenCL might be 2years old on PS3, but changing work flows isn't instant for developers or middleware companies, and can take two software iterations to become widespread. But for the last 12months I haven't agreed with too many of DF's conclusions, and would say changes started happening, but were still incorrectly analysed.
Unless these games like Crysis 2, etc are open source software to look through the actual code, LoT, iQgamer & DF aren't providing any more proof than we are by discussing stuff, are they? Resolution(pixel counting), frame-rate, screen-tearing and pad response is all they have extra.
Comment below viewing threshold Show
Comment below viewing threshold Show
We already have! FFXIII, all alpha effects is rendered in full resolution and the game is also running at 1280x720 with 2xMSAA + vsync. The same game on 360 is sub-HD, most likley NOT because of bad porting but because of the 10Mb eDRAM. 1280x720x2 gives us a 14Mb framebuffer.
/ Ken
Comment below viewing threshold Show
[link url=http://imagequalitymatters.blogspot.com/2010/03/tech-analysis-final-fantasy-xiii-ps3-vs.html
]http://imagequalitymatters.blogspot.com/...[/link]
"One area however, which is like for like across both platforms, is the use of Alpha to Coverage (A2C) for transparency effects and particles. When using A2C in order to render transparencies, instead of rendering a whole transparent texture, the A2C produces an interlacing style effect, creating an almost dithered look to things. It’s kind of like a mild screen door type effect, used to half the amount of bandwidth needed for such effects. The advantage is that you can render full resolution transparencies with lower cost than if you were rendering them as a whole solid effect.
All of the transparent elements of characters facial hair, except eyebrows are rendered using A2C, including the hair on their heads, and even eyelashes too. Also, numerous particle, and smoke effects are rendered this way, though not all, to keep bandwidth under control.
Unfortunately, the 360 version not only uses A2C in order to fit the framebuffer into EDRAM, it also renders lower resolution transparencies as well, due to the reduced overall rendering resolution, making the effects look even worse on that build than they should. The PS3 has no such issues, other than the interlacing style look to anything see-through, because all these effects are rendered in 720p. Quite how SQE couldn’t take advantage of the 360’s near limitless amount of bandwidth to deliver full resolution transparent effects is unknown, but we feel it’s a case of why bother, rather than how, given the short conversion time and rushed approach to 360 development."
Comment below viewing threshold Show
Probably because the streaming of geometry for the game world was add latency and costing bandwidth, and the real-time shadow mapping and lower resolution transparencies were already using up the Xenos 500M polygon/sec limit(30fps = 17Million), leaving very little bandwidth remaining to do a proper depth cueing fog.
Comment below viewing threshold Show
Every last post contains some version of the classic con: bait-and-switch. Clearest example: In a discussion pertaining to practical render-target bandwidth discrepancies, he misdirects you to the bandwidth "inside the SPUs"; which is indeed impressive and a remarkable feature of the PS3, but in no way does it change / compensate for the barely adequate video-RAM bandwidth. It has absolutely no bearing on the subject at hand, and is therefore effectively willful misinformation. But then he really seals the deal with this little gem:
vizzini - "The size of alpha blending FX on PS3 I would assume are the result of being done on SPUs without tiling(with less than 256KB for data). "
And such an idiotic assumption that is too... and I mean seriously, which is it vizzini? Either RSX is such a great little piece of kit, in reality superior to Xenos; OR it's so over-burdened that it's routinely resorting to offloading such a simple task as alpha-blending to the SPUs, which would represent a grossly inefficient use of their cycles anyway. This certainly isn't the fist time I've noticed you tripping over your own arguments, or their lack of integrity, but the most blatantly absurd.
Anyway thanks for that one, really, it provided a truly pristine example of how much BS all your professed expertise is; and how ultimately pure nonsense all your techno-babble is.
vizzini - 'I took the first image from the bioshock 2 link, visual inspected for crushed blacks, full frustum exponent fog, and then used a paint package to process A-B(gamma increase 4.0), B-A(gi 4.0), difference of A & B(gi 4.0). Which certainly suggests the opposite scenario for texture detail, as all the high frequency detail is in the PS3 image, and all the low band frequency differences were from the 360 image."
Still acting like you have any credibility against DF's after your own obstinate 'analysis' of those TFUII screens turned out to be garbage? I gotta give you credit for the balls at least...
It's exactly this kind of bullshit that underscores the genuine distinction between those relative "fanboys". All those "ten year olds" on the XBOT side are obvious and open targets (which is great actually, as they expose the motives of the people who engage them). Whereas being "SDF" really doesn't mean you're a flaming fanboy at all, which is why I will use the term because it's exceedingly descriptive and appropriate. There's just such a vast contingent who maintain this whole pretense that they're the reasonable, informed, and fair-minded ones, but all just as a cover for insinuating the same knee-jerk partisan claims in the end.
Comment below viewing threshold Show
The z-buffering, polygon rendering by the RSX are great, and I was pretty sure that is what I said all along. But if at any stage I've given the impression/said I thought the RSX had anywhere near the fragment shading(not vertex) capacity of the Xenos, then that was clearly wrong. As it can do it job well, clearly by how goodl Infamous looks utilising only 20% of the SPUs. But we've all been talking for some time about the offloading of fragment work to the 6xSPUs.
We all know the theoretical technical specs of the 360, and the Xenos has huge bandwidth, but it is very hard to avoid wasting that resource while producing a 720p@30fps(32bit depth buffered), HDR(deep colour), triple buffered and v-synched image, as the Crysis 2 demo and many other games (Halo 3, Reach, Alan Wake, Fable, Split/second, etc) have corroborated.
I do try to remain reasonable, but even people who have contributed articles to develop magazine over the last 2years aren't.
One developer had the audacity to make out that it was Sony's fault that he bought a shed load of entry level S model Bravia's LCDs that weren't suited to their Q/A needs(but failed to check the specs before buying); and then continued to complain about the additional resolution the screens above 1280x720p (1360x768) as though that is bad for customers and picture quality.
Comment below viewing threshold Show
And with the z-buffer again, which seems to be your new favorite straw man. There are work-arounds for any issues a 24-bit depth-buffer limitation may raise, and indeed even a 32-bit z-buffer doesn't eliminate those same issues, it just affords extra leeway before they may become unacceptable. But you keep bringing it up as some general IQ factor when in practice it's a complete non-issue for most games, and may amount to some virtually unnoticeable artifacts for almost all the rest. And for the rare cases where it's deemed a significant issue, well then it'll get worked around. Not without some cost of course, but then running a 32-bit z-buffer also comes at additional cost. Which is the core issue around which your whole tactic gets especially ridiculous:
"Name any recent AAA game(last 2years) on the 360 that you definitely know uses more than a fast/less accurate 24bit depth buffer with 24bit colour(with x8 HDR halo:reach), and then try and find any recent game on PS3 that uses less than a 32bit depth buffer with less than x128 HDR."
Lets be clear I'm even giving benefit of the doubt to your larger implication, that this is some hard deficiency of the Xenos or even 360's API. (I'm inclined to think it's just another assumption you've made based on ATI's PC cards, which even then may be limited at the driver level rather than the hardware.) But since you're the one making the challenge in those terms, it behooves you to actually source at least some of your case as well. Specifically I'll ask which PS3 games qualify as x128 HDR?
But as you didn't... two that seem most likely are GoW3 and LBP(2), which are both perfect examples of games that could gain literally nothing from that sensational 32-bit z-buffer. Neither exhibit any hard draw distance / FOV that could come close to taxing a 24-bit, so... if they're actually using a 32-bit z-buffer, it'd seem like really stupid engine coding or; if the PS3 API actually forces it's use, that would be another liability for the majority of cases?
I'm really just being gratuitous though, since that's rather moot in comparison to the core fallacy:
"We all know the theoretical technical specs of the 360, and the Xenos has huge bandwidth, but it is very hard to avoid wasting that resource while producing a 720p@30fps(32bit depth buffered), HDR(deep colour), triple buffered and v-synched image, as the Crysis 2 demo and many other games (Halo 3, Reach, Alan Wake, Fable, Split/second, etc) have corroborated."
That really reveals how off your perspective is vizzini: of course 360 games have to deal with the eDRAM limitation one way or another, but the bandwidth is effectively a non-issue. From everything I've picked up on, it would be much more accurate to say that it's impossible to tap more than a fraction that resource, unless tiling or other compositing techniques are being employed. The magnitude of that resource is precisely so it can be 'wasted' by the ultra-fast transfer to main RAM, MSAA or other internal logic operations. In practice, I'm sure even with 4x MSAA for the most part, the hard limits of Xenos' fillrate will come into play long before any ultimate limits of the bandwidth.
Which makes every one of your allusions to bandwidth saving factors for 360 games impertinent in general, and usually laughable in specifics, especially in a relative context to PS3. Source after source has referenced the geometry overhead as the drawback to predicated tiling, never once have I seen any lament of the bandwidth usage, because it's plainly not a general issue in the total equation.
Whereas source after source corroborates the bandwidth as a chronic hurdle for PS3 examples, yet you blithely suggest all these bandwidth overhead saving measures aren't being employed as well? And then bringing the 32-bit z-buffer up in that argument is true comedy, as though that extra 8-bits could account for any notable portion of the gulf in exhibited performance seen on the really alpha-heavy examples. (And again, if PS3 actually was using a 32-bit z-buffer in those examples, and to no visible gain, isn't that pretty damn stupid to waste the precious bandwidth?!?)
But you either already know all this - and I honestly think you may be literally a joke vizzini - or you're actually as fragmented in your understanding as you seem, in which case I'm sure I've not made a dent. But it's because of your own brand of propaganda (and of course others) that it's been so nearly impossible to have any genuine, and yes even genuinely biased, debate about actual technical concerns on this forum. Because it's been so polluted by misdirection that sounds like it makes sense... but more to the point that sounds like what people already want to believe.
Comment below viewing threshold Show
From article: Tech Analysis: Final Fantasy XIII
[link url=http://www.eurogamer.net/articles/tech-analysis-final-fantasy-xiii-article
]http://www.eurogamer.net/articles/tech-a...[/link]
Perhaps surprisingly, FFXIII in retail form still retains the full-resolution alpha buffers, and it is interesting to note that these still cause the game some issues - but the impact is definitely lower than it was before.
I also reread and rechecked the other articles and it seems like the hair on the characters is dithered. But it seems that Modern Warefare 2 on PS3 is running with the same alpha buffer as the 360 version
But Infinity Ward has not, the buffer sizes appear to be an exact match for the Xbox 360 game, and we can't help but feel that this has a lot to do with the performance difference.
[link url=http://www.eurogamer.net/articles/digitalfoundry-modern-warfare-2-face-off?page=3
]http://www.eurogamer.net/articles/digita...[/link]
Though with a slight difference in performance.
I also think Ninja Gaiden Sigma 2 on the PS3 runs with full resolution aplha buffers if I remember correctly.
/ Ken
Comment below viewing threshold Show
If you ever studied the maximum flow problem you'd understand that bottlenecks are to be avoided at all costs, because any time spent waiting (even for fast memory) is eating into the effective throughput.
If either the RSX/Xenos stalls for any time while vertex data is transferred from RAM, then the polys/sec and fragment/pixel shader throughput gets proportionally decreased. Or in the case of fragment shaders, is required to be given a longer task keeping busy until the stall + vertex pipeline processing feeds it new fragments.
That is what you meant by, “precisely so it can be 'wasted'” is it not?
Briefly on depth buffering comment; absent or sparse depth cueing fog(AC1/2/Brotherhood), or shorter frustum projections (using the same assets in games like Mafia 2's bonnet camera) indicate a smaller zbuffer size in use on the 360 than PC/PS3. As does overly large geometry seen in many 1st party 360 games like Kameo, Fable, Reach; which I assume is to avoid z-fighting on a smaller zbuffer, and not an art quality problem.
A lack of self shadowing at distance specific to the 360 demo/game of Sega All Stars Racing in an overhanging cliff, also suggests this to me; but you are right in saying I'm going mostly from PC programming with ATI/Nvidia cards.
But it could also be because I spent a year getting an algorithm to work using (h/w accelerated) 32bits for depth buffering, that yields the benefits of the painter's algorithm, but with very little of its overdraw problem. But if you can actually point to other evidence that supports a different conclusion then I'll look at it.
Comment below viewing threshold Show
ken from same article u posted
"Final Fantasy XIII is a 720p game, but features some pretty decent upscaling for 1080i and 1080p support.
One of the most noticeable compromises seen in the original demo has made it through to the retail version: an effect known as Alpha to Coverage. Rather than render a complete, seethrough texture, A2C utilises an interlacing style effect instead. Exactly why Square has utilised it here remains unknown, but it is safe to say that it is performance-related, and that it all comes back to the notion of those performance-sapping alpha buffers.
Comment below viewing threshold Show
One of the most noticeable compromises seen in the original demo has made it through to the retail version: an effect known as Alpha to Coverage. Rather than render a complete, seethrough texture, A2C utilises an interlacing style effect instead. Exactly why Square has utilised it here remains unknown, but it is safe to say that it is performance-related, and that it all comes back to the notion of those performance-sapping alpha buffers.
Essentially, transparent elements in characters' hair are rendered using A2C, and it's used for all types of facial hair (eyebrows aside) right down to the eye-lashes. It can be a fairly ugly effect, but the implementation in FFXIII allows for the layering of hair, all of which is animated, adding a lot more "life" to the characters.
Just because they are using A2C on the hair doesnt' mean that the game is not using full resolution alpha buffers on other stuff. For exmaple the spell effects is rendered with fully rendered alpha.
/ Ken
Comment below viewing threshold Show
Fair enough mate, my initial point was simply that most devs note bandwidth issues on ps3 as an issue and that alpha transparencies are bandwidth heavy and often reduced in quality on ps3 games, I know I asked Vizzini about whether we'd finally see them, but what I really meant i guess was would we see them in games which are heavy on bandwidth where traditionally the devs options have been to save bandwidth by reducing resolution on ps3 .
Comment below viewing threshold Show
From article : Face-Off: Battlefield: Bad Company 2
[link url=http://www.eurogamer.net/articles/digitalfoundry-faceoff-battlefield-bad-company-2
]http://www.eurogamer.net/articles/digita...[/link]
A "screen door" dither effect on some foliage, most obvious in the snow levels, looks a touch rough on Xbox 360 while it is not an issue at all on PlayStation 3. While it's not exactly a huge problem, it's undoubtedly the case that the PS3 version looks better without it, using a more conventional alpha-test technique for its transparency.
360 is using A2C so it gets the same dithering effect as you can see on the hair on FFXIII but PS3 is using a more conventional alpha-test technique. See pictures below.
360:
[link url=http://images.eurogamer.net/assets/articles//a/9/9/1/9/0/7/360_screendoor2.jpg.jpg
]http://images.eurogamer.net/assets/artic...[/link]
PS3:
[link url=http://images.eurogamer.net/assets/articles//a/9/9/1/9/0/7/PS3_screendoor2.jpg.jpg
]http://images.eurogamer.net/assets/artic...[/link]
/ Ken
Comment below viewing threshold Show
Comment below viewing threshold Show
/ Ken
Comment below viewing threshold Show
/ Ken
I never suggested that the ps3 couldnt do fully rendered alpha buffering mate, i was basically saying what you were saying, in a very ineloquent fashion due to my lack of tech lingo and understanding. You have summed up nicely what I wanted to say and kinda brought me back to where this conversation started with another poster, that both have strengths and weaknesses. I merely suggested that ps3 games often choose to not render them in full resolution, or use alternative methods to mimic these effects, to save on bandwidth. Vizzini suggested that the ps3 doesnt in reality have bandwidth issues, and that it in reality has a better bandwidth, but that this isnt evident due to the lack of programmers utilising ps3 properly, and that they can circumvent the issues it has with bandwidth via open cl developments.
I have no fan boy status as i own both consoles, but am curious about these issues and like to understand the basics, and keep my eye on material on subject. (not that i understand it all lol)
Comment below viewing threshold Show
"If you ever studied the maximum flow problem you'd understand that bottlenecks are to be avoided at all costs, because any time spent waiting (even for fast memory) is eating into the effective throughput."
Oh that's beautiful, I mean, I can't possibly argue against the MAXIMUM FLOW PROBLEM! Yeah I have to admit defeat now...
But wait, did I just find a glimmer of hope? Seriously, I'd love to see some sort of conclusive maximum flow problem result for both PS3 and 360, lets find out once and for all just at what costs Sony avoided bottlenecks, and how successfully.
But setting aside how freaking ridiculous that particular tactic was, in general it's just the same line you always take: trumping up every possible drawback, cost, or trade-off you can hypothesize about how 360 operates, while perpetually implying none of it would ever apply to all the fantastic features of the PS3. Taking any large sample of your contributions here, your own bias is concrete vizzini; and as such it's my main "evidence" for a possible "different conclusion": Your bias severely compromises you objectivity, and as such your questions / premises have negligible credibility as a starting point.
"Briefly on depth buffering comment; absent or sparse depth cueing fog(AC1/2/Brotherhood), or shorter frustum projections (using the same assets in games like Mafia 2's bonnet camera) indicate a smaller zbuffer size in use on the 360 than PC/PS3. As does overly large geometry seen in many 1st party 360 games like Kameo, Fable, Reach; which I assume is to avoid z-fighting on a smaller zbuffer, and not an art quality problem.
A lack of self shadowing at distance specific to the 360 demo/game of Sega All Stars Racing in an overhanging cliff, also suggests this to me; but you are right in saying I'm going mostly from PC programming with ATI/Nvidia cards."
First off, I'd just like to dispense with my acknowledgment of your point about such a technical showcase as Sega Allstars Racing! /facepalm
Now about the depth cueing in the AC games, are you just putting your own visual 'analysis' up against DF's and others again? (and ironically probably based on screenshots provided BY Digital Foundry and others) Because I can't say I really know what that's about, other than maybe you confusing the Quincunx effect with something else.
And the Mafia II, frustum thingy reference again... ahhh hell. I resisted this chore once before, though mostly cause I was preoccupied with <a href ="http://www.eurogamer.net/articles/digitalfoundry-call-of-duty-black-ops-faceoff/comments?comment_start=300">Another Fish Fry</a>, but as I'm now committed (and yeah, probably should be "committed" as well now):
"vizzini
17/11/10 @ 12:00
A quick question BuckEntropy, on the topic of Bayonetta; did you not thinking that the gaming areas looked and felt very small? like they were only using a 16 or 24 bit depth buffer? As a 16/24bit depth buffer is a massive help in reducing bandwidth demands on an ATI card from my own development work. Not to mention that the near frustum clip plane is kept a good distance away from the camera in this 3rd person game, which keeps the alpha blending contained in a smaller area.
I thought it felt like they'd used Kameo's poor old engine in Bayonetta, with everything being very sparse of geometry, or largely stretched polygons boxes that looked very last-gen. Trying both demos one after the other and they feel like they were the same game engine.
It also seems like they offset the alpha blending fillrate against the many fighting regions where the skybox was just background fill colour and there was no geometry to tax the GPU with the alpha blending.
As I have implied before(in other comments), the Xenos appears to be using integer 8bit per component RGBA colour and a 24bit zbuffer(with lossy 3Dc textures) in some games; which gives it a performance advantage in frame-rate tests, and I wouldn't be surprised if that is what is happening again here.
You can clearly see in games like Mafia 2 that the RSX is using a full 32bit zbuffer (from the car bonnet view) and the Xenos is using a 16/24bit with its reduced frustum. And the gamma incorrect colours using Directx9c on PC are consistent with integer 8 bit per component on the Xenos, where as the PS3 version colours are consistent with floating point colour of DirectX10/11 on the PC version.
It is easy to think the Xenos is better than the RSX, if you ignore the telling signs when you compare demanding PC games to PS3 versions, and PC to 360 versions for image quality."
First on the subject of Mafia II, WTF are you actually suggesting there? You've made an argument about "z-fighting artifacts" which seems a technically grounded point... but it seems to me the only possible relevance any minor difference in the overall view frustum could have is on those artifacts, yet it doesn't appear to be a relative issue in the first place. Otherwise, such a discrepancy as 256 times greater precision would certainly have to manifest in a glaring difference in the far-plane clipping, which I'm quite sure is not the case? But then the whole notion that this could only become an issue for the "car bonnet view" undermines any validity to the argument, or even at the least serves as reinforcement that it's an ultimately moot issue!
But that Bayonetta argument... like OMG! and WTF! and RUFR?!?! (
Comment below viewing threshold Show
Both of those assertions - that the alpha-blending is "contained in a smaller area" or that it's only prevalent against the "skybox" without any "geometry" concerns... are just the most abject nonsense, beyond my autonomous imagining. You CANNOT have even played the game, or you're just that cynical. Culling is not even in Bayonetta's technical vocabulary, and the transparencies may fill the entire screen in multiples without the slightest nod to discretion - and indeed to the severe detriment to performance, even for 360 in some places. And as for the geometry and general modeling... IMO the game qualifies as a case study for the point you seemed to be on in this thread about production overcoming subjective aliasing more than any nominal tech:
vizzini - "Greater geometry sub-division (ie more front facing polygons of smaller size) and better scenery/character modelling at the development level will provide higher fidelity imaging with smaller aliasing than this will."
Because I was actually caught off-guard by the revelation Bayonetta employs no MSAA, almost didn't want to believe it even. And in resolving that reaction to the truth, it was clear to me the judiciously exceptional geometry usage, as well as the superlative overall texture quality get main credit for the subjective response. Plus how the lighting is crafted - whether true HDR or not(?) - ties all the imaging together with a rare coherence. So to put it bluntly: BULLSHIT! Even in general Bayonetta is well above any median this gen for geometry and 'fragment' proficiency, and truly elite for general craftsmanship; but then the contemporary novelty of 60fps cannot even be overstated, I mean how many other games can it legitimately be compared to? Sure as HELL not God of War 3.
Which makes a perfect segway back to your non-issue about the z-buffer and it's rather negligible impact on RENDER TARGET bandwidth. You were suggesting Bayonetta's "gaming areas looked and felt very small" compared to WHAT? Maybe God of War 3?!? Fixed cameras and pre-scripted / on-rails set pieces do not make for a convincing testament to the PS3's depth precision prowess, SORRY, don't even try to pretend it's a contest there. Yet even the almighty UC2 is about as binned-off in reality as Bayonetta, if anything a few of the more showcase areas make it clear just how legitimate the draw distance is in Bayonetta, even by contrast.
Lest I come across as a total Bayonetta fanboy, I'll just admit it's become a centerpiece for my whole motivation here. But there's really nothing else like it, and rather than get the simple credit it's due, it's become mostly a symbol in this whole bogus controversy: "NO! We CANNOT accept that there's A SINGLE THING 360 might be able to do better / more efficiently than the PS3!" And you've made of yourself the epitome, nay a veritable rallying point for that grand fallacy vizzini. And my personal assertion: judged in context; or more so by the criteria of congruence, I'm not sure if I've ever witnessed a steamier pile of partisan pandering garbage than your contributions to this site vizzini, in aggregate.
Which is especially sad given that I have no problem believing you actually 'know' more than me about many specifics. But I still know bullshit when I see it, and the entirety of your understanding has been channeled into charlatan imperatives. As the axiom goes... grow your tree of falsehood from a small grain of truth... it's really too bad your agenda is more important than the grains of truth you clearly have, or I might be able to learn something a lot more interesting from these exchanges.
Comment below viewing threshold Show
I won't bother to re-iterate/re-substantiate the things I've previously wrote, or quote your last two comments yet, in case you wish to reduce and alter them.
Comment below viewing threshold Show
Seriously, that's the example you're going to use? And HokutoNoKen has just been flogging a single comatose horse for the entire year I've been reading, so yeah it'd be pretty easy to keep it simple when you're just repeating a talking point over and over.
So I'll decline to "reduce" anything vizzini, there's actually plenty more I could have said anyway. And as I've mentioned before, even when I do try to keep things simple, my antagonists always do their best to make me regret it.
Comment below viewing threshold Show
vizzini - "If either the RSX/Xenos stalls for any time while vertex data is transferred from RAM, then the polys/sec and fragment/pixel shader throughput gets proportionally decreased. Or in the case of fragment shaders, is required to be given a longer task keeping busy until the stall + vertex pipeline processing feeds it new fragments.
That is what you meant by, “precisely so it can be 'wasted'” is it not?"
Again you're on a fundamentally different issue from that of render-target bandwidth, and being impertinent to the discussion of alpha overhead anyway. But even if the 360 architecture was at some technical bandwidth disadvantage in asset fetch from main-RAM (and as always, I'll be interested in any source indicating that) it's exceedingly well established that it's no disadvantage in practice. And isn't that precisely the point of Xenos' unified shader architecture?
So no, and I'll illustrate the point I was making from the other direction, once again using the alpha / overdraw in Bayonetta as example: No MSAA and also no tiling is employed in the game, which means there's virtually no "waste" at all of the eDRAM's bandwidth; other than the single frame dump, presumably there's nothing else competing with the GPU for that render-target bandwidth. Now comparing to the PS3 version, even with it's reduced resolution alpha-buffers and 60-70% average framerate performance, we could guess the 360 version's overall bandwidth load is still only around twice that of the PS3 game. But lets just round up and say it's 3x, that would still indicate Bayonetta is soaking up barely half the bandwidth overhead available from the eDRAM. So then it's truly a wasted resource, when the GPU's fillrate is the only thing taxing it.
Comment below viewing threshold Show
But do you realise that very little of what you are saying makes sense(graphics wise)? And if you don't fully understand computer systems and computer graphics to be able to contribute properly there are two better options available to you:
1.Read what other more informed people are saying without getting offended or needing to comment.
2 Start learning by buying & reading books about the subjects. Computer graphics principles in practice(for reference), the Official OpenGL Red book, Orange book and developer.nvidia.com (to read techniques from GPU Gems 1, 2, 3)
Comment below viewing threshold Show
I certainly don't claim to have all the facts, but I'm also not the one continually implying 'the facts' are contrary to the manifest examples these systems have produced. If you, or anyone else, care to explain to me in good faith any specific claim I have made that doesn't make sense "graphics wise", I certainly wont take offense. But you haven't, and after that little cop-out, I honestly don't believe you really could even if you wanted to.
And don't pretend you're the shining example of other more informed people on this forum, and it's not like you haven't been at odds with others many times. And those others don't betray their credibility as a matter of routine.
Almost none of what you say ultimately makes sense logic wise vizzini, which is something I suppose I do take offense at...