Neverending Upscaling/Resolutions/AA etc Thread #2 *Rules: post: #616 *

[Quaz51]

the light seem more real, maybe more complex lighting or post-process on PS3?
http://www.vertigogaming.net/PSP/GTA/gtaps31.jpg

i would like to see this scene on X360 for compar:
http://img246.imageshack.us/img246/6668/wmplayer200804280007241zo6.png

[betan]

PS3 screenshots, quaz.

http://img170.imageshack.us/img170/6858/wmplayer200804272348001sz1.png


where are those from?
The lighting looks pretty good.
those are pretty lousy shots

Enough to confirm native resolution though.

[AlStrong]

http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png
That one *may* work.

Looks....blurry. :|

[Bellic]

http://dot50cal.the-horror.com/b3d/devshot1.png

Not seeing any difference

[Dot50Cal]

Quaz, I cant find those area's in the game (im not very far). Heres some shots that might help.
http://dot50cal.the-horror.com/b3d/gta4/360

[Juan Panson]

here a low quality 360 screen with similar lighting conditions. grabbed it from another forum.

http://i128.photobucket.com/albums/p177/blubbernalgas/vlcsnap-1689627.png

[MazingerDUDE]

judging by the screen captures provided, 360 version seems suffer terrible dithering



http://www3.telus.net/public/dhwag/400360.jpg

http://www3.telus.net/public/dhwag/4003602.jpg

http://www3.telus.net/public/dhwag/4003603.jpg




http://www3.telus.net/public/dhwag/400PS3.jpg

http://www3.telus.net/public/dhwag/400PS3B.jpg

http://www3.telus.net/public/dhwag/400PS3C.jpg



Since both version of GTA had the same rendering res, the same level of AA, IGN review comments like 'more AA on PS3, or PS3 version smoother' are probably related to dither in 360 verison.

If that's the case, what could have caused this in 360 version?

Could it be the level of HDR? (like the limited RGB pallete of FP10 HDR commonly used in 360 games)

Actually, this is not the first time I have noticed such difference, the Burnout Paradise had 360 version to dither a bit especially for distant object like trees.

[Quaz51]

Enough to confirm native resolution though.

yes :)

[inefficient]

judging by the screen captures provided, 360 version seems suffer terrible dithering




Hmmm that is pretty grainy looking. I guess we don't know yet if that was intentional artistic direction, or if it was just some kind of side effect of the filters they are running. Sam Houser did say the 360 version had a "certain look to it".

[Rangers]

yes :)

And ? :grin:

[inefficient]

Are we sure this is not a bullshot?
http://img170.imageshack.us/img170/6858/wmplayer200804272348001sz1.png

It looks very very different from the 360 version to my eyes. It looks very smooth and there is none of that posterization effect or grainy filter we see on the 360 frame grabs.

[Dot50Cal]

efficient, as far as I know that is from a video on IGN comparing the two. So I doubt its a bullshot. But I don't think its fair to compare screens off a lossy video to lossless shots off the framebuffer. Capture cards and the eventual re-encode would destroy fine dithering details like that.

[Quaz51]

And ? :grin:

it's 630p (1120x630) :oops:
regardless PS3 media look better for me, the light and atmosphere is more real (maybe HDR vs X360 MDR, post-process...), i agree with global reviews feeling but more comparison screenshots would be good
it's interesting, it shown that number of pixel isn't enough for determine better visual

[Quaz51]

Are we sure this is not a bullshot?
http://img170.imageshack.us/img170/6858/wmplayer200804272348001sz1.png

It looks very very different from the 360 version to my eyes. It looks very smooth and there is none of that posterization effect or grainy filter we see on the 360 frame grabs.

if this screenshot is bullshot then it's bullshot in 630p :)

[Rangers]

efficient, as far as I know that is from a video on IGN comparing the two. So I doubt its a bullshot. But I don't think its fair to compare screens off a lossy video to lossless shots off the framebuffer. Capture cards and the eventual re-encode would destroy fine dithering details like that.

It's not in the review or the insider head 2 head I dont think, anybody know which IGN vid it's from?

[Dot50Cal]

Thanks Quaz, as always a nice job!

[Betanumerical]

it's 630p (1120x630) :oops:
regardless PS3 media look better for me, the light and atmosphere is more real (maybe HDR vs X360 MDR, post-process...), i agree with global reviews feeling but more comparison screenshots would be good
it's interesting, it shown that number of pixel isn't enough for determine better visual

How much AA.?.

[Hardknock]

it's 630p (1120x630) :oops:
regardless PS3 media look better for me, the light and atmosphere is more real (maybe HDR vs X360 MDR, post-process...), i agree with global reviews feeling but more comparison screenshots would be good
it's interesting, it shown that number of pixel isn't enough for determine better visual

Quaz have you seen shots from identical locations and time of day? Both have HDR lighting. Please don't let bias start to creep into your findings. That wouldn't be good. Great work by the way.

[Quaz51]

Thanks Quaz, as always a nice job!

i know the resolution since yesterday and first IGN video grab but i don't sure that is the good moment for say it :)
i think it's not very pertinent in this case, for lot of people sub720p = worst version, it going to generate confusion

[Neb]

So the PS3 version is 630p... internet meltdown confirmed (NeoGAF explodes)!

Neverthless of the ss shown so far the PS3 version looks nicer IMO. Everything just blends in nicely and makes a seamless contrast between front objects and background objects.

[sevanig]

Maybe IGN capture video/images at 630p

Someone check a 360 IGN image...

[Hardknock]

i know the resolution since yesterday and first IGN video grab but i don't sure that is the good moment for say it :)
i think it's not very pertinent in this case, for lot of people sub720p = worst version, it going to generate confusion

so you've known since yesterday but didn't want to say in fear of hurting the PS3 version's sales?

[Dot50Cal]

Maybe IGN capture video/images at 630p

Someone check a 360 IGN image...

Thats not how it works dude.

[Rangers]

So the PS3 version is 630p... internet meltdown confirmed (NeoGAF explodes)!

Neverthless of the ss shown so far the PS3 version looks nicer IMO. Everything just blends in nicely and makes a seamless contrast between front objects and background objects.

Then it begs the question if people are going to say they prefer the look of the PS3 version, maybe Rockstar should have downscaled 360 and added more supposed postprocessing, HDR, AA, or whatever if possible?

[Quaz51]

Quaz have you seen shots from identical locations and time of day? Both have HDR lighting. Please don't let bias start to creep into your findings. That wouldn't be good. Great work by the way.

yes both "HDR lighting" but when i say HDR vs MDR i talk FP10 mode use on X360. for simulate HDR lighting on X360 games use generally FP10 (except Halo3 :) ) but FP10 isn't really HDR, it's rather MDR precision. On PS3 there is no FP10 mode then generally it's LDR or HDR (FP16 or other software method), maybe GTA PS3 use HDR (or not) that make little difference or maybe post-process, just hypotesis, anyway PS3 version look different. i would like to see screenshots in same context

[warb]

so you've known since yesterday but didn't want to say in fear of hurting the PS3 version's sales?
Don't do that.

[sevanig]

Thats not how it works dude.

Hmmm ok... I assumed they just used a HD capture card...

[Quaz51]

so you've known since yesterday but didn't want to say in fear of hurting the PS3 version's sales?

in truth i waited others PS3 screenshots for confirm but PS3 screenshots are unobtainable

[Rangers]

Suddenly there is 598 viewing the console forum :lol:

Yep, GAF has been alerted..

Quaz you're like a net celebrity nowdays..and consulting on the eurogamer stuff!

Are you in the games line of work? How know so much?

[sevanig]

Suddenly there is 598 viewing the console forum :lol:

Yep, GAF has been alerted..


Insert Animated Gif

[Quaz51]

Maybe IGN capture video/images at 630p

Someone check a 360 IGN image...

the rescaling (video or screenshot) don't corrupt the native resolution

[Rangers]

in truth i waited others PS3 screenshots for confirm but PS3 screenshots are unobtainable

I wonder if there were no PS3 screens available just to avoid you, Quaz.

Seriously, look at the hubbub Halo 3 being sub 720 caused..I'm sure publishers noticed.

Although arguably it might have been better to get th whole 630P thing out of the way six months ago, rather than it break the day before release..such as it matters at all to the casual public.

[-tkf-]

How about Eurogamer screens?
http://www.eurogamer.net/article_gallery.php?article_id=134493&image=assets/articles//a/1/3/4/4/9/3/5.jpg.jpg&position=4

[Quaz51]

How about Eurogamer screens?
http://www.eurogamer.net/article_gallery.php?article_id=134493&image=assets/articles//a/1/3/4/4/9/3/5.jpg.jpg&position=4

it's X360 screenshots (i look MSAA pattern to know)

[Betanumerical]

it's X360 screenshots (i look MSAA pattern to know)

About AA, how much did the PS3 shot have? 2xQ I presume.

[function]

So the PS3 version is somewhat lower resolution, which explains why the PS3 version looks sharper? Also you can't make out the clearly intended "artistic" grain effect, so that's another bonus because it looks cleaner! ;)

As a developer it must be disheartening to develop a nice over-sharpened grain effect for your game to finish off the look you've spent years developing, and have journalists desperate to prove how good they are at "spotting stuff" say if makes the game look "less clean" and give the nod to the low res version. :p

I bet Nintendo were glad to duck under this kind of stuff this round.

[-tkf-]

it's X360 screenshots (i look MSAA pattern to know)

Yeah, after looking again it looks like many of those screenshots have the weird dithering pattern. It must look very noisy in motion.

[FirewalkR]

Post #513 is the beginning of the end of the Internets. :lol:

[zarbaj]

Quaz, now that GTA4 is kinda outta the way, can you please tell me about Ninja Gaiden 2 res/aa?

some images:

http://www.gamersyde.com/gallery_8146_en.html
http://www.gamersyde.com/gallery_7706_en.html

and sorry for troubling you so much.

[Quaz51]

Yeah, after looking again it looks like many of those screenshots have the weird dithering pattern. It must look very noisy in motion.

no i talk the MSAAx2 pattern, according to the aspect of the AA on the edge i know if it's a PS3 screenshot or a X360 screenshot (on any game, not only GTA, but need good quality screenshot)

[FirewalkR]

no i talk the MSAAx2 pattern, according to the aspect of the AA on the edge i know if it's a PS3 screenshot or a X360 screenshot (on any game, not only GTA, but need good quality screenshot)

That's due to differences between Nvidia and ATI rendering processes, right?

[AlStrong]

Quaz, now that GTA4 is kinda outta the way, can you please tell me about Ninja Gaiden 2 res/aa?

some images:

http://www.gamersyde.com/gallery_8146_en.html
http://www.gamersyde.com/gallery_7706_en.html

and sorry for troubling you so much.

http://forum.beyond3d.com/showpost.php?p=1150738&postcount=400

[Galduta]

it's 630p (1120x630) :oops:
regardless PS3 media look better for me, the light and atmosphere is more real (maybe HDR vs X360 MDR, post-process...)l

the HDR FP16 + MSAA working in the RSX ? Because in some games like Assesin,s Creed (http://www.pcgameshardware.de/aid,637652/News/Assassins_Creed_Technical_QundA/) or R6 Las vegas, the game don´t have HDR

Perdon if already explained in Beyond :oops:

[zarbaj]

http://forum.beyond3d.com/showpost.php?p=1150738&postcount=400

shame.

thanks to both of you btw.

[Quaz51]

That's due to differences between Nvidia and ATI rendering processes, right?

yes, reverse pattern msaa

[inefficient]

So the PS3 version is somewhat lower resolution, which explains why the PS3 version looks sharper? Also you can't make out the clearly intended "artistic" grain effect, so that's another bonus because it looks cleaner! ;)

As a developer it must be disheartening to develop a nice over-sharpened grain effect for your game to finish off the look you've spent years developing, and have journalists desperate to prove how good they are at "spotting stuff" say if makes the game look "less clean" and give the nod to the low res version. :p


Well Rockstar co-founder Sam Houser himself wrote "The 360 games have a certain look to them; PS3 games have a certain look to them. I like the way [the PS3] renders. There's a certain kind of softness without being blurry -- some warmth to it -- and then there's a certain more clinical element to how the 360 looks."

They obviously are well aware of the visual differences. Those differences exist to taylor to the strengths of each machine. The actual artists and coders are probably smiling with glee at how much attention individual pixels of their work are getting right now.

[Shifty Geezer]

judging by the screen captures provided, 360 version seems suffer terrible dithering

Since both version of GTA had the same rendering res, the same level of AA, IGN review comments like 'more AA on PS3, or PS3 version smoother' are probably related to dither in 360 verison.

If that's the case, what could have caused this in 360 version? My theory is this is part of the 'painter' rendering style. You displace some pixels and add an average filter, like a median (not as dumb as a straight median though) and it softens the content. I'm pretty certain it's a post-processing difference. HDR depth etc. wouldn't produce dithering like that.

Actually, this is not the first time I have noticed such difference, the Burnout Paradise had 360 version to dither a bit especially for distant object like trees.This could be a case of differences in how the averaging effect is selectively applied between platform. Dunno what hardware differences could account for it though.

[Bellic]

Both versions had different teams and it looks like they made different decisions.

[function]

Well Rockstar co-founder Sam Houser himself wrote "The 360 games have a certain look to them; PS3 games have a certain look to them. I like the way [the PS3] renders. There's a certain kind of softness without being blurry -- some warmth to it -- and then there's a certain more clinical element to how the 360 looks."

They obviously are well aware of the visual differences. Those differences exist to taylor to the strengths of each machine. The actual artists and coders are probably smiling with glee at how much attention individual pixels of their work are getting right now.

It's crazy isn't it? When there are no important differences to argue about any more the fanboys will manufacture them! Discussing the differences is interesting. Trying to spin every difference in your "teams" favour is nuts.

[Edit - Oh no, got confused with the GTA4 console game thread for a second!]

[Shifty Geezer]

It even has commentaries about how the thread's going to the dogs. Which is extremely unwelcome. This is a serious thread for serious people, not for griping about people biases and whatnot, nor writing editorials about the Fanboy Warz. I might go through and prune back to pure upscaling and related discussion when I have a moment, but certainly any OT posts after this one can expect a short life.

[AbbA]

So is it correct?

Ps3= 630P@2xAA(FP16?)
360=720P@2xAA(FP10)

[Hazard]

I don't see any proof that the PS3 version is only 620p. I would love to have a link to this fact.

[grandmaster]

I would like to see screenshots in same context

Tomorrow if my copies arrives.

[Dominik D]

So is it correct?

No, read previous page. It's stated that AA cannot be determined from current caps'. FP used is an educated guess too.

[Shifty Geezer]

I don't see any proof that the PS3 version is only 620p. I would love to have a link to this fact.
The facts come from research by our own Quaz51, so you're not going to find lnks anywhere. Quaz51 is the source that everyone else references! By looking at the framegrabs and analysing patterns of intensity around polygon edges, when you know what to look for it's possible to determine AA and upscaling amounts. If you really want to know the science of it, it'll be in one or other of these threads. In fact do a member search for Quaz51 and find his earliest posts, I think. And whether you care to know how it's done or not, the science and methods are accepted here as legitimate. It would be nice to have corroborating investigators, but even without that you can place your faith in the research and its findings.

[Chisholm]

I'm a 36 year old gamer looking for some proof to help me decide which version to buy. I'm not sure why you neede to insult me for asking a honest question. Please learn some manners guy.

As far as I can remember, facts don't come in links. If you don't like Quaz's pixel-counting skills, please start doing the work yourself. Having said that, if you had followed the discussion you would know that resolution is only one of the factors that come into play here.

Cheers - Chisholm

[Hazard]

The facts come from research by our own Quaz51, so you're not going to find lnks anywhere. Quaz51 is the source that everyone else references! By looking at the framegrabs and analysing patterns of intensity around polygon edges, when you know what to look for it's possible to determine AA and upscaling amounts. If you really want to know the science of it, it'll be in one or other of these threads. In fact do a member search for Quaz51 and find his earliest posts, I think. And whether you care to know how it's done or not, the science and methods are accepted here as legitimate. It would be nice to have corroborating investigators, but even without that you can place your faith in the research and its findings.

Thanks for the answer and I aprreciate it.:grin:

[Shifty Geezer]

If you don't like Quaz's pixel-counting skills, please start doing the work yourself. Having said that, if you had followed the discussion you would know that resolution is only one of the factors that come into play here.I think you're reading more into Hazard's post than he wrote, as is often the way with the written word. You can add emphasis in reading his post and end up with...

I don't see any proof that the PS3 version is only 620p. I would love to have a link to this "fact" :roll:

...and some readers probably do read it that way, but I think you'll find it was an honest question from a visitor who hasn't followed months of discussion and doesn't know the ways we employ. An honest answer is that was needed. Remember, not everyone's a steamin' great fanboy trying to get people's heckles up!

[Darkon]

No, read previous page. It's stated that AA cannot be determined from current caps'. FP used is an educated guess too.

From the zoomed in shots few pages back, it does look like Quincunx AA

[Hazard]

I think you're reading more into Hazard's post than he wrote, as is often the way with the written word. You can add emphasis in reading his post and end up with...

I don't see any proof that the PS3 version is only 620p. I would love to have a link to this "fact" :roll:

...and some readers probably do read it that way, but I think you'll find it was an honest question from a visitor who hasn't followed months of discussion and doesn't know the ways we employ. An honest answer is that was needed. Remember, not everyone's a steamin' great fanboy trying to get people's heckles up!

Thank you Geezer and I'm sorry if I started something by not being clearer in my post.

[Shifty Geezer]

Thank you Geezer and I'm sorry if I started something by not being clearer in my post.It's not really your bad. Just a pitfall of written communications where the lack of intonation loses much of the intention from the words, and for whatever reason human nature tends in such cases to read words in the harshest way. Learning to speak softly on the internet is part of being a useful community member, as is learning to read without mis/pre-conceptions.

This is OT but as I can't PM you, I'll say it here publicly as a reminder for everyone, and clean up later.

NOTE : Moved a couple of posts on which version to get into the console game thread : http://forum.beyond3d.com/showthread.php?t=47802

[ex0du5]

What's up with this shot? (from a GAF poster's PS3 version)

http://xs226.xs.to/xs226/08181/img_0395408.jpg

Looks like there's no anti-aliasing being applied at all in some spots. The heck?

[ex0du5]

Crap...need an edit button (am I missing it?). There's definitely 2x AA that I can spot (top of the car's window is the most visible spot to me). Still...doesn't look too good. Maybe it's just a bad choice for a shot. Kind of scared as we haven't really seen any direct feed PS3 stuff yet.

[AlStrong]

Crap...need an edit button (am I missing it?). There's definitely 2x AA that I can spot (top of the car's window is the most visible spot to me). Still...doesn't look too good. Maybe it's just a bad choice for a shot. Kind of scared as we haven't really seen any direct feed PS3 stuff yet.

You'll get editing privileges over time and as you post. :)

As for the AA, 2xMSAA is only does so much with the two sample points, and only then certain geometry angles will receive AA dependent upon the sample positions.

[D3B]

IGN said PS3 is noticeably smoother in jaggies side by side in its head to head.

[keptcopsup]

IGN said PS3 is noticeably smoother in jaggies side by side in its head to head.

Well maybe there's some blurring filter over the whole image, as some of the shots seem to suggest.

'Smoother' isn't always better, if that's the case.

[Chisholm]

I think you're reading more into Hazard's post than he wrote, as is often the way with the written word. You can add emphasis in reading his post and end up with...

I don't see any proof that the PS3 version is only 620p. I would love to have a link to this "fact" :roll:

...and some readers probably do read it that way, but I think you'll find it was an honest question from a visitor who hasn't followed months of discussion and doesn't know the ways we employ. An honest answer is that was needed. Remember, not everyone's a steamin' great fanboy trying to get people's heckles up!

Sorry for that. I clearly overreacted. Back to my reading comprehension lessons, I guess... :-|

Cheers - Chisholm

[deepbrown]

The facts come from research by our own Quaz51, so you're not going to find lnks anywhere. Quaz51 is the source that everyone else references! By looking at the framegrabs and analysing patterns of intensity around polygon edges, when you know what to look for it's possible to determine AA and upscaling amounts. If you really want to know the science of it, it'll be in one or other of these threads. In fact do a member search for Quaz51 and find his earliest posts, I think. And whether you care to know how it's done or not, the science and methods are accepted here as legitimate. It would be nice to have corroborating investigators, but even without that you can place your faith in the research and its findings.

Yes and Quaz doesnt have a good PS3 screenshot yet.

[ex0du5]

So what's up with the dithering exactly? I mean, we're not 100% sure if it's evident in the PS3 version or not, but we are sure it is in the 360 version. Is this an art-style choice perhaps, or is it more likely a side effect to their rendering/post-processing methods?

[Neb]

I wonder if new HDTVs only have sharpen filters or they can do blurs as well...?

Most if not all HDTV's with a sharpen slider lets you set it to a negative value, blurring the image. I actually did this with a SD TV of HQ for some xbox games to hide the jaggies ( lol )!

[ex0du5]

Most if not all HDTV's with a sharpen slider lets you set it to a negative value, blurring the image. I actually did this with a SD TV of HQ for some xbox games to hide the jaggies ( lol )!

The problem is when the TV's have a value, let's say 0 to 30, and the correct value is 15 (whereas 0 would blur the image). It's annoying to have to use a test pattern to determine the correct neutral value.

[-tkf-]

Its worth considering that this generation of Consoles are so close to each other on graphics that the actual reviews have to take great care on the calibration of their tv sets and the settings on the consoles themselves..

Not to mention viewing these 2 games on a small screen at a certain distance will reveal little. Viewing it on a 110 inch screen from 3 meters will be a completely different story.. which i intend to tell from the PS3 perspective :-)

Welcome to the (sub) HD era :-)

[D3B]

Well maybe there's some blurring filter over the whole image, as some of the shots seem to suggest.

'Smoother' isn't always better, if that's the case.

Whatever the effect it definitely sounded like it was a plus.

[ex0du5]

Its worth considering that this generation of Consoles are so close to each other on graphics that the actual reviews have to take great care on the calibration of their tv sets and the settings on the consoles themselves..


That's why I always *facepalm* when I hear "The 360 version seems to have richer and more vibrant colors" for every single game out there.

[Gholbine]

Since when is 630p an allowable resolution on the PS3? As far as I was aware, it's very strict with the native resolutions they allow.

And how is it upscaling? It can't be line doubling, because it's got to scale the horizontal resolution too. What, suddenly the machine can upscale in hardware? Is it a software solution?

Which media are we using to make this assessment?

[Laa-Yosh]

Plenty of games use irregular resolutions on the PS3, from Jericho to Call of Duty 4... GTA4 is not the first and probably not the last either.

[Dominik D]

@Gholbine: Could you please move back to page 21? Most of your answers ("native resolution" one aside) are there. Technique used and some info about Quaz's technique can be extracted from his posts (this was pointed out on previous page too).

[Shifty Geezer]

Since when is 630p an allowable resolution on the PS3? As far as I was aware, it's very strict with the native resolutions they allow.There are no, and have never been any, restrictions on framebuffer size AFAIK. I've certainly never, ever seen any document that specifies a minimum limit for framebuffer, not for any platform. In any generation!

[Extra411]

Sorry, I tried the search function but couldn't find what I was looking for. In this thread http://forum.beyond3d.com/showthread.php?t=46241 it is said that VF5 for 360 is 4xAA? I seem to recall reading it was 2xAA a long time ago in the first version of this thread. Could someone confirm it for me please? Thanks :)

[Statix]

Since when is 630p an allowable resolution on the PS3? As far as I was aware, it's very strict with the native resolutions they allow.

And how is it upscaling? It can't be line doubling, because it's got to scale the horizontal resolution too. What, suddenly the machine can upscale in hardware? Is it a software solution?

Which media are we using to make this assessment?

They probably use a software upscaling solution, just like for COD4 (COD4 is 1024x600).

[Gholbine]

There are no, and have never been any, restrictions on framebuffer size AFAIK. I've certainly never, ever seen any document that specifies a minimum limit for framebuffer, not for any platform. In any generation!

I seem to recall something a while ago, where an update to the PS3 SDK 'allowed' the usage of 960x1080 for pixel doubling. That implied to me that native resolutions were very strict. It's news to me that any PS3 games so far have been sub-720p, so if that's the case then that answers my question.

[Mintmaster]

I seem to recall something a while ago, where an update to the PS3 SDK 'allowed' the usage of 960x1080 for pixel doubling. That implied to me that native resolutions were very strict. It's news to me that any PS3 games so far have been sub-720p, so if that's the case then that answers my question.You can render at any resolution you want and scale it to the the specific PS3 output resolutions yourself. It's much like how a 256x512 texture of a door can be any size on the screen depending on how close you are in the game. There's just no automatic scaling done by the hardware (or maybe the OS) as on the 360.

The slight blur suggests a custom (i.e. using a pixel shader) filter to go from 630p 2xAA (assuming that's correct) to 720p or 1080p, and then Rockstar draws the text/HUD on top of that at native res. You don't get any more information, though. It's just like scaling in Photoshop.

EDIT: Missed your earlier post. Yes, in short it's a software solution, just like it's always been.

[Mintmaster]

Well Rockstar co-founder Sam Houser himself wrote "The 360 games have a certain look to them; PS3 games have a certain look to them. I like the way [the PS3] renders. There's a certain kind of softness without being blurry -- some warmth to it -- and then there's a certain more clinical element to how the 360 looks."

They obviously are well aware of the visual differences. Those differences exist to taylor to the strengths of each machine. The actual artists and coders are probably smiling with glee at how much attention individual pixels of their work are getting right now.You don't see that as the PR spin that it is? :lol:

Man, it's even worse than joker454 painted it. Devs are actually scared of backlash from the 360 version looking better!

[chachi]

You don't see that as the PR spin that it is? :lol:

Man, it's even worse than joker454 painted it. Devs are actually scared of backlash from the 360 version looking better!
I can't tell if you're joking but that's what I thought when I found out the news, they were spinning early because they knew the internets would blow up when the native resolution was found to be less than 720p.

The PS3 version does look good though, and in some shots the lighting looks better. This is the kind of choice that I prefer developers to make rather than just code to the lowest common denominator in order to avoid any controversy.

edit - I mean the choice between better lighting and outright resolution, play to the strength of the respective consoles.

[Mintmaster]

judging by the screen captures provided, 360 version seems suffer terrible dithering .At first I thought it was the same stunt that Lair pulled in using 2xAA and then expanding the picture to 1 sample per pixel, but then you'd see a clear doubling pattern in textured areas. There's no way in hell that FP10 or even 8-bit causes that, as the dithering error an order of magnitude too big.

I guess it's an artistic effect. Maybe it looks better in motion. Do the owners see this while playing?

[Mintmaster]

I can't tell if you're joking but that's what I thought when I found out the news, they were spinning early because they knew the internets would blow up when the native resolution was found to be less than 720p.No, I'm not joking. joker454 is a dev on these boards, and actually seems to specialize in PS3 now, but check his comments:
http://forum.beyond3d.com/showpost.php?p=1153604&postcount=569 (last paragraph)
The PS3 version does look good though, and in some shots the lighting looks better. This is the kind of choice that I prefer developers to make rather than just code to the lowest common denominator in order to avoid any controversy.Is the lighting any different, though? Seems like anyone who claims this is comparing apples to oranges.

Look at this post:
http://forum.beyond3d.com/showpost.php?p=1153699&postcount=506

That lighting "feel" looks very similar to a PS3 screen at the same time of day:
http://img170.imageshack.us/img170/6858/wmplayer200804272348001sz1.png

(I'm sure someone can find a better comparison)

[zaftxx]

Quaz can you test the resolution on a different PS3 image? i guess to double check.

[tia86]

i know the resolution since yesterday and first IGN video grab but i don't sure that is the good moment for say it :)
i think it's not very pertinent in this case, for lot of people sub720p = worst version, it going to generate confusion


Hi.
Are you sure that the source of the PS3 screenshot (grab from a IGN video) is ok for check the native resolution?
I know that it's impossible to record via HDMI in PS3, you must use the videocomposite, and with the rencoding you lose much.

Thank you.

[chachi]

No, I'm not joking. joker454 is a dev on these boards, and actually seems to specialize in PS3 now...
Oh, I knew that. I just meant about them being afraid of the 360 version being seen as better due to a very vocal minority.

Is the lighting any different, though? Seems like anyone who claims this is comparing apples to oranges.
I don't think the PS3 is using FP10 and the 360 likely is. How much of a visible difference that makes, no idea. We probably won't ever see direct framebuffer grabs from the PS3 version but I'm sure we'll get lots of offscreen comparison shots of similar locations/lighting conditions very soon. The shot you posted does have a very similar sky which suggests it is the same lighting condition so if we had a bigger screen we could compare better. I'm not sure this is the proper thread for all the lighting and comparison talk, maybe a mod will clean some of this up and move it to the GTA IV tech one. Or maybe it's fine here, what do I know? :)

[Galduta]

Hi.
Are you sure that the source of the PS3 screenshot (grab from a IGN video) is ok for check the native resolution?
I know that it's impossible to record via HDMI in PS3, you must use the videocomposite, and with the rencoding you lose much.

Thank you.


http://www.eurogamer.net/article.php?article_id=88461

And I believe that this is via HDMI

http://dot50cal.the-horror.com/b3d/uncharted/

[GameBoiye]

Untrue. You can't tell the native resolution of a game based on screenshots from random websites (as that's what he's basing it off of) seeing as the screenshots are outputted from the system at 720P (monitors would not be able to handle an off standard resolution) and who knows how the pictures were taken or what was applied to them before putting them on the internet. Even if the native resolution isn't 720P, you would need to have the game in order to tell what the system is upscaling from by doing multiple tests using yoru own hardware, not looking at screenshots taken by someone else.

[Laa-Yosh]

It's funny how we haven't had so many sceptical new users around here back then when the issue of Halo 3 being sub-720p came up. And Quaz haven't even made a picture saying "GTA4 PS3 is not HD, it's SD" yet...

His methods are reliable and have been proven to be so several times now. After the initial shock, it turned out to be that higher resolution isn't that important, even though it's of course good to have. So please stop questioning him, and if you have doubts, use the search function and read his previous posts with the explanations.

[GameBoiye]

Oh I'm not questioning his skills at all, but I believe his use of screenshots from other people/video game websites to prove his point isn't right. He didn't do that back when he proved halo 3's resolution to my knowledge.

If he had the game and could take screenshots himself and use those to determine the resolution before upscale, it wouldn't be a big deal, but the way he's going about doing it now is kind of shady compared to his past actions.

[V3]

So is it correct?

Ps3= 630P@2xAA(FP16?)
360=720P@2xAA(FP10)

Can RSX even do AA when using FP16 ?

[GameBoiye]

Can RSX even do AA when using FP16 ?

Technically yes, but theoretically, I don't think I've seen it used before.

[Gitaroo]

what kind of HDR were used in Motorstorm, Uncharted, and Warhawk? All NAO32?

[Mintmaster]

Oh I'm not questioning his skills at all, but I believe his use of screenshots from other people/video game websites to prove his point isn't right. He didn't do that back when he proved halo 3's resolution to my knowledge.There's no reason for these shots to show lower resolution. Cropping a 360 pic to make a face is hard because of the HUD, and the pattern that you look for isn't really affected by compression artifacts or bad captures.

Bungie was sneaky and released screenshots that were 720p without AA, suggesting that they were in-game. I don't think Quaz even posted about resolutions before Halo3, and it was precisely because of Bungie's sneakiness that it was big news (well, that and the fact that it was MS's flagship title). If anything, the higher resolution screenshot is the one to suspect, but it seems solidly confirmed that 360 is truly 720p 2xAA.

You can question it if you want, but it's pretty much settled now with a few independent sources having similar results. People with PS3s and expensive HDMI or component capture cards will only confirm what we've seen.

[GameBoiye]

There's no reason for these shots to show lower resolution. Cropping a 360 pic to make a face is hard because of the HUD, and the pattern that you look for isn't really affected by compression artifacts or bad captures.

Bungie was sneaky and released screenshots that were 720p without AA, suggesting that they were in-game. I don't think Quaz even posted about resolutions before Halo3, and it was precisely because of Bungie's sneakiness that it was big news (well, that and the fact that it was MS's flagship title). If anything, the higher resolution screenshot is the one to suspect, but it seems solidly confirmed that 360 is truly 720p 2xAA.

You can question it if you want, but it's pretty much settled now with a few independent sources having similar results. People with PS3s and expensive HDMI or component capture cards will only confirm what we've seen.

I'll still wait till I get more pictures of the actual game. I can do the tests myself. I wouldn't doubt there is some form up upscaling, but I'm not stupid enough to base my findings off of pictures from gaming websites.

What you just said proved my point. Had he checked halo's screenshots before the game was released he would have guessed wrong, and there's a possibility he's guessing wrong here too. I would have rather him waited till after the game was released so that we can all confirm it and there would be no bickering about it. Shame on Quaz for that.

[Mintmaster]

what kind of HDR were used in Motorstorm, Uncharted, and Warhawk? All NAO32?You don't necessarily need HDR formats to get HDR results. Bloom can be applied on any image, but you get a few artifacts, though. In GT5p, for example, I've seen screenshots where a white license plate blooms as much as the sun reflecting off the rear window because the post-process filter can distinguish between white and super-bright-white.

Even when you want more range, you can do much of it by storing an extra value in the alpha channel to indicate overbrightness. It doesn't mix properly for AA, but that only affects a few edges and often bloom eliminates the jaggies there anyway.

HDR is much more about artistry than framebuffer format.

[dabomb665m]

I'll still wait till I get more pictures of the actual game. I can do the tests myself. I wouldn't doubt there is some form up upscaling, but I'm not stupid enough to base my findings off of pictures from gaming websites.

What you just said proved my point. Had he checked halo's screenshots before the game was released he would have guessed wrong, and there's a possibility he's guessing wrong here too. I would have rather him waited till after the game was released so that we can all confirm it and there would be no bickering about it. Shame on Quaz for that.

To be fair, quaz did say he has known for a few days but didn't want to mention it in fear of the shitstorm that would ensue... Seems like it was triggered by the 400% zoom comparisons to show dithering. What's done is done though, if you want more proof I'm sure it'll be brought forth eventually, this is B3D afterall.

Speaking of which, the dithering is quite distracting in the 360 videos... wish there were some HD videos of the PS3 in action (that didn't require subscriptions like IGN's insider). This video: http://www.mediafire.com/?kdkmditng1i posted in the closed GTA4 thread shows it's pretty much everywhere when in action. Look at the fence in the beginning... ugg.

Question: how was the 360's resolution calculated? I'm not an expert on this at all but just looking at the dithered 400% zoom of the 360 version seems like it's hard to count the pixels.

[Mintmaster]

What you just said proved my point. Had he checked halo's screenshots before the game was released he would have guessed wrong, and there's a possibility he's guessing wrong here too. I would have rather him waited till after the game was released so that we can all confirm it and there would be no bickering about it. Shame on Quaz for that.No, it doesn't prove your point, because Halo was the other way around. Screenshots are never rendered at lower resolution. Ever. Even downscaling a 720p image to any size and upscaling it again won't lead you to a 630p conclusion, because the pattern would different.

He is not getting his shots just from websites:
http://forum.beyond3d.com/showpost.php?p=1153685&postcount=501
http://forum.beyond3d.com/showpost.php?p=1153657&postcount=497

For you to bitch at Quaz after all the work he's done for us is utterly classless, especially when he did wait a day or two after he first figured out the 630p before posting here, probably so that he'd be more sure. If you don't like it, get out of the thread and off B3D.

[Statix]

I'll still wait till I get more pictures of the actual game. I can do the tests myself. I wouldn't doubt there is some form up upscaling, but I'm not stupid enough to base my findings off of pictures from gaming websites.

What you just said proved my point. Had he checked halo's screenshots before the game was released he would have guessed wrong, and there's a possibility he's guessing wrong here too. I would have rather him waited till after the game was released so that we can all confirm it and there would be no bickering about it. Shame on Quaz for that.

You're really, really way off base here to criticize Quaz's methodology. The guy has no agenda, and is the most reliable pixel counter there is. He's smart enough to be able to tell what screenshots are legitimate in-game ones and which ones aren't. You are in absolutely no position to question him.

[GameBoiye]

No, it doesn't prove your point, because Halo was the other way around. Screenshots are never rendered at lower resolution. Ever. Even downscaling a 720p image to any size and upscaling it again won't lead you to a 630p conclusion, because the pattern would different.

He is not getting his shots just from websites:
http://forum.beyond3d.com/showpost.php?p=1153685&postcount=501
http://forum.beyond3d.com/showpost.php?p=1153657&postcount=497

For you to bitch at Quaz after all the work he's done for us is utterly classless, especially when he did wait a day or two after he first figured out the 630p before posting here, probably so that he'd be more sure. If you don't like it, get out of the thread and off B3D.

Just wonderying then why he couldn't wait one more do, to get his own copy before releasing "definite" proof is all I'm wondering.

[Mintmaster]

No but they are all from IGN:lol:I didn't know that, but do you think there is some conspiracy where vertigogaming and IGN are both managed to elaborately crop out the HUD, fill in the holes, crop, upscale, put back the HUD, and do all this despite claiming the PS3 looks better in their review?

Just wonderying then why he couldn't wait one more do, to get his own copy before releasing "definite" proof is all I'm wondering.Just wondering? You don't say "shame on you" if you were just wondering.

This is a forum. Someone gave Quaz a screenshot, asked what res it was rendered at, and he kindly answered. What's wrong with that? He's not publishing a friggin scientific paper. Where's your outrage at all these retard gaming sites informing millions of viewers that a blurred, lower-resolution image is "cleaner" and superior in IQ?

[Pete]

Just wonderying then why he couldn't wait one more do, to get his own copy before releasing "definite" proof is all I'm wondering.

We've got like 5 pages of ppl begging him to analyze a PS3 shot and you want him to wait even longer? Why? If you suspect an agenda, it's not enough that he says he both prefers the PS3 version's look and hesitated to post the res he calculated precisely b/c ppl would focus on the raw #s rather than the end result? :???:

I particularly enjoy you calling Quaz stupid. How about shame on you for not showing the maturity in discussing his "guesses" (as you yourself called them, as opposed to definitive conclusion) that he expects? I'm sure he's just gutted that you didn't get your money's worth from his work.

And he wouldn't have "guessed" wrongly about the Halo 3 PR shots, he would've assessed them accurately. That they were bullshots and so not representative of the shipping game wouldn't have been his fault, and your implication that his first stab at it would be his final and that he wouldn't have reassessed actual screengrabs is farce, given his efforts on these last few pages alone. This forum is meant to be mature enough to discuss the unexpected intelligently, without resorting to running in circles waving our arms over our heads.

There's no need for Quaz to spoon-feed us surprising results, nor for him to quintuple-check and source anything he posts for the good of the internets. Most of us would much prefer a forum that assumes intelligence on the part of its participants, and that's the environment we're in. Really, it's amazing that you would call him stupid and irresponsible for not assuming all of us are stupid.

Feel free to voice reasoned skepticism, but impugning Quaz's motives given his history is a bit much (and the only reason I bothered to elaborate on Mint's reply). Asking him to think of the children before he posts anything is unnecessary and insulting. We're (most of us) big kids here.

[Mintmaster]

I dont know 360 looks blurred more to me.

http://www.neogaf.com/forum/showpost.php?p=10963199&postcount=4831You mean the depth of field?

Why don't you look at the whole gallery?
http://www.gamersyde.com/gallery_8316_en.html
There are plenty of sharp lines on the object of focus.

Also, "blurred more" compared to what? Show me the PS3 shots of similar scenes with less blur.

[GameBoiye]

We've got like 5 pages of ppl begging him to analyze a PS3 shot and you want him to wait even longer? Why? If you suspect an agenda, it's not enough that he says he both prefers the PS3 version's look and hesitated to post the res he calculated precisely b/c ppl would focus on the raw #s rather than the end result? :???:

I particularly enjoy you calling Quaz stupid. How about shame on you for not showing the maturity in discussing his "guesses" (as you yourself called them, as opposed to definitive conclusion) that he expects? I'm sure he's just gutted that you didn't get your money's worth from his work.

And he wouldn't have "guessed" wrongly about the Halo 3 PR shots, he would've assessed them accurately. That they were bullshots and so not representative of the shipping game wouldn't have been his fault, and your implication that his first stab at it would be his final and that he wouldn't have reassessed actual screengrabs is farce, given his efforts on these last few pages alone. This forum is meant to be mature enough to discuss the unexpected intelligently, without resorting to running in circles waving our arms over our heads.

There's no need for Quaz to spoon-feed us surprising results, nor for him to quintuple-check and source anything he posts for the good of the internets. Most of us would much prefer a forum that assumes intelligence on the part of its participants, and that's the environment we're in. Really, it's amazing that you would call him stupid and irresponsible for not assuming all of us are stupid.

Feel free to voice reasoned skepticism, but impugning Quaz's motives given his history is a bit much (and the only reason I bothered to elaborate on Mint's reply). Asking him to think of the children before he posts anything is unnecessary and insulting. We're (most of us) big kids here.

"Based on these screenshots the game looks like it runs at 630P" is all it would have taken. To definitively state that the game runs definitely at 630P without having concret pictures is what bugged me. Who knows, those pics could have been from beta version of the game that the companies got months ago. R* did say they were taking extra time to work on the PS3 version.

I have a lot of respect for him for his skills, which are greater than mine, yet all he did by posting early (by 12 hours, maybe less) is incite flame bate all over the web, with multiple websites pointing to this forum as a definitive source that the game does indeed run at 630P.

Maybe it's the way he worded it that just bugged me. Yes it's most likely true (though I counted 620P on those screenshots, but I'll admit my estimations aren't as good as his) but doesn't mean he should post it as truth without seeing a final pic of the final release of the game.

I've been a long time reader of these forums, and for some reason nothing seemed to warant me posting till this.

[dabomb665m]

"Based on these screenshots the game looks like it runs at 630P" is all it would have taken. To definitively state that the game runs definitely at 630P without having concret pictures is what bugged me. Who knows, those pics could have been from beta version of the game that the companies got months ago. R* did say they were taking extra time to work on the PS3 version.

I have a lot of respect for him for his skills, which are greater than mine, yet all he did by posting early (by 12 hours, maybe less) is incite flame bate all over the web, with multiple websites pointing to this forum as a definitive source that the game does indeed run at 630P.

Maybe it's the way he worded it that just bugged me. Yes it's most likely true (though I counted 620P on those screenshots, but I'll admit my estimations aren't as good as his) but doesn't mean he should post it as truth without seeing a final pic of the final release of the game.

I've been a long time reader of these forums, and for some reason nothing seemed to warant me posting till this.


I'm a long time lurker as well... and your opinion on the matter is quite alright... even to the ones criticizing you. I think what really triggered them was your abrasive attitude on the matter directed towards quaz.

I think what we all want is just a small apology to quaz so we can continue to discuss this maturely as you did in this last post. Afterall, we are all entitled to disagree but making personal attacks aren't warranted.

[Mintmaster]

To be fair to the "retard gaming sites" they've been waiting so long to say that the PS3 has proven itself (and validate their E3 2005 monkey dancing) that they can almost be forgiven for getting ahead of themselves.*

(*The PS3 has already proven itself, but I guess that doesn't matter when you're waiting to call a clear winner in a major head to head contest.)I totally agree, and only Sony could have pulled it off. I'm glad they did even though they probably could have wiped XBox off the map if they skipped BR and launched a year earlier.

Nonetheless, it's amusing to see people - even professional, supposedly neutral reviewers - trying so hard to see "proof" that the PS3 is more powerful than the 360. Here we have a game that's so "obviously" built to take advantage of PS3's strengths. :lol:

I'd be quite interested in knowing what resolution these games run at on a SD television. If people can't tell the difference between 720p and 630p (or thereabouts), I wonder how they'd feel about the difference between 630p (or thereabouts) and a PAL-esque 1024 x 576? Or what about 853 x 480 but with 4XAA and a high level of aniso (and maybe a bit of extra blurring for kicks)?Calling 1024x576 SD-esque is a bit of a stretch. It's almost twice the resolution of 640x480, and many games ran at even lower resolution last gen.

I do agree, however, that 4xAA and 16xAF at SD res is nothing to sneeze at for image quality, particularly when compared to 720p w/o AA or AF at reduced framerate (not that this is the case for GTA, of course). I've long been a proponent of rendering at lower res - say 30% fewer pixels - to have those two enabled. AF adds so much detail, and on top of reducing aliasing AA also makes scaling better.

Anyway, it looks like reviewers could indeed see a difference between 720p and 630p. Some just thought that because this was the game that would make the PS3 shine, that version must have been the better, cleaner, intended look.

Personally, I don't think that 630p is clearly worse than 720p. Suggesting that 630p has cleaner lines and holds the visual edge, though, is laughable. Hilary Goldstein is such a tool.

[GKchikan]

Curious is, when the gaming sites including IGN which has previously claimed the PS3 version had a better resolution start to call it as the 'lesser' one. Of course Quaz still may need more screen shots, but...

[dabomb665m]

Personally, I don't think that 630p is clearly worse than 720p. Suggesting that 630p has cleaner lines and holds the visual edge, though, is laughable. Hilary Goldstein is such a tool.

I'm sorry if I'm reading this wrong but... suggesting that 630p has cleaner lines and holds the visual edge [over 720p] is only laughable with all things being equal right?

We still don't know all the details at work in the PS3 build though, even quaz wants to see certain lighting conditions from the same scene compared on the two systems to see where that "better" (or more favorable to many) look comes from. Personally I don't buy the blurring from upscaling to be the source of this bias... but maybe I'm giving the reviewers/masses too much credit.

I'm 3 hours away from picking up either copy (both pre-ordered) and I STILL can't make up my mind.

[-tkf-]

Maybe kill all the posts here that doesn´t have anything to do with Native Resolution?

There is a "what gta4 will you buy" thread and a "technical blabla GTA4 engine" thread that is usefull. For insults and cheap shots go read the neogaf thread of doom.

I hope i can provide some photos/screenshots in a few hours of the GTA4 PS3 version so that Quaz and squash any doubts about his work.

[Kinan]

there are already good quality screens posted by dot50cal on neogaf:

http://i26.tinypic.com/mkahc9.jpg

http://i27.tinypic.com/11kfcar.jpg

[dabomb665m]

Maybe kill all the posts here that doesn´t have anything to do with Native Resolution?

There is a "what gta4 will you buy" thread and a "technical blabla GTA4 engine" thread that is usefull. For insults and cheap shots go read the neogaf thread of doom.

I hope i can provide some photos/screenshots in a few hours of the GTA4 PS3 version so that Quaz and squash any doubts about his work.

Yeah my bad. How do you guys usually do the screen shots? I have an 8mp canon quickshot type camera and just got my ps3 version. Anyone have any tips?

[Shifty Geezer]

This thread is being temporarily closed until I or some other mod can sort it out. There's a lot of OT banter that's taking away from this thread as a reference point for upscaling etc. It's needs to be managed better, and the beginnings of that is to stop it growing more unwieldy especially with interested new visitors!

[StefanS]

Ok, I did my best to clean up the mess left behind by some posters and I am reopening the thread with the following terms and rules:

RULES:

A.) To the persons who think they can roam around calling people idiots and liars etc.: We won't tolerate this behaviour and we will very gladly accomodate your wishes for a break from this forums, if you continue asking for it.

Let's me make one thing very clear: This is now a zero tolerance thread and we, mods, will act accordingly.


B) This is a thread about the visual IQ of games, like resolution, AA, texture filtering.
it's not the appropriate place
* to discuss reviews
* to ask for advice on which version to get, etc.
* to voice emotional appeals

Such posts will be removed. Continued derailing or derailing attempts will also result in stricter mod actions.




That said: I'd like to take the opportunity to make a personal statement here:

I find it very disturbing and fairly ungrateful, how some posters - a few of them even being long time posters - have behaved in this thread here. Quaz51 is providing this service here for free and has done so over a considerable amount of time. He has been absolutely reliable and has never shown anything that portrays a bias. In fact, he has been addressing the issues of his analysis regarding the PS3 version of GTAIV in his first posts.
Some people definitely owe him an apology for trying to misconstrue his posts as biased.

[Heinrich4]

it's 630p (1120x630) :oops:
regardless PS3 media look better for me, the light and atmosphere is more real (maybe HDR vs X360 MDR, post-process...), i agree with global reviews feeling but more comparison screenshots would be good
it's interesting, it shown that number of pixel isn't enough for determine better visual

Quaz51 thanz a lot for patience,excelent and precision information.

(at ps3 version maybe post processing filters and method AA can determinate a sensation of better graphics despite little less resolution and this is most important at all)

[GAZOman]

I think it would be nice if someone would just add all the info into one post because the thing gets kind of confusing.

well, I have both versions at home - its GTA so I would like to have the best version anyway.

I don't need to count pixels to tell you that neither of the versions is native 720p - 360 is NOT 720p and has almost as much aliasing as halo3 did. just compare a real 720p game to the 360 version of GTA4 side by side and you could end all the speculation within a second.

if you ask me, the differences are very minor - I will look into it tonight again. one thing is for sure: the PS3 has not better AA like IGN and other reviews were implying and the framerate in general is horrible in both versions. after the dithering debate I will mostly look into that, the HDR thing and of course a deep comparison of framerates.

PS: this thread is up on eurogamer and n4g
http://www.eurogamer.net/article.php?article_id=134736

[warb]

From those pics and the video comparison on Gamersyde, it looks as though the only real difference is in resolution. The 360 version looks a little sharper.

[GAZOman]

imo there's no difference in resolution at all between the too. these are both no 720p games - look at the aliasing (even if there's 2x) the jaggies are so big that it's impossible that any of the two is real 720p. again ppl may think the 360 has more resolution because of the absence of overscan in a scaled 1080p image on there FullHD's but if you have an "underscan" function on your set like me - its the same.

[Quaz51]

Good news, dot50cal send me very good screenshot (thanks) and finally it's 1152x640p (and no AA) exactly like Halo3, perfect Dark or Dark sector PS3 :)
even though i has clearly identified 7/8 recurrence on many different edge and screenshot but insufficient edge lenght and bad screenshot lead astray me, sorry, it's finally 8/9 and 9/10 recurrence, it's the final result! :D

[Betanumerical]

Good news, dot50cal send me very good screenshot (thanks) and finally it's 1152x640p (and no AA) exactly like Halo3, perfect Dark or Dark sector PS3 :)
even though i has clearly identified 7/8 recurrence on many different edge and screenshot but insufficient edge lenght and bad screenshot lead astray me, sorry, it's finally 8/9 and 9/10 recurrence, it's the final result! :D

No AA?, that doesn't sound right it looked like it had plenty of AA when i played my friends copy today on my screen.

[AbbA]

Good news, dot50cal send me very good screenshot (thanks) and finally it's 1152x640p (and no AA) exactly like Halo3, perfect Dark or Dark sector PS3 :)
even though i has clearly identified 7/8 recurrence on many different edge and screenshot but insufficient edge lenght and bad screenshot lead astray me, sorry, it's finally 8/9 and 9/10 recurrence, it's the final result! :D

Both version or only Ps3?

[GAZOman]

I expected there's no real AA - its just because of the scaling and blur that we think there's AA imo.

confusing - the 360 with its hardware AA and HDR capabilties has no AA and no real HDR!?

[GKchikan]

Maybe they used a good scaling method that virtually works as sorta AA, instead of that horisonal-only built-in scaler?:)

Besides you, more than one review has so far claimed the PS3 version has a better AA.

[GAZOman]

I know but my eyes tell me there is no AA - just look at the rear bumper of a car in any version. there is no 1:1 pixelmapping because it is scaled but not because auf AA used.

[Jazz9]

Quaz, is the 360 version still confirmed to be 720p 2XAA or not?

[Quaz51]

I think it would be nice if someone would just add all the info into one post because the thing gets kind of confusing.

well, I have both versions at home - its GTA so I would like to have the best version anyway.

I don't need to count pixels to tell you that neither of the versions is native 720p - 360 is NOT 720p and has almost as much aliasing as halo3 did. just compare a real 720p game to the 360 version of GTA4 side by side and you could end all the speculation within a second.

if you ask me, the differences are very minor - I will look into it tonight again. one thing is for sure: the PS3 has not better AA like IGN and other reviews were implying and the framerate in general is horrible in both versions. after the dithering debate I will mostly look into that, the HDR thing and of course a deep comparison of framerates.

PS: this thread is up on eurogamer and n4g
http://www.eurogamer.net/article.php?article_id=134736


no, 360 version is clearly 720p AAx2 but good post-process filter on PS3 version do a good job for hide aliasing and imrove visual, and the little more blurry aspect with more hot tint make a more cinematic feeling, probably what some tester like on PS3 version, and with less pop-in, finally every version is good (phew.... it's delicate topic :) )

[Quaz51]

Both version or only Ps3?


sorry, indeed i talk PS3 of course

[Dopefish]

no offence to quaz but imo there's no way that the 360 version is 720p for real... judged by my eyes and not any screenshots from anywhere.

http://imk.cx/screenshots/xbox360/Grand%20Theft%20Auto%20IV/

Here are screenshots. These are taken directly from the frame buffer before any scaling is applied.

This game is 100% 720p on the Xbox 360.

[chris100]

no, 360 version is clearly 720p AAx2 but good post-process filter on PS3 version do a good job for hide aliasing, and the little more blurry aspect with more hot tint make a more cinematic feeling, probably what some tester like on PS3 version, and with less pop-in, finally every version is good (phew.... it's delicate topic :) )


no AA in ps3 version ?

It seems the presentation of PS3 version is just like Sam Houser(r* co-founder) said.

Warm and soft feeling.

[DJ12]

Very odd, I am no expert like Quaz, but I could've sworn I saw AA in the picture inside the bowling alley on the stairs and railings and that is confirmed PS3 and in game.

[Hazard]

Good news, dot50cal send me very good screenshot (thanks) and finally it's 1152x640p (and no AA) exactly like Halo3, perfect Dark or Dark sector PS3 :)
even though i has clearly identified 7/8 recurrence on many different edge and screenshot but insufficient edge lenght and bad screenshot lead astray me, sorry, it's finally 8/9 and 9/10 recurrence, it's the final result! :D


Okay sorry for my ignorance but does that mean it is in fact 630p? Also I bought the PS3 version and the Jaggies are seriously killing the game grphically for me. I have a older 34" widescreen Sony Wega CRT tv.

[Quaz51]

Quaz, the aliasing in the ps3 version is very visible on my LCD tv. The post-processing doesnt hide it at all.

So far im really disappointed with the iq of this game. It's the worst ive seen this gen.

look PS3 game like the Darkness or MX vs ATV is lot of worst :)
even Halo3 aliasing is logically more visible (same resolution but no filter)

[Butta]

For all those that are getting bad aliasing try turning off 1080i on your PS3. It seems that it may produce some lesser results according to a few people in other forums. It would be interesting to see a screencap from a 1080i setting and a 720 just to find out what is going on.

[Dopefish]

http://imk.cx/screenshots/xbox360/Grand%20Theft%20Auto%20IV/

Here are screenshots. These are taken directly from the frame buffer before any scaling is applied.

This game is 100% 720p on the Xbox 360.

I also want to point out that other than converting these images from BMP to PNG, I have also not applied any kind of post-processing to the images.

Quaz, if you would like any specific screenshots (for comparison) from the 360 version, just let me know and I'll try to capture them.

[Quaz51]

Very odd, I am no expert like Quaz, but I could've sworn I saw AA in the picture inside the bowling alley on the stairs and railings and that is confirmed PS3 and in game.

this?
http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png

no AA, just filter and upscale

[GAZOman]

imo aliasing is not worse on the PS3 than on 360 but its very disturbing in both games. and if you have no underscan feature you will see more aliasing on the PS3 because you loose resolution in the not shown part that is cut from the overscan.

[GAZOman]

this?
http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png

no AA, just filter and upscale


look at the stair steps - no AA, just scaling softness.

[Chisholm]

Okay sorry for my ignorance but does that mean it is in fact 630p? Also I bought the PS3 version and the Jaggies are seriously killing the game grphically for me. I have a older 34" widescreen Sony Wega CRT tv.

640p.

[Hazard]

640p.

Thank you very much. :grin:

[djskribbles]

So i went ahead and bought the PS3 version anyway and it has the worst jaggies so far of this generation. Even worse than Halo 3, imo.

Havent seen the 360 version yet.
imo aliasing is not worse on the PS3 than on 360 but its very disturbing in both games. and if you have no underscan feature you will see more aliasing on the PS3 because you loose resolution in the not shown part that is cut from the overscan.

yup, aliasing on both. IGN even says the anti-aliasing is better on the ps3, while other sites say the ps3 version looks cleaner. maybe the post processing effects paired with the slightly lower resolution and scaled, hides the aliasing better than the AAx2 on the 360 version.

[Neb]

No AA?, that doesn't sound right it looked like it had plenty of AA when i played my friends copy today on my screen.

I saw some screenshots where the aliasing was very clear and it sure didn't look like any AA was in use. Instead it looked like a blur filter.

[warb]

I wonder what kind of performance improvement we'd see with GTAIV at 1152x640 with no AA (or tiling) on 360.

[GAZOman]

remember - gears of war, UT3 engine - no AA - but not that much jaggies due to some clever filters applied.

[Juan Panson]

I wonder what kind of performance improvement we'd see with GTAIV at 1152x640 with no AA (or tiling) on 360.

4xaa would have been nice.

[Neb]

remember - gears of war, UT3 engine - no AA - but not that much jaggies due to some clever filters applied.

Reminds me of Crysis and how I in many situations find the use of edge aa (also blurs edges) to give a softer look and harder to find jaggies than 2xAA in many situations.

EDIT: Though the edge AA in Crysis is the best I've seen in any game so far so maybe it is not a 'good to use example'.

[Neb]

4xaa would have been nice.

That makes me wonder if any of the version is using anisotropic filtering and if so to what level.

[catisfit]

I just counted these myself and got the same findings as Quaz:

The opposite kerb in this 360 image (http://imk.cx/screenshots/xbox360/Grand%20Theft%20Auto%20IV/gta4_02.png) gives 720p

The neon bar above the bowling lane, and the bottom edge of the "Memory Lanes" sign in this PS3 image (http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png) give 640p.

[AlStrong]

Alrighty, there seems to be some confusion as to what is considered to be technical and non-technical.

Technical: measurable specifications such as resolution, anti-aliasing, AF, frame rate, algorithms, shadowing, lighting etc

Technical opinion & questions: e.g. supposition of choices by the game developers with the intent of understanding the technical details - afterall, only they know why they chose to do something a certain way or chose a spec. It's a privilege to get opinions from resident developers to help understand what is being shown. Could be a good choice, could be a bad choice, but leave the dramatic stuff out.

non-technical opinion - judging games with some emotional attachment - "worst something ever" That's just fluff exaggeration.

I hope this helps some of you.

[AlStrong]

4xaa would have been nice.

This is something that's been floating around in my head because they've clearly got tiling working, and 4xAA would have meant only one extra tile.

At the same time, I wonder if they are using deferred shading or lighting or shadowing. Particularly for a large city with lots of objects or potential light sources... it should help. Crackdown used a deferred lighting scheme for instance.

[Hazuki Ryu]

AlStrong when is the list of rendering resolutions going to be updated? Looks like there's so much missing there. :)

[Quaz51]

indeed the eDRAM size (and ROPs) are more MSAAx4 compliant but still rarely use even now

[Hazuki Ryu]

Don't think this will help a lot but its a comparison so i guess it will tell us something about the final results though its not nearly as good as the real thing. http://www.gamersyde.com/news_6399_en.html

[Dr. Nick]

Since the PS3 version doesn't use AA, it is possible that they are using FP16 for HDR, right? Also could they have used Quantix(I'm not saying they did I'm asking could they have.) and been able to use FP16 at the same time?

[AlStrong]

indeed the eDRAM size (and ROPs) are more MSAAx4 compliant but still rarely use even now

hm....yeah. I also wouldn't suspect geometry re-processing issues... Maybe it's an issue with multiple render targets and how they're handled on the 360. i.e. AA'd render targets and so it's not just one extra tile, but an extra tile per render target. (?)



AlStrong when is the list of rendering resolutions going to be updated? Looks like there's so much missing there. :)

What's missing in particular :?: I've been updating things as the info arrives.

[AlStrong]

(To Wabe Walker: I've moved your post to Site Feedback as it was more appropriate. :) )

Since the PS3 version doesn't use AA, it is possible that they are using FP16 for HDR, right? Also could they have used Quantix(I'm not saying they did I'm asking could they have.) and been able to use FP16 at the same time?

IIRC, FP16 flat-out negates the use of any hardware AA mode for G7x. They would have to do some other method to achieve polygon AA, such as NAO32, but this method requires a bit of hacking to do alpha blending.

[Hazuki Ryu]

What's missing in particular :?: I've been updating things as the info arrives.

Uncharted for example I think it was tested here wasn't it?

Edit: I just noticed it has a lot of new stuff, its my brain that needs an update sorry

[AlStrong]

Uncharted for example I think it was tested here wasn't it?

Edit: I just noticed it has a lot of new stuff, its my brain that needs an update sorry


No worries. :)

[Asher]

Uncharted for example I think it was tested here wasn't it?

Edit: I just noticed it has a lot of new stuff, its my brain that needs an update sorry
It might be best to just store this info in a public google spreadsheet.

I can help set that up if needed...

[Quaz51]

if you want all 720p game in the list i can make a list with 300 games that i checked :D but it's not pertinent, consider that all games that aren't in the list are 720p

[Lairliar]

http://images.gamersyde.com/gallery/public/8319/1077_0014.jpg
http://images.gamersyde.com/gallery/public/8319/1077_0009.jpg
http://www.gamersyde.com/gallery_8319_en.html

Dose this help?

Oh,and if you are looking for better screenshot,why not check back this side later.
So there won't be any missing detail or transition problem.
http://ruliwebimg.empas.com/data/news8/10m/20/ps3/ruliweb_gt5_11.jpg
http://ruliwebimg.empas.com/data/news8/10m/20/ps3/ruliweb_gt5_12.jpg

Mod Note: Just a heads up that we don't allow hotlinking media from other sites unless they are hosted on free sites such as image-shack or photobucket. The media sites rely on advertising hits to a degree, and it would be unfair to use up their bandwidth as well.

-AlS

[Scott_Arm]

if you want all 720p game in the list i can make a list with 300 games that i checked :D but it's not pertinent, consider that all games that aren't in the list are 720p

Wow ... I can't say I've ever heard of someone that enjoys counting pixels as much as you do.

[Dopefish]

I just counted these myself and got the same findings as Quaz:

The opposite kerb in this 360 image (http://imk.cx/screenshots/xbox360/Grand%20Theft%20Auto%20IV/gta4_02.png) gives 720p

The neon bar above the bowling lane, and the bottom edge of the "Memory Lanes" sign in this PS3 image (http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png) give 640p.

Since you brought that PS3 image up, I decided to try and take as close a shot as possible with the 360 version.

Compare these two again:
PS3 (http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png)
Xbox 360 (http://imk.cx/screenshots/xbox360/Grand%20Theft%20Auto%20IV/compare.png)

[Hazuki Ryu]

This may seem odd but I wonder whats the level of AF shenmue uses on the dreamcast if it even uses any.

I wonder if this screens are good enough to determine anything. Can't find any good quality screens :(

Would someone kindly look at them for me :)?

http://image.com.com/gamespot/images/screenshots/dc/shenmue/shenmue_b2_screen015.jpg

http://image.com.com/gamespot/images/screenshots/dc/shenmue/shenmue_b2_screen044.jpg

[warb]

if you want all 720p game in the list i can make a list with 300 games that i checked :D but it's not pertinent, consider that all games that aren't in the list are 720p
What about just the 720p games featuring AA?

[GAZOman]

Since you brought that PS3 image up, I decided to try and take as close a shot as possible with the 360 version.

Compare these two again:
PS3 (http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png)
Xbox 360 (http://imk.cx/screenshots/xbox360/Grand%20Theft%20Auto%20IV/compare.png)

this is it! it proves that the 360 has 2xAA, the PS3 none and that the 360 renders in native 720p instead of the 640p from the PS3. (yes, I counted myself)

I was wrong - but the 360 image looks sterile and explains the quote from sam houser about PS3 version looks softer, warmer.

well I would prefer a better framerate and the IQ of the PS3 instead of the choppiness that I have now on my elite.

[Quaz51]

Wow ... I can't say I've ever heard of someone that enjoys counting pixels as much as you do.

it's not counting, just aproach my head to my TV (very close) and with experience in two second i see if it's 720p or not and it's 720p in 90% case or more
and i'm a gamer which download all the demo on Us,Euro,Jap PSN and US, Euro, Jap XBL , couting or not couting i DL all demo ( i check also all comparison screenshots of Richard leedbetter on eurogamer face off and some other screenshot for some case)
and when couting for no 720p game it's just one minute, sometime more in rarely difficult case

[Dopefish]

it's not counting, just aproach my head to my TV (very close) and with experience in two second i see if it's 720p or not and it's 720p in 90% case or more
and i'm a gamer which download all the demo on Us,Euro,Jap PSN and US, Euro, Jap XBL , couting or not couting i DL all demo ( i check also all comparison screenshots of Richard leedbetter on eurogamer face off and some other screenshot for some case)
and when couting for no 720p game it's just one minute, sometime more in rarely difficult case

I made somewhat of a small list for 360 games.

http://imk.cx/screenshots/xbox360/43vs169/

For any image that doesn't have a 4-3 or 16-9 at the end, that game renders in 1280x720 and downscales it to 640x480 (which is why the actual image pulled out of the framebuffer is still 1280x720).

Read the text file at the top for possibly more information.

None of the games I tested render in the framebuffer at 1920x1080. All games upscale to that. Some arcade games (which aren't pictured), though, run in a true 1080p.

[Juan Panson]

Saw this posted at another forum.

Seems PS3 has more aliasing in the distance?

http://i128.photobucket.com/albums/p177/blubbernalgas/vlcsnap-148044.png

[tongue_of_colicab]

Might as well be the angle, the bridge seems to be under a turned angle so the ropes get more in sight rather than more behind each other so it might seem more aliased.

Anyway, I havnt read much of this thread but does all the gta comparisons really matter? the idea that I get is that you will hardly notice any difference if any at all when in motion anyway.

[Quaz51]

I made somewhat of a small list for 360 games.

http://imk.cx/screenshots/xbox360/43vs169/

For any image that doesn't have a 4-3 or 16-9 at the end, that game renders in 1280x720 and downscales it to 640x480 (which is why the actual image pulled out of the framebuffer is still 1280x720).

Read the text file at the top for possibly more information.

None of the games I tested render in the framebuffer at 1920x1080. All games upscale to that. Some arcade games (which aren't pictured), though, run in a true 1080p.


Good, it's direct framebuffer grab, it's confirm all my numbers (Oblivion 1024x600, PG3 1024x600 and most of all COD3 in 1040x624, this case was difficult for me, just some bad screenshot and not conformist sub720p resolution. EDIT: i see in the Alstrong list old COD3 number before my final correction, my personal list is more accuracy ;) )

but framebuffer grab isn't the panacea. some game make rescale in software during post-process (in your list VF5 and COD4 apparently) and in this case final framebuffer is in output resolution not native resolution, counting is necessary :)
interesting...

[Juan Panson]

Might as well be the angle, the bridge seems to be under a turned angle so the ropes get more in sight rather than more behind each other so it might seem more aliased.

Anyway, I havnt read much of this thread but does all the gta comparisons really matter? the idea that I get is that you will hardly notice any difference if any at all when in motion anyway.

Yeah i just watched the gamevideos comparison and you cant even tell the difference aside from that bridge shot. They look pretty much identical.

[Dopefish]

but framebuffer grab isn't the panacea. some game make rescale in software during post-process (in your list VF5 and COD4 apparently) and in this case final framebuffer is output resolution not native resolution, counting is necessary :)

Yeah. These shots won't show any kind of scaling they do before it hits the frame buffer. Those shots you'll just have to inspect yourself. Call of Duty 4 is 1280x720 in the framebuffer, but you can see signs of upscaling. Two games out of the entire list I've made do upscaling beforehand; It's a very small fraction. The average game runs in a true 720p.

[Tap In]

yup, aliasing on both. IGN even says the anti-aliasing is better on the ps3, while other sites say the ps3 version looks cleaner. maybe the post processing effects paired with the slightly lower resolution and scaled, hides the aliasing better than the AAx2 on the 360 version.

yes which is exactly the same plan that Halo 3 used and with the excellent effects (ie lighting) that they gained from it was the right decision IMO.

[AlStrong]

i see in the Alstrong list old COD3 number before my final correction, my personal list is more accuracy ;) )

Yeah, there was some ambiguity in that number I remember. I forgot to ask about it. :p Anything else inconsistent?

[Heinrich4]

Reminds me of Crysis and how I in many situations find the use of edge aa (also blurs edges) to give a softer look and harder to find jaggies than 2xAA in many situations.

EDIT: Though the edge AA in Crysis is the best I've seen in any game so far so maybe it is not a 'good to use example'.

Good point here!

And no one talk or critized Gears of Wars without AA at this time,cause Epic could use some good effect to hide jaguies and same could set here in GTA4 ps3 with your post processing filters and others technics.

(im remember DoomIII at launch in 2004 with many machines cannot run AA and could show good graphics with less jaggies)

[Heinrich4]

Yeah. These shots won't show any kind of scaling they do before it hits the frame buffer. Those shots you'll just have to inspect yourself. Call of Duty 4 is 1280x720 in the framebuffer, but you can see signs of upscaling. Two games out of the entire list I've made do upscaling beforehand; It's a very small fraction. The average game runs in a true 720p.

But COD 4 is does 1024x600 60 fps 2XFSAA in both versions?

[Quaz51]

But COD 4 is does 1024x600 60 fps 2XFSAA in both versions?

yes

[Heinrich4]

yes

Thanx


(we allways say thanx to Quaz51 to find information about resolutions like Halo 3,COD4 etc and now GTA4)

[Quaz51]

Yeah, there was some ambiguity in that number I remember. I forgot to ask about it. :p Anything else inconsistent?

yes
- skate PS3 was an error, lens deformation effect has misleaded my counting, it's just 1280x720 no AA, i'm more experience now, i don't fall in trap ;)
- erase calling all cars in the PS3 list, it is already in the PSN list
- they choose 1040x624 AAx2 for COD3 on X360 because it's the maximum native resolution with basic ratio upscale that can put in the EDRAM (= 9.9Mio) but for the PS3 version they apparently choose a resolution more isomorphic with little more horizontal pixel (1040x624 is anamorphic), i never find good screenshot but my last estimat was ~1088x624

[catisfit]

Since you brought that PS3 image up, I decided to try and take as close a shot as possible with the 360 version.

Compare these two again:
PS3 (http://img294.imageshack.us/img294/4913/wmplayer200804280036497sh8.png)
Xbox 360 (http://imk.cx/screenshots/xbox360/Grand%20Theft%20Auto%20IV/compare.png)

Thanks - confirms 360 at 720p.

[Quaz51]

yes
- skate PS3 was an error, lens deformation effect has misleaded my counting, it's just 1280x720 no AA, i'm more experience now, i don't fall in trap ;)
- erase calling all cars in the PS3 list, it is already in the PSN list
- they choose 1040x624 AAx2 for COD3 on X360 because it's the maximum native resolution with basic ratio upscale that can put in the EDRAM (= 9.9Mio) but for the PS3 version they apparently choose a resolution more isomorphic with little more horizontal pixel (1040x624 is anamorphic), i never find good screenshot but my last estimat was ~1088x624


and..
- Sega tenis isn't 4xAA, it's AAx2 on X360 and Quincunx on PS3
- Lost planet PS3 is temporal AAx2 on the demo
- write "1280x720" for army of two PS3 for avoid confusion

[AlStrong]

Gotcha, thanks :)

[zarbaj]

so...

gta4 360 = 720p 2xaa
gta4 ps3 = 630p no aa?

am i correct so far?

[Arwin]

Holy crap, Eurogamer mentions the 630p and directly links to this thread here (two pages back). :D

[Jesus2006]

I've put GTA4 (PS3) to 1080i mode and counted some pixels.

If i did it correctly i came to 548 lines (plus probably overscan), is that ok?

Here's the shot:

http://img501.imageshack.us/img501/9581/p1000458nq7.jpg

http://img501.imageshack.us/img501/9581/p1000458nq7.th.jpg (http://img501.imageshack.us/my.php?image=p1000458nq7.jpg)

EDIT: Could it be the game's outputting in 640x5xx (PAL?) while in 1080i mode???
At least that's what i counted...but i might need a better screenshot to judge that.

EDIT2: No my wrong it's definately also 1xxx something (horizontal) ;)

[AlStrong]

so...

gta4 360 = 720p 2xaa
gta4 ps3 = 630p no aa?

am i correct so far?
Yes. You can also check out the stickied threads in this forum for information on other games as well.

Holy crap, Eurogamer mentions the 630p and directly links to this thread here (two pages back). :D

They should reference the final confirmed resolution now. :p

I've put GTA4 (PS3) to 1080i mode and counted some pixels.

If i did it correctly i came to 548 lines (plus probably overscan), is that ok?

That's not really a good way to go about it, particularly because of the overscan.

[deepbrown]

Yes. You can also check out the stickied threads in this forum for information on other games as well.



They should reference the final confirmed resolution now. :p



That's not really a good way to go about it, particularly because of the overscan.

I thought it was 640p not 630p

[AlStrong]

I thought it was 640p not 630p It is. Eurogamer is apparently referencing 630 instead.

[assen]

But COD 4 is does 1024x600 60 fps 2XFSAA in both versions?

Neither of the versions is locked 60 fps. I've heard somebody trustworthy quote "a bit below 60 fps on the 360, around 40 on the PS3", but you have to trust me trust him :-)

[-tkf-]

How do you scale "1152x640p" to 1280x720p on the PS3? Run it through a SPU? the infamous "scale" chip? Or another trick i missed?

It has to cost something, why not just render at 720P and take the framerate hit?

It would be so cool to have the developers actually explain how and why they ended at 640

[AlStrong]

How do you scale "1152x640p" to 1280x720p on the PS3? Run it through a SPU? the infamous "scale" chip? Or another trick i missed?

It has to cost something, why not just render at 720P and take the framerate hit?

It would be so cool to have the developers actually explain how and why they ended at 640

http://forum.beyond3d.com/showpost.php?p=1154211&postcount=583

It's a relatively inexpensive operation compared to rendering the extra couple hundred K pixels.

[Laa-Yosh]

32" samsung LCD here, 1366*768 native, connected with HDMI.
I can't see any dithering, sitting about 2 meters from the TV.

Aliasing is noticeable, particularly on car bumpers while driving them, though.

[Butta]

This high res video clearly demonstrates that differences between the 2 versions of GTA are negligable and not even worth all the attention they have been getting:

http://www.gamevideos.com/video/watch?video=18625&largeFormat=mov

Overall Rockstar has managed a very close port.

[Nate Finch]

Obviously total newbie here, but I think I know what all the confusion stems from with the screen shots vs gameplay differences everyone's seeing.

When reviewers and gamers look at 640p on their 720p or 1080p TV, the TV and/or the PS3 are doing scaling from the native resolution of the picture to the native resolution of the TV/PS3. Scaling induces a reduction of aliasing by its very nature, which could be why reviewers have said the PS3 version looks softer and less aliased, even though the PS3 version has no aliasing and the 360 version does.

The 360 version is likely being viewed at native 720p resolution on many people's screens, so doesn't get the "benefit" of scaling from the TV/console.

However, when you take screenshots from the actual output of the two consoles, they're not getting scaled, so you lose some of the "free" anti-aliasing you get when the picture gets scaled to HD resolutions.

So, screenshots of the PS3 look more aliased than they actually would on an HD TV.

It also explains why some people with CRT TVs might see worse aliasing, because most of them can adjust to display any resolution and therefore do not get aliasing via upscaling.

Just a thought.

-Nate

[GKchikan]

Is it possible?

PS3 630p noAA FP16 (multi-render target)
360 720P 2xAA FP10

[Dominik D]

However, when you take screenshots from the actual output of the two consoles, they're not getting scaled, so you lose some of the "free" anti-aliasing you get when the picture gets scaled to HD resolutions.

The problem is (as it's been stated about 5 posts earlier) upscaling happens before HUD drawing, which is done in 720p, which means that direct grab still has scaled graphics. It all probably boils down to personal preferences.

Is it possible?

PS3 630p noAA FP16 (multi-render target)
360 720P 2xAA FP10
Is it possible you read at least two pages of this thread? It's 640p noAA. As for the render target formats - it's a guess and it will probably stay that way.

[AlStrong]

Is it possible?

PS3 630p noAA FP16 (multi-render target)
360 720P 2xAA FP10


It's possible (by the way, it's 640p ala Halo 3/Perfect Dark Zero, and Dark Sector PS3). It would be a difficult analysis based on screenshots.

RSX should be able to support FP16 with alpha blending, but at the same time it will not do MSAA. On the 360 side, FP10 is the highest precision it supports while maintaining the ability to perform alpha blending.

[Rangers]

So at 720P 2XAA it must be using tiling on 360?

I wonder if some of the tiling issues are solved now, since formerly that was supposedly a performance hit that caused some 360 games to stay <720P to avoid tiling.

And what about the fact most 720P displays are natively 1366X768. So you are automatically getting upscaling on every game there.

Then if you are displaying 720P game on a 1080P native display, of course much more upscaling.

It seems we make such a big deal of rendering resolution, when display scaling is at least as much an issue but unnoticed.

Although you can still see why 720P is better, of course it involves less upscaling to 1366X768.

[JaylumX]

It's possible (by the way, it's 640p ala Halo 3/Perfect Dark Zero, and Dark Sector PS3). It would be a difficult analysis based on screenshots.

RSX should be able to support FP16 with alpha blending, but at the same time it will not do MSAA. On the 360 side, FP10 is the highest precision it supports while maintaining the ability to perform alpha blending.

Just out of curiousity is FP10 the highest it supports with antialiasing or the highest it supports overall (without AA)?

[AlStrong]

Just out of curiousity is FP10 the highest it supports with antialiasing or the highest it supports overall (without AA)?

FP16 is usable with MSAA, but without hardware alpha blending (tricky software support).

[betan]

RSX should be able to support FP16 with alpha blending, but at the same time it will not do MSAA.
Why?
Or are you talking about 720p only?
Laa-Yosh already did a test where he applied a blur filter and partial upscalling to a cropped 360 image and it comes out looking pretty much the same as the PS3 image. The conclusion is that the upscaling on the PS3 image is actually removing the dithering.

Why would anyone conclude upscaling is responsible for getting rid of "dithering" when blur would obviously do the job?
Upscaling would not get rid of any noise more than it removes jaggies.

BTW term dithering implies intentional noise to avoid contouring, I object the term until it's confirmed that this is to combat a color range problem (including shadow aliasing).

[AlStrong]

Why?
Or are you talking about 720p only?


Hardware limitation. G7x could not do MSAA with FP16.

[betan]

Hardware limitation.
G7x could not do MSAA with FP16.
someone here claimed it was a BW issue which was the reason I asked,
Now that I checked Nvidia doc, I know better now. :)

[Gitaroo]

can't developers just use the hdr method that they can achieve anti aliasing on both platform (mainly PS3)? The HDR in HL2 looks really good and I know ppl said its not real HDR but it doesnt look bad at all. Im actually supprised that the 360 version of GTA4 has anti aliasing and running at higher res cuz the game looks just as jaggy as the PS3 version from what I have seen. And I thought they would use temporal AA or something like that for the PS3 version, cause some part looks too smooth to be true compare to many ps3 games that run at a sub 720p resolution with no AA.

[22psi]

Is the 360 version of GTA4 using tiling to get 720p and 2xAA?

[Asher]

Is the 360 version of GTA4 using tiling to get 720p and 2xAA?
It has to, else it won't fit in the 10MB eDRAM.

[22psi]

It has to, else it won't fit in the 10MB eDRAM.

yeah that's what i thought, just wanted to confirm... glad to see a game using tiling

[zarbaj]

Yes. You can also check out the stickied threads in this forum for information on other games as well.

thanks, that list rox.

though i noticed 2 games not there, dead or alive 4 and dead rising, would love to know their res. (both games had demos iirc)

[AlphaWolf]

thanks, that list rox.

though i noticed 2 games not there, dead or alive 4 and dead rising, would love to know their res. (both games had demos iirc)

Chances are if they are not on the list they are 1280x720, Quaz51 checked most of the demos.

[zarbaj]

i see.

that gives me hope for Ninja Gaiden 2 being 720p lol.

thanks again for the list, splendid work by Quaz and AlStrong.

[Galduta]

http://www.videogamer.com/features/article/29-04-2008-389-6.html

http://img518.imageshack.us/img518/1713/vsapril29large4yj9.jpg

http://img153.imageshack.us/img153/8470/vsapril29large3ec7.jpg

[mech]

Anyone know what res Rock Band is using? Checked the other thread and did a search but came up with nothing...

[keptcopsup]

http://img518.imageshack.us/img518/1713/vsapril29large4yj9.jpg

http://img153.imageshack.us/img153/8470/vsapril29large3ec7.jpg

Really poor captures, the 360 shot at least. Take a look at the devkit shots right from the framebuffer here. Even the HUD shows how poor those captures are, the thing doesn't have that much aliasing on my TV by far.

[Galduta]

Are jpg ;) .

[deepbrown]

Which platform was the lead development platform?

[woundingchaney]

Which platform was the lead development platform?

Im thinking that two teams were assigned one for each console though I am not completely sure. Although I would imagine that the 360 team were more familiar with their given hardware. Im not aware of the game having a "lead platform" in the traditional "online" sense of the word today. It could also be that the PS3 version had more manpower and time allocated for its creation.

[deepbrown]

Im thinking that two teams were assigned one for each console though I am not completely sure. Although I would imagine that the 360 team were more familiar with their given hardware. Im not aware of the game having a "lead platform" in the traditional "online" sense of the word today. It could also be that the PS3 version had more manpower and time allocated for its creation.

Ah ok. I'm sure they had Sony tech wizards sent in - it would be wise, especially since some of the best are based in the UK (London/Liverpool/Cambridge). I would be unsurprised if Sony was involved in the 640p deicision - the question being "do you want a dodgy framerate, or do you want a solid framerate at a resolution that most consumers won't notice the difference?"

[GKchikan]

Is it possible you read at least two pages of this thread? It's 640p noAA. As for the render target formats - it's a guess and it will probably stay that way.

I do think it's still possible that you don't pounce on a vertex 10 pixels difference, but maybe there is more to it than 10 pixels? (For 360, that really matters given its eDRAM is just 10MB). Sorry if my inane question gave you a bad day.


It's possible (by the way, it's 640p ala Halo 3/Perfect Dark Zero, and Dark Sector PS3). It would be a difficult analysis based on screenshots.


True, but not a few people feel that the PS3 version has the 'better' rendering output, so I wondered if it could be attributable to FP16 HDR. I know it was a tad bad question though.

As for AA on FP16 by pre-G80 nVidia GPU what I meant was the one Naughty　dog had used (http://www.naughtydog.com/corporate/press/GDC%202008/UnchartedTechGDC2008.pdf). Since GTA4 PS3 seems to have no AA though, it's totally another story.

[liolio]

I do think it's still possible that you don't pounce on a vertex 10 pixels difference, but maybe there is more to it than 10 pixels? (For 360, that really matters given its eDRAM is just 10MB). Sorry if my inane question gave you a bad day.




True, but not a few people feel that the PS3 version has the 'better' rendering output, so I wondered if it could be attributable to FP16 HDR. I know it was a tad bad question though.

As for AA on FP16 by pre-G80 nVidia GPU what I meant was the one Naughty　dog had used (http://www.naughtydog.com/corporate/press/GDC%202008/UnchartedTechGDC2008.pdf). Since GTA4 PS3 seems to have no AA though, it's totally another story.

I'm not that clever but I feel that the choice of resolution has more to due with frame buffer memory print than lack of computational power, clearly I feel like FP 16 is unlikely.

[Dot50Cal]

http://www.videogamer.com/features/article/29-04-2008-389-6.html

http://img518.imageshack.us/img518/1713/vsapril29large4yj9.jpg

http://img153.imageshack.us/img153/8470/vsapril29large3ec7.jpg

This looks like they de-interlaced a progressive image...

[JaylumX]

I find it ironic that an effect that a lot of forum members on numerous boards complained about - excessive bluring caused by blooming and hdr/post-processing - seems to be the more prefered effect today. Anyone who has played Dirt may know what i mean.

EDIT: Another quick question. I don't fully understand the importance of FP16 with alpha blending. Does that relate to calculating HDR within transparant surfaces? Why is the alphablending componet so important and also processor hungry.

[Pugger]

I find it ironic that an effect that a lot of forum members on numerous boards complained about - excessive bluring caused by blooming and hdr/post-processing - seems to be the more prefered effect today. Anyone who has played Dirt may know what i mean.

EDIT: Another quick question. I don't fully understand the importance of FP16 with alpha blending. Does that relate to calculating HDR within transparant surfaces? Why is the alphablending componet so important and also processor hungry.

I have played the 360 version for around 4 hours and got to play the PS3 version for around an hour or two last night. There no doubt to me that the PS3 version is blurry compared to the 360 version. I don't care if its blooming or post processing effects the result is difficult on the eyes. The IQ is just plain poor and it makes it really tiring to play. The 360 version is better it terms of blurring but is by no means good in terms of IQ. Playing the game remainds me of working on a monitor that is slightly out of focus, its jaring. Further although they are different in terms of how they portray an open world none the less Crackdown is similar in scale and is miles beyond GTA4 in terms of IQ.

[AlStrong]

As for AA on FP16 by pre-G80 nVidia GPU what I meant was the one Naughty　dog had used (http://www.naughtydog.com/corporate/press/GDC%202008/UnchartedTechGDC2008.pdf). Since GTA4 PS3 seems to have no AA though, it's totally another story.

From what I can tell (quick passover), that's essentially NAO32 HDR. (logluv HDR). They lose hardware alpha blending, and they have a couple slides on how they implemented AB. The upshot is that they can enable MSAA through a custom resolve.

Evidently, Rockstar didn't go down that route. The process does involve some pixel shader cost though, and considering the reduction in resolution, it may be that they were already too pixel shader bound.

There are other things to consider of course.. it could be an entirely different reason.

EDIT: Another quick question. I don't fully understand the importance of FP16 with alpha blending. Does that relate to calculating HDR within transparent surfaces? Why is the alphablending componet so important and also processor hungry.

http://en.wikipedia.org/wiki/Alpha_compositing

[ex0du5]

Man...although the game is lots of fun, I'm really disappointed at the lack of AA in the PS3 version. The jaggies are quite distracting. Though, not moreso than the texture shimmering going on in the background.

I guess I'll have to wait for the PC version to be fully content.

[AlStrong]

Uno momento. Thread split in-progress.

[Mefisutoferesu]

From what I can tell (quick passover), that's essentially NAO32 HDR. (logluv HDR). They lose hardware alpha blending, and they have a couple slides on how they implemented AB. The upshot is that they can enable MSAA through a custom resolve.

Evidently, Rockstar didn't go down that route. The process does involve some pixel shader cost though, and considering the reduction in resolution, it may be that they were already too pixel shader bound.

There are other things to consider of course.. it could be an entirely different reason.



http://en.wikipedia.org/wiki/Alpha_compositing

Off topic, I know, but why is it that so many 360->PS3 games seem PS bound? I was under the impression PS3 had a fairly reasonable advantage in this area. I vaguely recall something about PS length/width differences on the platforms... is it related?

[XeRo]

http://i27.tinypic.com/4vlg8j.jpg

Hi Quaz, do you know what resolution this is running at? If screenshot doesn't work there is a demo on JPN PSN. The game is called Battle Fantasia. Thanks

[btdvox]

Im still interested to hear from devs how they are doing HDR + AA- I understand the NAO32 that was used in Heavenly sword first i beleive but I wouldnt think devs would use the same method to create this effect on the PS3.

*Snip : derailing comment*

[novcze]

So how rendering works in GTA4 and what is best for GTA4 on full hd tv?

1. 640p render target -> upscale to 720p and HUD rendering -> upscale to 1080p in TV

2. 640p render target -> upscale to 1080p (and HUD rendering?) -> TV

3. 640p render target -> upscale to 720p and HUD rendering -> upscale to 1080p in PS3 -> TV

One of them is of course wrong.

[Quaz51]

http://i27.tinypic.com/4vlg8j.jpg

Hi Quaz, do you know what resolution this is running at? If screenshot doesn't work there is a demo on JPN PSN. The game is called Battle Fantasia. Thanks

the demo is 1280x720

[makaveli87]

here you go

360:
1280*720
2xAA
2XAF
FP10

PS3:
1152*640
0xAA
8xAF
FP16

good day.

[AlStrong]

here you go


Evidence or source please?

[Asher]

What would be the point in using AF on the PS3 version if they're using a blur filter anyway?

[crabbz]

Sorry for the silly question, but what is FP in FP10/FP16? Some googling indicates it is related to HDR but I can't find much else.

[Neb]

What would be the point in using AF on the PS3 version if they're using a blur filter anyway?

I was just thinking about that, barely any improvement vs perfomance impact. And neither one seems to have any AF judging by the screenshots (goes to check again).

[22psi]

here you go

360:
1280*720
2xAA
2XAF
FP10

PS3:
1152*640
0xAA
8xAF
FP16

good day.

where did you get this?

[Dominik D]

Sorry for the silly question, but what is FP in FP10/FP16? Some googling indicates it is related to HDR but I can't find much else.
Floating point lighting calculation precission (more or less). As far as I understand image quality impact this translates to contrast ratios and how "vivid" dynamic range image is. I haven't seen any side-by-side comparisons of the same scene with HDR in both formats and I am to lazy to write one myself, so I trust ppl saying, that there is a noticeable difference.

[liolio]

At some point a mod gently explained me the whole fp10/16 stuff.
If I remember properly, the memory foot print for the framebuffer was the same (ie fp10 just save computational power or it's the other way around).
Given the scale of the game and memory constraint I'm sure that none of the version use fp10/16.

Mintmaster or joker454 (others devs of insightful members opinion are evidently welcome it's just that they are active on this topic) what is your opinion about that?

[Dominik D]

@liolio: memory footprint has to be different: fp16 requires 64bit buffer (16bits per channel) while fp10 uses 32bit one in 10-10-10-2 schema (10 bits per color channel, 2 bits for alpha).
References:
http://www.beyond3d.com/content/articles/4/4
http://forum.beyond3d.com/showthread.php?t=21553

[AlStrong]

At some point a mod gently explained me the whole fp10/16 stuff.

:)


RGBA8, FP10/16 etc are pixel formats dictating the precision and storage for each component.

RGBA8 = 8 bits for Red, 8 bits for Green, 8 bits for Blue, 8 bits for alpha = 32 bits per pixel

FP10 = 10 bits for RGB components, and 2 bits for alpha = 32 bits per pixel

FP16 = 16 bits for each component = 64 bits per pixel

The upside to FP10 is the storage, and slightly higher range than RGBA8. The downside to FP10 is that the data is expanded to FP16 in order to maintain precision, and that affects the rendering speed.

Oh and there are differences in how the data is represented with regards to mantissa and exponent, integers vs floating point.

[grandmaster]

I'm running both versions of GTA here side by side. The texture artifacting on 360 is off-putting, but its impact seems to be limited to faraway items (for the most part). The lack of AA and the lower resolution is immediately apparent on PS3.

There have been reports of higher frame rates on PS3 but I'm not seeing it at all to be honest - both versions seem to be as good/bad as each other based on my initial observations.

[liolio]

thanks guy my memory was wrong so fp10 saves space but have the same computational as fp16 (will ry to remeber properly next time).

[timboritus]

Has anyone noticed the similarities between GTA4 and the game Wreckless: The Yakuza Missions for the original XBOX. GTA 4 really doesn't appear to be much of a leap graphically from what was capable on the original xbox console with this game. Both had similar looks, heavy filters and fx, a ton going on onscreen at once, but also had blurry low-rez looking graphics. GTA 4 is impressive when you step back and see how realistic it looks, (although blurry) but it still amazes me that this was the best they can do for the current gen consoles. When you see what they're able to achieve with other games: COD4, Drake's Fortune, GOW etc. It just really dissapoints me to think that 640p and 720p is really pushing the limits of the consoles.

[Dominik D]

The upside to FP10 is the storage, and slightly higher range than RGBA8. The downside to FP10 is that the data is expanded to FP16 in order to maintain precision, and that affects the rendering speed.

Oh and there are differences in how the data is represented with regards to mantissa and exponent, integers vs floating point.

How would expanding FP10 to FP16 affect performance in a significant way? You expand mantissa with 0s and add/sub something to/from the exponent. That's as cheap as conversion can get. ;-) Obviously The other way round it's a little bit trickier as you have to deal with denormalized FP16 values (as long as internal FP16 math is IEEE 754 compliant) and values out of FP10 range. But I think that the overal overhead of the conversion process is close to the gain from smaller set to read/write. Calculations will obviously take as much as in FP16 case, so it seems that there is memory gain, quality loss and no perf impact really. Or am I wrong and packing/unpacking data is THAT expensive?

[Juan Panson]

the light seem more real, maybe more complex lighting or post-process on PS3?
http://www.vertigogaming.net/PSP/GTA/gtaps31.jpg


Has the lighting been confirmed to be different on PS3? I found these 360 shots which look to be the same time of day as that PS3 screen.

http://i128.photobucket.com/albums/p177/blubbernalgas/untitled.jpg

[patsu]

They look similar in these screens (although the PS3 one looks brighter and warmer). IMHO, the city looks better on the PS3 one. Then again, both versions show jaggies here, but the 360 one should have less.

[AlStrong]

How would expanding FP10 to FP16 affect performance in a significant way?

With filtered FP10 targets/textures the expansion is done in the hardware automatically and follows the FP16 route.

http://forum.beyond3d.com/showpost.php?p=914211&postcount=4

[Mintmaster]

The upside to FP10 is the storage, and slightly higher range than RGBA8. The downside to FP10 is that the data is expanded to FP16 in order to maintain precision, and that affects the rendering speed.I wouldn't say "slightly higher range". It's 8192:1 (with a sign bit too) compared to 256:1 ro RGBA8. Considering that most quality cameras have 8-9 stops of dynamic range, FP10 is plenty for photorealism plus lots of room for overbright if the rest of your renderer is up to the task.

The speed isn't really a downside, either. FP10 renders at full speed for opaque pixels and half speed for blending. FP16 renders at half speed (no blending available). FP10 on Xenos is as fast as RGBA8 on RSX. It's only when you use hardware filtering in a post-process pass (and that isn't always useful anyway) that texturing speed drops to half for that particular part of the renderer.