Wow. The tread is still going on. Why don't people just say that a Radeon 7790 (XB1) cannot keep pace with a Radeon 7870 (PS4). It won't work, no matter how you spin it.
Dat extra 5-10 fps on the 7870: life changing.
Wow. The tread is still going on. Why don't people just say that a Radeon 7790 (XB1) cannot keep pace with a Radeon 7870 (PS4). It won't work, no matter how you spin it.
It does actually. 2x anyAA isn't all that great, by the way. You're unlikely to see these artifacts with the game in motion, these are probly a byproduct of some post processing.
Look! You can't even see the jaggies down the track!as clearly
How much framebuffer would take for 1920x1080 without AA ?
I don't get that relative ESRAM to DDR3 bandwidth is a key factor, that doesn't make sense as to why the XB1 is worse as a system then the 360 and couldn't use tiled deferred rendering. That sounds like the wrong metric to me.
For example, lets say you split your 1080p frame into 4, and went for 1080p30 and you somehow filled all 32MB with each of the 4 tiles, that would be 32*4*30 (*2 as you have to save the current tile from ESRAM to DDR3 and load the next tile into ESRAM from DDR3). That is 7.6GB/s (around 11%).. of your frame time spent copying.. But once in ESRAM, you then have a theoretical 272GB/s (204 + 68) of bandwidth to the system, so the CPU can be doing other stuff while the ESRAM is being used by the GPU.. for the other 89% of the rendering time..
I quoted his post in my posts, and I agree that ESRAM is cited as a problem, but he also uses the statement "If MS makes this transparent in the API" which sounded like it was an API issue rather then hardware, but I also support it could be either.
Images
IDK about you guys but I'd take upscaled 900p with proper AA than this shit
That car is painful to look at when you zoom in like that.
Maybe a ridiculous question from someone stuck in the 90s but can we get a "bittage" comparison of each consoles CPU?? I guess we're in MEGA-bits now right? Or giga-bits...?
I did it for people who know how 1920x1080 looks like. If you compare it to Portal 2 screen i presented you can judge role of artifacts from video. Portal2 screen at 200% looks much better than still from FM5 video.
I did it for people who know how 1920x1080 looks like. If you compare it to Portal 2 screen i presented you can judge role of artifacts from video. Portal2 screen at 200% looks much better than still from FM5 video.
It could if MS is making them rush it out for launch... which they are.If that Forza 5 pic is indicative of what the game actually looks like... goddamn Microsoft is in trouble.
It can't look that bad. Turn 10 is really talented, large and well funded. It can't look like that.
Please use ANY other game than Portal, which has very simple geometric environments and is de facto less impacted by the lack of AA...
Out of curiosity is that console portal or pc?
That is absolutely false. Aliasing is especially visible on straight diagonal lines. Portal 2 is game where there is absolutely shit ton of one-two pixel wide straight diagonal lines.
Apologies if this has been posted already but there is a video in this Kotaku post showing what appears to be the latest Madden game in action on the Xbox one, rather than direct marketing vids,,, doesn't look too hot, although to be fair the vid is 360p ...
http://kotaku.com/espn-says-its-been-using-next-gen-madden-footage-it-l-1445158516
While Portal does have diagonal straight lines, it doesn't have many because the environments are really simplistic, which makes it tolerable to no AA.
Little to no AA is especially bad in the little details, where a lot of edges are involved, making the whole screen a jagged mess, not just a few lines here and there like in Portal.
http://i2.minus.com/iqPboMkXUyFRW.png
Can you find here any straight diagonal lines ?
Can you make me a Crysis 2 screenshot with no AA and understand my point?
If you are talking about distant detail then yes adding more complex models to less square/pixel ratio will mean more noticeable aliasing and more noticeable shimmering as you play.
Gadzooks! That's hideous. Can you do one of Driveclub's cars for comparison's sake?
fuck sakes
Grimløck;86104861 said:Gadzooks! That's hideous. Can you do one of Driveclub's cars for comparison's sake?
Oh God! My eyes!!!!! Please don't do that again?
Good question. Paging Skeff!
From the gamersyde video, exported one frame, cropped the car, 200% resize.
Lighting conditions can make a big impact on aliasing and it does look bit better (definetely some form of AA going on). I am downloading currently gameplay video with full sun same as in forza. That should be representative (compared to Forzavid)
I for one cannot understand developers and their total lack of care for image quality. It's like they are dev'ing on 15 inch screens and expect no one to notice how totally facking awful their game looks!
Thank you! We worked really hard on it.
1) We actually used to use deferred rendering, and one of the reasons for switching was so that MSAA would be more affordable and less error-prone. I'm a firm believer in MSAA, and I think it's required if you want really good image quality. The shader-based techniques like FXAA and MLAA are certainly a lot better than nothing and they make a still screenshot look really nice, but they're fundamentally limited by a lack of sub-pixel information and it really shows in motion. They can still be nice though when used in conjunction with MSAA in order to improve the quality further and to cover up some of the places where MSAA doesn't work as well, and there's also some shader things you can do on modern GPU's that make MSAA more effective overall.
From the gamersyde video, exported one frame, cropped the car, 200% resize.
I for one cannot understand developers and their total lack of care for image quality. It's like they are dev'ing on 15 inch screens and expect no one to notice how totally facking awful their game looks!
Look what we have guyz! Umm, hold on while I power up my PC. I speak as a rabid console fan but this is just too much for me to bear for 500 hundred or 600 hundred bones.
I think it is to early to judge IQ based on Gamersyde video. of FM5. There is a lot more shimmering and aliasing problems than it should have at 1920x1080. My assumtion is compression artifacts are making it look a lot worse than it should.
I also wonder why Forza devs didn't realease any material with OVERCAST weather lighting. Overcast is for driving games like "this looks photo-realistic" paint. No overcast lighting don't need to be dynamic. IBL can do that.
HEre is GT5 gif at spa in overcast weather:
Your right man. GT5 overcast looks incredible
PGR used that a lot. It is one of reasons why it looked so good.
Just watched gamersyde videos of both Forza and Drive Club and DC looks better to me. Of course we won't know until the games are out and we can see actual gameplay in person but the lighting in DC just does it for me. There seems to be a depth Forza doesn't have. Imho
Another contributing factor is that for some reason the environmental shadow mapping is broken.
Lighting conditions can make a big impact on aliasing and it does look bit better (definetely some form of AA going on). I am downloading currently gameplay video with full sun same as in forza. That should be representative (compared to Forzavid)
I don't get that relative ESRAM to DDR3 bandwidth is a key factor, that doesn't make sense as to why the XB1 is worse as a system then the 360 and couldn't use tiled deferred rendering. That sounds like the wrong metric to me.
For example, lets say you split your 1080p frame into 4, and went for 1080p30 and you somehow filled all 32MB with each of the 4 tiles, that would be 32*4*30 (*2 as you have to save the current tile from ESRAM to DDR3 and load the next tile into ESRAM from DDR3). That is 7.6GB/s (around 11%).. of your frame time spent copying.. But once in ESRAM, you then have a theoretical 272GB/s (204 + 68) of bandwidth to the system, so the CPU can be doing other stuff while the ESRAM is being used by the GPU.. for the other 89% of the rendering time..
I quoted his post in my posts, and I agree that ESRAM is cited as a problem, but he also uses the statement "If MS makes this transparent in the API" which sounded like it was an API issue rather then hardware, but I also support it could be either.
Oh God! My eyes!!!!! Please don't do that again?
Good question. Paging Skeff!
Gemüsepizza;86116102 said:
We made a list of confirmed resolutions, but Albert sure as shit wouldn't want to look at it:
Forza: 1080p
Ryse: 900p
KI: 720p
DR3: Dynamic bullshit
Driveclub: 1080p
KZ:SF: 1080p
Knack: 1080p
Infamous: 1080p
the Order: 1920x800 (artistic- 2.40 cinema aspect ratio, retains 1:1 pixel mapping)
ACIV: 1080p
Resogun: 1080p
Flower: 1080p
They are just about the only confirmed resolutions out there.
Is DR3 confirmed dynamic now?We made a list of confirmed resolutions, but Albert sure as shit wouldn't want to look at it:
Forza: 1080p
Ryse: 900p
KI: 720p
DR3: Dynamic bullshit
Driveclub: 1080p
KZ:SF: 1080p
Knack: 1080p
Infamous: 1080p
the Order: 1920x800 (artistic- 2.40 cinema aspect ratio, retains 1:1 pixel mapping)
ACIV: 1080p
Resogun: 1080p
Flower: 1080p
They are just about the only confirmed resolutions out there.
We made a list of confirmed resolutions, but Albert sure as shit wouldn't want to look at it:
Forza: 1080p
Ryse: 900p
KI: 720p
DR3: Dynamic bullshit
Driveclub: 1080p
KZ:SF: 1080p
Knack: 1080p
Infamous: 1080p
the Order: 1920x800 (artistic- 2.40 cinema aspect ratio, retains 1:1 pixel mapping)
ACIV: 1080p
Resogun: 1080p
Flower: 1080p
They are just about the only confirmed resolutions out there.