Image Quality and Framebuffer Speculations for WIP/alpha/beta/E3 games *Read the first post*

[djskribbles]

I've noticed this since DF's last article, the PS4 has considerably better AA/IQ even when considering the resolution differences. SMSRenderTeam have stated that both console versions use EQAA, so it must be the extra TAA pass.

 

[Allandor]

I've noticed this since DF's last article, the PS4 has considerably better AA/IQ even when considering the resolution differences. SMSRenderTeam have stated that both console versions use EQAA, so it must be the extra TAA pass.

Well, I would say it has more blur. It seems to be a little bit more washed out, maybe thanks to that broken "motion blur". I still hope they fix this problem.

 

[AlNom]

Maybe the best place for analysis is in the car viewing?

 

[London Geezer]

Wow. Haven't been on this thread in a little while and sweet baby jesus this looks shite now.

 

[AlNom]

Bad for screenshots, certainly. I'm not sure it's that bad while playing given the conditions.

 

[hesido]

Bad for screenshots, certainly. I'm not sure it's that bad while playing given the conditions.

It surely is bad also in motion, it's worse IQ, even the smallest movement creates a "blur". It's inclusing also meant they've dialed back the proper motion blur to the point that it is not even affecting scenery and roadside objects, which reduces the sense of speed. The ghosting does not necessarily add anything to the sense of speed other than be a distraction.

Just as everything relating to perception, it's effect is different for each individual. In this case most seem not to not notice at all, some don't care, some find it bad, for a very small amount of people it's a deal breaker.

Spoiler: Ghost

In the above image, there's not a single frame that properly displays the color of the beams for the bridge. If this is not butchering image quality, I don't know what is.

 

[AlNom]

MSAA and FSAA were usually very costly back in the day. But again the point is that they stated EQAA is used but then they also talked about temporal AA. Quite confusing, actually.

AMD put out a document a while back for the various EQAA modes. Looks like there's more than I thought originally.
http://developer.amd.com/wordpress/...%20for%20AMD%20HD%206900%20Series%20Cards.pdf

The "8 fragments 4 samples" comment from SMS is a bit weird.

 

[Shortbread]

Apparently many people noticed the weird ghosting (and excessive blurring) of Project Cars, PS4 edition, day one patch rumored.

http://www.thegamescabin.com/projec...ne-patch-and-how-quality-is-fcking-important/

Oddly enough, I think SMS was going for a post processing effect similar to a camera (24/30fps) panning slowly across, trying to capture cars at high speeds - giving the viewers at home that delayed blurring/ghosting effect on their TVs.

Honestly, this effect would be great if it was used within the context of a 80's-90's camera panning slowly across replay scheme... not so much from the actual driver perspective.

 

[Cyan]

AMD put out a document a while back for the various EQAA modes. Looks like there's more than I thought originally.
http://developer.amd.com/wordpress/media/2012/10/EQAA%20Modes%20for%20AMD%20HD%206900%20Series%20Cards.pdf

The "8 fragments 4 samples" comment from SMS is a bit weird.

I read that document a few days ago. The question is whether they use some kind of temporal AA or not. From that document EQAA doesn't have a temporal variant --puzzling.

 

[Globalisateur]

Like I feared before, Project Cars ground textures: PC 16xAF | PS4 2xAF | XB1 4xAF. Typical Benjamin Button case of missing AF (well retrospectively we can now be pretty sure that it has always being a case of low 2xAF level in all previous #PS4noAF games).

It's hard to see the difference between the 4xAF used on XB1, but it's there, the filtering is slightly below XB1 levels, it's definitely under 4xAF so 2xAF. Even DF are clueless:

Ground textures on console make use of what appears a match for PC's 4x anisotropic filtering mode, though PS4 produces blurrier results than Xbox One overall

Because of the blur from the ghosting this has probably being overlooked by the devs, they even really think PS4 uses 4xAF like XB1 as written in this post by one dev! and because of the very small difference between 4x, it will probably stay like this. Also the native resolution of PS4 makes it harder to spot the difference because the XB1 textures are automatically very slightly blurred because of the upscaling.

 

[Shifty Geezer]

Is PS4 sampling weighted towards a lower mip level?

 

[AlNom]

hm... curious.

 

[hesido]

The AF loses its value when you are indiscriminately pasting the previous frame on top of the last one.

 

[Shifty Geezer]

That shouldn't be affecting static scenes.

 

[hesido]

That shouldn't be affecting static scenes.

True, but you are almost always moving in this game, and a single pixel motion ends up in blur. Only in the distance when you are on a long straight it would be beneficial, and only for the lines (not much for the road textures)

 

[AlNom]

Is PS4 sampling weighted towards a lower mip level?

AF seems lower in this comparison, which is puzzling against the developer comments, unless said comments were in reference to a "patched" build.

Cockpit shadows for XO were supposed to have been a bug as well, oddly not fixed in retail.


hm....................

 

[Globalisateur]

The AF loses its value when you are indiscriminately pasting the previous frame on top of the last one.

My second screen is taken while the camera is totally static, before the start of the race. And again, the fact that the XB1 game is upscaled is slightly reducing the 2x / 4x difference. That's why filtering difference here is very small, but it's there.

If you are unsure, compare with the high quality DF images (opened in 2 tabs in order to quickly compare both pics). Be sure to compare the grass on the left of the road...It's even more noticeable somehow there.

http://images.eurogamer.net/2013/articles//a/1/7/5/3/8/2/9/PS4_002.bmp.jpg
http://images.eurogamer.net/2013/articles//a/1/7/5/3/8/2/9/XO_002.bmp.jpg

EDIT: On another note the AA used is really good, particularly on the PS4 version.

 

[hesido]

If you are unsure, compare with the high quality DF images (opened in 2 tabs in order to quickly compare both pics). Be sure to compare the grass on the left of the road...It's even more noticeable somehow there.

I'm not saying there's no difference or not noticable, it's just that you are moving most of the time and even if PS4 was 16x AF, it would probably end up looking blurry in motion. The IQ hit by having a frame blend is worse in motion compared to any AF setting.

edit: Oh sorry! Now I understand what you mean. I think you got me wrong. I wasn't talking about you pasting screenshots I was talking about the frame blending ("pasting" pixels from previous frames on the last one)

edit2:

EDIT: On another note the AA used is really good, particularly on the PS4 version.

Oh no, you didn't!!

 

[djskribbles]

Minus the ghosting, the IQ is pretty good on PS4 IMO. Not everyone is as sensitive to the ghosting it appears... only some people seem to be bothered by it. Maybe it depends on the display type? I notice it in motion, but I'm not really bothered by it.

 

[hesido]

Minus the ghosting, the IQ is pretty good on PS4 IMO. Not everyone is as sensitive to the ghosting it appears... only some people seem to be bothered by it. Maybe it depends on the display type? I notice it in motion, but I'm not really bothered by it.

Well, there was at least one post that is in line with display type having an effect.
http://www.neogaf.com/forum/showpost.php?p=163289398&postcount=703

As for me, I notice it both on my Led and my laptop LCD.