Digital Foundry Article Technical Discussion Archive [2011]
- Thread starter user542745831
- Start date
-
- Tags
- digital foundry
- Status
ShadowRunner
Veteran
Hey now! I never explicitly said here, but with all the hyperbole being thrown around there's not much technical discussion going on atm.
Anyway so few games...well racing games I mean employing dynamic framebuffers. Any thoughts?
Probably because few games render above 720p so cant afford to drop many pixels from there. at 1280x1080 they can drop resolution quite a bit without such a significant drop in IQ.
Wipeout HD is another racer that that uses a dynamic framebuffer but again its a 1080p game.
I predict that next gen when dynamic framebuffers will be used much more often.
Ive found it impossible to find any other non sony PR 1080p shots of this game. Unusual since there is a demo out that everyone with a PS3 has access to. Dont know if my word means much, but i can assure you the game has very good IQ, whereas the IQ in that pic is indeed horrible. Id hold off on IQ analaysis until we can get some new pics.
This screens are causing some lulz worthy comments here, that's for sure. PS2, N64, what's next... Atari 2600?
In 3D mode it looks very Virtual Boy to me...
Last edited by a moderator:
When most people couldn't tell the difference between the lower res buffer and the higher res buffer, I'm not so sure this will happen. I would imagine sticking with the lower res buffer and applying more effects to the remaining pixels would have a more positive impact for most people. :smile:
Well, at some point you'll need to fix the resolution as an upper bound for memory considerations, so it's actually not a bad idea to have the ability to switch to lower resolution rendering (be it the entire frame or specific effects) as there's undoubtedly going to be the oddball event that sees the framerate dip below the target render time for several frames. Depending on the game, they could even hide the lower res by bombarding the player's screen with *cough* "cinematic" post effects (bloooooom, motion bluuuuuur ).
One of the reasons you don't see it used as much for the entire frame on 360 is because of a quirk with the hardware scaler necessitating a couple frames delay when changing resolution so there are limited situations where it shouldn't interrupt the gameplay experience (cut-scene transitioning with black screens for instance). Although if they're software scaling (for whatever reason), it becomes a different problem.
Devs could actually cheat a bit if they decide to go with a wider aspect ratio than 16:9 for cut-scenes (black bars cinematic presentation++++).
).One of the reasons you don't see it used as much for the entire frame on 360 is because of a quirk with the hardware scaler necessitating a couple frames delay when changing resolution so there are limited situations where it shouldn't interrupt the gameplay experience (cut-scene transitioning with black screens for instance). Although if they're software scaling (for whatever reason), it becomes a different problem.
Devs could actually cheat a bit if they decide to go with a wider aspect ratio than 16:9 for cut-scenes (black bars cinematic presentation++++).
There is a huge difference in terms of edge smoothing....look at the artifacts around some of the geometry in the 360 shots. Is it compression?
The rest of the shots:
http://www.eurogamer.net/gallery.php?game_id=13829&article_id=1347185#anchor
I agree in movement its probably a different story.
The rest of the shots:
http://www.eurogamer.net/gallery.php?game_id=13829&article_id=1347185#anchor
I agree in movement its probably a different story.
Looking at the videos paints a completely different picture. Compared to 4xMSAA, MLAA sometimes produces better results, sometimes worse, but the popping in more distant cars is very evident and annoying. Interestingly I think that the X360's MSAA must have something wrong with it too, it should be a bit cleaner IMHO - could there be an issue with the engine, ie. deferred lighting or so, that interferes with the process?
Anyway, the clear conclusion is that for a title like this, MLAA is definitely not the best solution at all. I wonder how TFU2's DLAA would fare with the content.
Also, the compression artifacts and banding on the headlight lensflares on the X360 destroy the effect completely and utterly, it looks as if there was a piece of transparent paper attached to the front of the car.
Anyway, the clear conclusion is that for a title like this, MLAA is definitely not the best solution at all. I wonder how TFU2's DLAA would fare with the content.
Also, the compression artifacts and banding on the headlight lensflares on the X360 destroy the effect completely and utterly, it looks as if there was a piece of transparent paper attached to the front of the car.
Yeah, if you look at dem legs in this screenshot it looks like they mucked up the MSAA swizzle that happens in edram. A lot of edges in this shot are very messy when they shouldn't be. Wonder how that made it through QA.
Screenshot viewer is an epic fail for me on Opera so I don't really look at all of them...
Anyway, if the 4xMSAA has such an obvious bug in the implementation, then it's safe to assume that a correctly working version would reduce flickering and popping edges significantly, compared to MLAA.
Anyway, if the 4xMSAA has such an obvious bug in the implementation, then it's safe to assume that a correctly working version would reduce flickering and popping edges significantly, compared to MLAA.
TheWretched
Regular
Yeah, the small details in the background really are a problem with MLAA... but actually... I still find it better than only 2xAA or QAA.
I had access a few days ago to look through.
It's really bizarre considering it was ok (afaik) in Shift 1.
Yes... QAA. Where did you see blur filter?
I just click on "view full resolution" then middle click to open in new tab.
It's a curious statement regarding the 360 - if you look at the same image I linked before, you can see just how well the wires are rendered in the background compared to the PS3 version.Or maybe I'm reading this wrong
link to DF Shift 2 comparison
thanks for the discussion on this guys... been wondering for two weeks now if they broke their system or just made a bad decision (no choice based on night racing?) compared to S1
in motion S2 has a lot of edge shimmering (360 on 50" tv) but night looks good
thanks for the discussion on this guys... been wondering for two weeks now if they broke their system or just made a bad decision (no choice based on night racing?) compared to S1
in motion S2 has a lot of edge shimmering (360 on 50" tv) but night looks good
Ah, I think he misremembered or confused it with another racer. The actual Shift comparison (Round 22) mentions QAA (not that the difference between a full screen blur filter or Quincunx averaging is much of a distinction).
i nerini del buio
Newcomer
yup, QAA+blur filter is Blur (awesome pun XD) by Bizarre Creations (what an awesome game it is, if you ask me )
)- Status
Similar threads
- Locked
- Deleted member 11852
- Gaming Technology
- Locked