I don't know... in Home the MLAA work great in some edge, terrible in others aspects, especially in the long draw distance... I mean, the aliasing is almost inexistent (a part in the fine objects) but the shimmering it's horrible in the objects in the long distance... I think the MLAA is perfect for the characters & large objects, but for the fine details it's almost better no AA after all... I'm pretty curious to see 2xmsaa+MLAA, probably the best alternative on the ps3 to the 4xAA... but even 2xAA temporal+ MLAA could be interesting to see in action...
Digital Foundry Article Technical Discussion Archive [2010]
- Thread starter liolio
- Start date
-
- Tags
- digital foundry
- Status
I don't know... in Home the MLAA work great in some edge, terrible in others aspects, especially in the long draw distance... I mean, the aliasing is almost inexistent (a part in the fine objects) but the shimmering it's horrible in the objects in the long distance... I think the MLAA is perfect for the characters & large objects, but for the fine details it's almost better no AA after all... I'm pretty curious to see 2xmsaa+MLAA, probably the best alternative on the ps3 to the 4xAA... but even 2xAA temporal+ MLAA could be interesting to see in action...
Billy Idol
Legend
@grandmaster: are you planning a big DF GT5 tech round up (similar to the famous KZ3 tech round up?)...I would love to read about it (and I guess that I am not the only one)...although I can imagen that GT5 is to much of a mammoth to discuss every aspect.
Me to!Seeing KZ3 example it looks great on some edges while some others dont look quite well and when everything around is smooth and then one thing shimmers or is aliased it kinda draws attention.
Thats why i want to see Cryteks implantation.Postprocessing AA on near objects,temporal on foliage and further objects.
I can see MLAA working fine for Assassin's Creed.
The renaissance era art would mean a lot less thin geometry and specularity (hence less shimmery stuff overall) when compared to a futuristic industrial setting full of metallic architecture as seen in Killzone 3.
The renaissance era art would mean a lot less thin geometry and specularity (hence less shimmery stuff overall) when compared to a futuristic industrial setting full of metallic architecture as seen in Killzone 3.
However I suspect too the Home implementation is light MLAA or something of similar, I can't explain better... even in the large object same almost a 2xAA more a MLAA seen in GOW 3 for example, looking Kratos smoothness compared to the Home avatar...
i nerini del buio
Newcomer
Why? In uncharted 3 there aren't no problemI mean, theorically could be possible or not in AC Brotherhood? And why not in this case?
actually in U2 there is as much input lag as in every other triple buffered games
maybe you're just not enough sensitive to notice?
wat
I think it's the first time to read something of that, forgive me, but I have play enough unchy 2 to say the triple buffering not cause any evident problem of lag ... I have seen even the DF analisys & it seem almost pretty in line with the double buffer games...
Last edited by a moderator:
What I've meant was that platforming in AC2 is different, and I've already had plenty of cases where I've fallen or - in the special side missions - failed tricky jumps*, because of the controls' clumsiness. I don't think that adding another frame of latency would help with that.
* I get so angry at such things, that I had to keep a plastic bottle of water near my seat - just to have something other than the controller to throw around...
* I get so angry at such things, that I had to keep a plastic bottle of water near my seat - just to have something other than the controller to throw around...
I think it would've been fine if the input lag is consistent in the tricky spots you mentioned.
It's no Tekken, but iIt's the most responsive and consistent triple buffered game out there. Consistency being the operative word since most games don't quite achieve that.
Compare it to something like RE5 where the input lag becomes noticeable when the frame rate drops (the public assembly being the prime example), or Castlevania on both platforms. I've noticed that with the latter in more scenarios since I'm more sensitive to lag with fighters and action games since my experience with shooters in general is minimal.
Anyway, here's the NFS Faceoff:
http://www.eurogamer.net/articles/digitalfoundry-nfs-hot-pursuit-face-off
I've got a question regarding audio options. I know file sizes are often comparable, and in some cases like Dead Rising 2, the PS3 version is actually smaller (albeit there is a mandatory install). Is the HDMI 1.2 spec the only limitation that keeps developers from adding multi channel LPCM to 360 games? If so, why don't we see any DTS used? All you need is a Toslink for that. The hardware is licensed to do it (at least for movies)
It's no Tekken, but iIt's the most responsive and consistent triple buffered game out there. Consistency being the operative word since most games don't quite achieve that.
Compare it to something like RE5 where the input lag becomes noticeable when the frame rate drops (the public assembly being the prime example), or Castlevania on both platforms. I've noticed that with the latter in more scenarios since I'm more sensitive to lag with fighters and action games since my experience with shooters in general is minimal.
Anyway, here's the NFS Faceoff:
http://www.eurogamer.net/articles/digitalfoundry-nfs-hot-pursuit-face-off
I've got a question regarding audio options. I know file sizes are often comparable, and in some cases like Dead Rising 2, the PS3 version is actually smaller (albeit there is a mandatory install). Is the HDMI 1.2 spec the only limitation that keeps developers from adding multi channel LPCM to 360 games? If so, why don't we see any DTS used? All you need is a Toslink for that. The hardware is licensed to do it (at least for movies)
Last edited by a moderator:
I've seen the effect of 4x MSAA + MLAA on Half life 2 and the result isn't that much different at all to just MLAA by itself. MLAA would get rid off the blurring of QAA for starters thus allowing clearer textures and that's actually more important over some shimmering here and there. I have never had problems with shimmering but fuzzy textures are a big turn off to me. One of the reason why killzone 3 looks so much better than killzone 2. I just think the pros of MLAA far outweighs its cons in either killzone 3 or AC.
It's a limitation in the Xbox 360 itself that prevents the Xbox 360 from outputing 8 channel LPCM not the HDMI 1.2 spec.
http://www.avsforum.com/avs-vb/showthread.php?p=10177967#post10177967
Do note that the sounds and music are encoded as an lossy format since using uncompressed audio would probably waste memory for basically no benefit.
4xMSAA + MLAA in HL2 tests where borked. Visual result was worse combining both than either alone. This should be ATI driver bug/problem and maybe resolved now. Though each by itself it was clear how 4xMSAA was immensly superior in handling vegetation and finegrain detail. 2xMSAA + MLAA could be a good combination though other examples than Hl2 should be used (which there are and it looks very good). :smile:
Lucid_Dreamer
Veteran
When last I checked, the 360 only bitstreams DTS (like for movies). It doesn't encode or decode DTS. The only exception being the defunct HD-DVD. That's all done in software (in the 3+million lines of code written for HD-DVD to work on the 360). Needless to say, that would take up more resources. The HDMI 1.2 spec doesn't limit multi-channel LPCM from 360 at all.
Ah, thanks for the answers guys.
Isn't that the case with the PS3 as well? Noting my earlier examples of LPCM supporting third party games, the files sizes are the same or even smaller than the 360 game. Or does the method of compression vary? First party games do have noticeably higher quality audio, but I assumed because the files were compressed less.
Isn't that the case with the PS3 as well? Noting my earlier examples of LPCM supporting third party games, the files sizes are the same or even smaller than the 360 game. Or does the method of compression vary? First party games do have noticeably higher quality audio, but I assumed because the files were compressed less.
Billy Idol
Legend
@grandmaster: if you are planning a DF GT5 special I have a suggestion:
I would really like to know how the actual GT5 ingame graphics compare to GT4 photomode?
You know why? Because I think that the GT5 photomode shots look great and sometimes astoundingly real. But I wonder when we will see GT5 photomode ingame...PS4 or PS5? So I guess a good estimation would be to compare GT4 photomode to GT5 ingame and see how much progress has been made...especially since people state that "standard cars" in GT5 are derived from GT4 models?!?
I would really like to know how the actual GT5 ingame graphics compare to GT4 photomode?
You know why? Because I think that the GT5 photomode shots look great and sometimes astoundingly real. But I wonder when we will see GT5 photomode ingame...PS4 or PS5? So I guess a good estimation would be to compare GT4 photomode to GT5 ingame and see how much progress has been made...especially since people state that "standard cars" in GT5 are derived from GT4 models?!?
Well that explains it then, but doing more aa on top of MLAA would be too much trouble for a console though, I'd rather save the precious bandwidth for other effects instead.
@Grandmaster
When You will compare GT 5 with GT 5 please use Fuji as an example, its the track that has changed the most [based of the movies/screens from events].
And if it will be possible it would be nice to see fps analysis of photo mode and photo travel, and car also LOD comparison between replay and photo mode replay.
Oh and there are that same cars in premium and standard model - it would be nice comparison in direct feeds in replay and gameplay ;]
Thx in advance ;]
When You will compare GT 5
with GT 5 please use Fuji as an example, its the track that has changed the most [based of the movies/screens from events].And if it will be possible it would be nice to see fps analysis of photo mode and photo travel, and car also LOD comparison between replay and photo mode replay.
Oh and there are that same cars in premium and standard model - it would be nice comparison in direct feeds in replay and gameplay ;]
Thx in advance ;]
I would like to run double blind tests and see if humans can actually detect even 200+ms of lag. I suspect the answer is no. Much like audiophiles cant really tell the difference between various HQ audio formats and low quality MP3's in such tests.
Doesn't stop me from being a bit obsessed with display lag, but at the same time part of me doubts it makes any difference. If something like Onlive is playable, which is going up to 1000 miles out to a server and back, it seems our paltry local lag is hardly worth quibbling about.
Silent_Buddha
Legend
Depends on the game and player. 200+ ms of lag is definitely very noticeable however. The easiest place to notice this was in online games back when Modems (usually 250 ms best case) were still popular compared to DSL (on average around 100 ms) versus some cable as well as direct T1's (sub 30 ms).
Display + controller lag is also quite evident in PC FPS where there is a direct translation (exact 1:1 or greater corollation) between the movement of your hand and the cursor on the screen, unlike a console controller where, at least for me, there's a bit of disconnect between pushing a stick in a direction and watching the cursor move at varying speeds. Back in the early days of USB mouses, there were people that could immediately notice the 60hz polling rate of USB mice compared to the 120+ hz polling rate of some of the nicer PS/2 mice.
30 fps versus 60 fps also being quite noticeable for twitch players who rely on being able to aquire targets in one frame. People watching me play back then thought they were watching a slideshow (90 degree and 180 degree turns in 1 frame before turning again), but for me it was fluid as I only cared about whatever I could glimpse in that 1 frame before checking other attack vectors.
Certainly not common. And you're right that casual players may not be able to tell definitively whether they are playing at 30 fps or 60 fps, for example. However, 200+ ms difference is incredibly noticeable.
After all, there's only about a 30-50 ms difference between Move and Kinect, but if you go from one to the other (or just look for the lag) it's certainly a noticeable difference. However, if you just start using Kinect and playing its games without actively looking for the lag (for example casuals that don't console game much if at all), it's unlikely that they'll notice much if any lag.
As well control/display lag is obviously more noticeable if you're actually playing a game. So the 30/60 fps example I used above would be pretty indistinguishable to almost everyone if they were only watching it. The distinction for visuals (versus control) is further blurred (pardon the pun) when you start introducing motion blur and other visual tricks that can fool the eyes into thinking the action is much more fluid than it necessarily is. Control however remains infuriating for some even with that. For example Crysis, which looked quite fluid and fast visually (20-30 fps) with motion blur enabled, but felt quite wooden and unresponse (again 20-30 fps) when controlling and aiming.
Regards,
SB
Exactly. Twitch based FPS's and fighters get hit noticeably with any lag, and naturally, casual players don't notice it.
However, with fighters, there is a reason why we don't see any 60fps ones. The genre needs to be as responsive as possible and any decline is immediately noticeable.
- Status
Similar threads
