New D3D FSAA Viewer
- Thread starter Colourless
- Start date
dr3amz said:MuFu said:
not doing anything wrong
with 2xT and 4xAA at 1024*768@100hz i get around 34fps average in farcry.
aliasing is working, i have tempthreshold set to 0 (so its always on) and i get NO flicker.
I dunno - but it is working i know how to run it
Ok.
You must be doing it wrong, most of us use size 1 so it's totally unlegible to the naked eye *wink*Chalnoth said:
Joe DeFuria said:digitalwanderer said:
Yeah...maybe a little....though it leaked only "one week" before launch, which isn't too bad.
Not sure if Ati is that pissed about this leak. People were moaning because they didn't see the benches 26th of April. Now those same people have something to play with for the next week and they're quite happy.
Considering the amount of talk on 3D IQ on this forum it is rather surprising to find so many people using LCD displays
I get the feeling a lot of people here prefer to talk about video cards rather than actually using them
This seems a rather nice conjuring trick from all reports, will be interesting to see how benchmarking is done.
I get the feeling a lot of people here prefer to talk about video cards rather than actually using them
This seems a rather nice conjuring trick from all reports, will be interesting to see how benchmarking is done.
dizietsma said:Considering the amount of talk on 3D IQ on this forum it is rather surprising to find so many people using LCD displays
I get the feeling a lot of people here prefer to talk about video cards rather than actually using them
This seems a rather nice conjuring trick from all reports, will be interesting to see how benchmarking is done.
indeed - will 2xT/3xT be paired off against normal AA? as the effect is the same - but better with 2xT/3xT.
I think it has to be - if ATI have revolutionised AntiAliasing again, then this sets a new standard.
ie: x800 does 60fps with 2xT+4aa and 6800multi does 60fps with 2xAA - i know what is better.
dr3amz said:
As you'll be running 1600x1200 at this fps and AA you'll probably not really notice the difference anyhow unless you stand and stare.
Run nature in 2001 at 1600x1200 and 4xAA ( 6800u's sweet spot ) and see the AA quality, now if you run at x8AA you will be hard pushed to see much improvement . Where this might have a big impact is at lower reslutions and here X800 and 6800u are more likely to be cpu limited, hence this feature IMO is a real boon for the r3xx guys.
dr3amz said:
I was just thinking about how my 9800XT performance with temporal AA would compare to the 6800u at comparable settings when I remembered the Hardocp 6800 preview.
http://www.hardocp.com/image.html?image=MTA4MTc0NzQ0ODZxTE1PbWV1dFNfNF81X2wuZ2lm
http://www.hardocp.com/image.html?image=MTA4MTc0NzQ0ODZxTE1PbWV1dFNfNF82X2wuZ2lm
They are benching at 1024 x 768 at full IQ settings in farcry.
6800u is at 4xAA x 8xaf
9800xt is at 2xAA x 8xAF ( add 2xt with no performance hit=4XAA)
-----------------------------------------------------------------------------------
Test 1
-----------MIN FPS-----MAX FPS--- AVG FPS
6800U ------21----------112---------40.7
9800XT------21----------101---------38.6
-----------------------------------------------------------------------------------
Test 2
-----------MIN FPS-----MAX FPS--- AVG FPS
6800U ------22---------52---------33.6
9800XT------24---------44---------34.1
-----------------------------------------------------------------------------------
So with 2t added these cards are being benched equally at equal IQ settings and getting nearly the same FPS!!!!
(Of course the ATI card is still rendering more PS2.0 because of the Farcry 1.1 patch!)
Not exactly a fair comparison.. but food for thought!
hmmmm ......... yummy!
Joe DeFuria
Legend
The problem there, is that 30 FPS average is not going to be fast enough from an absolute sense to enable temporal AA. You're likely to get way to much "flickering" going on.
You're really going to need 60 FPS average at a minimum to get use out of Temporal AA.
You're really going to need 60 FPS average at a minimum to get use out of Temporal AA.
Yah, 2T looks like ass at anything less than 60FPS (and even then, it's possible to see flicker when using 2x2T and sometimes 4x2T). If you pick a scene with just the right complexity level you can "toggle" the flickering on and off just by looking around.
I still can't get the damn TemporalAAFrameThreshold setting to kick in.
I still can't get the damn TemporalAAFrameThreshold setting to kick in.
At 2xAA2T I get almost no noticeable flickering at all. It is barely visible to me in games that have many straight lines at off angles (like UT2004) and in Farcry I don't see it at all (outdoors at least) because there are just so many lines going in all directions.
I must say that I'm very impressed with Temporal AA so far.
I must say that I'm very impressed with Temporal AA so far.
No threshold setting at all.
Yeah I know the numbers will be different because of vsync but from what I'm seeing so far its working pretty well for me.
I'll have to go to some indoor Farcry levels to see if the flickering is more noticeable or not. Outdoor scenes are so "busy" that I really don't see it at all.
Yeah I know the numbers will be different because of vsync but from what I'm seeing so far its working pretty well for me.
I'll have to go to some indoor Farcry levels to see if the flickering is more noticeable or not. Outdoor scenes are so "busy" that I really don't see it at all.
I almost can't stand to look at CRTs anymore after using LCDs for so long. Even at high refresh rates, the best CRTs look dim and shimmery to me. (I used to have a Cornerstone 21" at 100+Hz 1600x1200)
And of course, the amount of space and weight that CRTs take up really annoy me. I banished them from my house.
And of course, the amount of space and weight that CRTs take up really annoy me. I banished them from my house.
Joe DeFuria
Legend
Moose said:
Yeah....it's going to be dangerous...they're going to actually have to use their heads and WATCH a little bit (and heaven forbid, actually sit and play some), instead of just running push-button benchmarks and throwing up graphs.
But then, they've really needed to do that for quite some time now...and it hasn't stopped the majority of 'em from throwing up crap.
You can't really review it since it only works at high framerates (greater than 40 if i'm not mistaken).
digitalwanderer said:
It supports two patterns: ordered grid and rotated grid, which can be switched every other frame for "2T" in the parlance being used here.
But moreover, programmable sampling positions are not required. The effect can also be achieved via pixel center offset.
How do you know other IHVs already didn't implement this along time ago but decided against exposing it because it didn't really improve IQ in most cases, but made it worse?