New D3D FSAA Viewer
- Thread starter Colourless
- Start date
MuFu said:
Perhaps on old games, might be good for Counter-Strike, but current games? Current games are often CPU limited and far from stable with minimum FPS differing greatly from the median. I don't think you'll be running HL2 at a rock solid 90Hz, nor D3, which is hard limited to 60. FC and UT2k4 also seem CPU bottlenecked.
Graphics_Krazy
Newcomer
Actually LCD users should benefit from this since they usually have a ghosting delay.
On mine it looks fine at 60hz at any 2X Temporal setting.
Thats with Painkiller, Ut2k4 etc (farcry works only if I dumb the settings to Low and keep a constant frame rate over 60)
On mine it looks fine at 60hz at any 2X Temporal setting.
Thats with Painkiller, Ut2k4 etc (farcry works only if I dumb the settings to Low and keep a constant frame rate over 60)
{Sniping}Waste
Regular
I need to know were to add the string for Windows 2000.
Graphics_Krazy said:
That's the same for me on this LCD, although you can see a slight flicker with 2x2T sometimes. 6x2T looks excellent.
All 3T modes flicker detectably at 60Hz. This is not nearly as bad as what happens when the FPS drops - it's like somebody shoved a vibrator up your butt (I guess...).
Althornin said:
nevermind think you misread what i was saying though
joe said you would lose quality at 3temp - but i didnt notice that, but performance stayed the same. so i tested with 6xAA to check whether 3xtemp had an impact using the highest AA setting. results proved (for me at least) that using 3xtemp and 6xaa was no different to 2xtemp and 6xAA - in IQ or performance.
just going through the process of elimination here, don't shoot me for that eh
just can't see any point in running 3xtemp at this stage..?as such it seems that 2xtempaa = the best solution atm, on less intensive games than farcry you may want to try out 4xaa with tempaa=2 as this should still give you very playable frame rates, at some 6xAA quality for much less of a performance hit
can someone else confirm?
i set TEMPAA to 1 with 2xAA and it looks better.....?
er scrap that, looking at too many lines
noticed that the tempaa method on the fsaa test prog takes a few secs to kick in (if you watch the lines at the bottom you notice the aliasing isnt great, and then it kicks in properly and its nice and smooth
but what do the various settings for TEMPAA relate to? ie 1/2/3 ?
i set TEMPAA to 1 with 2xAA and it looks better.....?
er scrap that, looking at too many lines
noticed that the tempaa method on the fsaa test prog takes a few secs to kick in (if you watch the lines at the bottom you notice the aliasing isnt great, and then it kicks in properly and its nice and smooth
but what do the various settings for TEMPAA relate to? ie 1/2/3 ?
no, i understood you quite clearly.dr3amz said:nevermind think you misread what i was saying though
joe said you would lose quality at 3temp - but i didnt notice that, but performance stayed the same. so i tested with 6xAA to check whether 3xtemp had an impact using the highest AA setting. results proved (for me at least) that using 3xtemp and 6xaa was no different to 2xtemp and 6xAA - in IQ or performance.
just going through the process of elimination here, don't shoot me for that eh
as such it seems that 2xtempaa = the best solution atm, on less intensive games than farcry you may want to try out 4xaa with tempaa=2 as this should still give you very playable frame rates, at some 6xAA quality for much less of a performance hit
Joe said you MIGHT lose quality, and that performance would definitly stay the same. You then proceeded to comment that you didnt see any IQ difference, and then expounded on the lack of performance difference - indicating to me that you hadn't seen that Joe said there wouldnt be.
As i said, whatever, it doesnt really matter, you seem to understand now.
3x temp COULD result in better IQ if high enough frame rates/refresh rates are attainable (which they might be for much older games).
how often the sample pattern changes.dr3amz said:can someone else confirm?
i set TEMPAA to 1 with 2xAA and it looks better.....?
er scrap that, looking at too many lines
noticed that the tempaa method on the fsaa test prog takes a few secs to kick in (if you watch the lines at the bottom you notice the aliasing isnt great, and then it kicks in properly and its nice and smooth
but what do the various settings for TEMPAA relate to? ie 1/2/3 ?
2 means it flips back and forth between two different patterns.
3 means it flips through 3 different patterns.
One pattern per frame.
The "TemporalAAFrameThreshold" setting has no effect for me (assuming it is what we think it is). I have it set to 60 and you can pick certain gaming conditions in which the drop from 60=>30=>15FPS etc is clearly accompanied by distinct levels of flicker severity. This is with v-sync on or off in the control panel.
MuFu.
MuFu.
Althornin said:how often the sample pattern changes.dr3amz said:can someone else confirm?
i set TEMPAA to 1 with 2xAA and it looks better.....?
er scrap that, looking at too many lines
noticed that the tempaa method on the fsaa test prog takes a few secs to kick in (if you watch the lines at the bottom you notice the aliasing isnt great, and then it kicks in properly and its nice and smooth
but what do the various settings for TEMPAA relate to? ie 1/2/3 ?
2 means it flips back and forth between two different patterns.
3 means it flips through 3 different patterns.
One pattern per frame.
thanks, so i'm thinking what does 1 do - because it seems to have the same effect... hmm think i might try a reboot instead of the `smart shader reset` for d3d.
just tried 1/2/3/4 and all settings produce the same patterns within the fsaa tester - is that right? so for being thick here, just trying to eliminate things
*EDIT*
ah boll*x - just rebooted with 3tempaa and now i see the diff doh - seems the smartshader reset doesnt work, thought it did - i'll rebench 3dmark using 3xtempaa now its working correctly.
as you say - same speeds, but some flickering (especially noticable on FSAA Viewer) and not really a lot of difference in quality...
keep playing
Wow!! this is really cool!!!
It works great too!!!!
hehe, now I can set 2xaa and temp aa to 2 and simulate the quality of 4xaa. Sweet!!!
I always have had vsync running so I'm getting even more FPS than usual because I can back off on the AA and still have pretty much the same efect.
It works great too!!!!
hehe, now I can set 2xaa and temp aa to 2 and simulate the quality of 4xaa. Sweet!!!
I always have had vsync running so I'm getting even more FPS than usual because I can back off on the AA and still have pretty much the same efect.
dr3amz said:
tempaa at 1 basically turns it of.
MuFu said:
hmm wouldn't the body mass between someones butt and their eyeballs absorb most of the vibration :?
. I just tried ons_torlan with no bots (my xp2000+ can do around 60fps with that) and it does indeed look bad.2x3t looks fine at something that can maintain 120fps (3dmark2000 demo for example), I find about 100fps is the minimum before it starts to get noticably flickery.
Joe DeFuria
Legend
I'm trying to get it to work...
The FSAA viewer shows the frame by frame differences in windowed mode. However, when I switch to full-screen, the "temporal" aspect of it stops, and reverts back to plain AA.
I switch back to windowed....and the temporal mode returns. Anyone else have similar problems? (Using latest official Cats...and DVI interface to flat panel)
The FSAA viewer shows the frame by frame differences in windowed mode. However, when I switch to full-screen, the "temporal" aspect of it stops, and reverts back to plain AA.
I switch back to windowed....and the temporal mode returns. Anyone else have similar problems? (Using latest official Cats...and DVI interface to flat panel)