Page 2 of 7

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 16:34
by Cooknn
Chief Blur Buster wrote: ...Set "Max pre-rendered frames" to 1
Forgive me for taking this off topic for a moment, but what would you set this variable to with a G-Sync monitor? In game it defaults to "0", but I always change it to "1".

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 16:38
by Chief Blur Buster
At the drivers, 0 and 1 seems to be the same for VSYNC ON, as far as I am aware (0 rounds up to 1). But I haven't tried the difference during GSYNC mode, so I'd just leave it alone (keep it at 0) during GSYNC mode. Are you noticing any difference?

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 16:39
by Cooknn
Chief Blur Buster wrote:At the drivers, 0 and 1 seems to be the same for VSYNC ON, as far as I am aware (0 rounds up to 1). But I haven't tried the difference during GSYNC mode, so I'd just leave it alone (keep it at 0) during GSYNC mode. Are you noticing any difference?
It seems more fluid when I set it to 1. I swear I saw stutter at 0 which I believe is equal to 3 pre-rendered frames. These are set in an INI file for my game (Assetto Corsa). Sorry for the Hijack :?

Congratulations on the new monitor rob_IDDQD! It must be amazing :!:

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 17:01
by mello
Chief Blur Buster wrote: VSYNC ON Input lag goes down significantly when:
- Using a higher refresh rate such as 120Hz instead of 60Hz
- Set "Max pre-rendered frames" to 1
- Setting an in-game framerate cap a very tiny bit below Hz
Example: In Counter Strike:GO, enter "fps_max 119" or even "fps_max 119.8" at the console
(Important note: Driver capping, or any other external capping, will always have more lag than in-game capping, due to the way things work).

Doing all the above massively reduces input lag for VSYNC ON (and even helps GSYNC a bit too, for the capped-out situations). It does not eliminate all of it. But 80% to 90% of the VSYNC ON lag penalty disappears when all three techniques are done simultaneously.
And what about triple buffering ? How does it affect input lag in the best case scenario (VSYNC ON) that you just desribed ?

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 18:10
by Chief Blur Buster
mello wrote:And what about triple buffering ? How does it affect input lag in the best case scenario (VSYNC ON) that you just desribed ?
Triple buffering usually has less lag. But triple buffering has pros and cons. It introduces microstutters caused by the framerate-refreshrate match. If you are prioritizing motion clarity over input lag, then you need to use double buffered VSYNC and a powerful-enough GPU to maintain the framerate-refreshrate sync.

However, if your framerate varies a lot, triple buffering can look better than double buffering, and with less lag. But it isn't motion clarity nirvana, for the motion nirvana-seekers.

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 18:20
by flood
Chief Blur Buster wrote: Triple buffering usually has less lag.
the type with two backbuffers? yea that would reduce lag
but I've never seen that actually working in any program.

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 19:14
by ManuelG_NVIDIA
Cooknn wrote:
Chief Blur Buster wrote:At the drivers, 0 and 1 seems to be the same for VSYNC ON, as far as I am aware (0 rounds up to 1). But I haven't tried the difference during GSYNC mode, so I'd just leave it alone (keep it at 0) during GSYNC mode. Are you noticing any difference?
It seems more fluid when I set it to 1. I swear I saw stutter at 0 which I believe is equal to 3 pre-rendered frames. These are set in an INI file for my game (Assetto Corsa). Sorry for the Hijack :?

Congratulations on the new monitor rob_IDDQD! It must be amazing :!:
There is no difference between 0 and 1 since technically it must always be in process of rendering the next frame.

It is not mentioned much but colors on this monitor are so much better than my old 3D Vision Asus 27" monitor. TN panel image quality sure has improved a lot over time.

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 01 Aug 2014, 20:03
by nimbulan
ManuelG_NVIDIA wrote:It is not mentioned much but colors on this monitor are so much better than my old 3D Vision Asus 27" monitor. TN panel image quality sure has improved a lot over time.
That is quite true. I was very impressed with the image quality on my VG248QE compared to my old monitors, though it needed some significant calibration to get there. Unfortunately nVidia's drivers do not allow forcing color profiles in games so I only get the best colors on the desktop.

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 02 Aug 2014, 03:06
by RealNC
mello wrote:
Chief Blur Buster wrote:And what about triple buffering ? How does it affect input lag in the best case scenario (VSYNC ON) that you just desribed ?
Tripple buffering will always increase lag by at least one frame. The kind of triple buffering the Chief is talking about, which actually reduces lag, is not used anymore in games. It was last used about 15 years ago or so. These days, triple buffering means just one more render-ahead buffer compared to double buffering.

Re: I got the ROG Swift (PG278Q) -- my tests

Posted: 02 Aug 2014, 04:47
by Rothron
I'm seeing fairly strong, I'm guessing, inversion artifacts that are especially visible in 3D-Vision mode. I can see it in non-3d modes as well, mostly when scrolling fast on web pages with lots of images. It appears as alternating light and dark vertical lines.
It's also visible when viewing something that moves horizontally 1 pixel per refresh such that my eyes will follow the alternating dark and light lines.

How is your screen in this regard? I'm still wondering if I have a bad panel or if that's just the way they are. Otherwise I'm very impressed with the screen.