forii wrote: ↑14 Apr 2020, 16:05
1) Alright, but why would someone use extra v-sync (with g-sync!!) if it does not active due to -3fps cap? + g-sync already give the pernament "v-sync experience)
is the v-sync actually working or not if it doesnt reach the equal fps of hz or above?
Even if you choose a frame rate cap, you still have to choose a fallback sync technology. By default, the fallback sync tech is VSYNC ON.
Caps are often aiming at averages. Capping is not always accurate.
A 141fps cap may create some frames that are 1/130sec and other frames 1/150sec. One of those frames gets the G-SYNC treatment, and the other frame gets the fallback sync technology (latency for that specific frame, or tearing for that specific frame).
A bigger differential means capping imperfections will occur less often.
A more accurate frame rate cap means capping imperfections will occur less often.
Depending on the game, the cap accuracy, and the capping differential --sometimes it will happen only instantaneously once every few seconds. Other times, it will happen continuously. Remember, for 200fps, there are potentially up to 200 different sync-technology treatments per second. If all frametimes are all within VRR range, then all of them got all the G-SYNC treatment. That's assuming all frametimes were permanently within the VRR range.
Please note: There is sometimes a pick-poison choice. In-game framerate caps are less accurate but have less lag. While RTSS frame rate caps sometimes have more lag than in-game frame rate caps, but is more accurate. Sometimes prioritization becomes necessary -- deciding which frame rate capper to use is a strategic move that often depends on the game and on the priorities you have. For
VRR capping, however, it's
usually better to use the lower-lag frame rate cap.
The more imperfect a frame rate cap is, the more often the frametimes will deviate to a frametime faster than a refresh cycle. That's why the common "cap a few below" advice for VRR. For very perfect frame pacing (e.g. RTSS and emulators), it is possible to use tighter cap (0.5fps or 1.0fps below). For very imperfect in-game framerate caps, it is possible to cap about 10fps below and still occasionally get fallback sync treatment (rare tearing or rare millisecond-lagged frames)
forii wrote: ↑14 Apr 2020, 16:05
And if so, why it gives some input lag if it isn't active like you said due to 3-fps bellow 240hz.
VSYNC OFF does the equivalent of real-time mid-raster splicing of frame slices, which means that frame rates higher than refresh rates, can sometimes have multiple frameslices per refresh cycle. This can bypass part of scanout latency. For more information, see
High Speed Videos of Scanout and
Advantages of Frame Rates Above Refresh Rates.
If pushed to the limit, VSYNC OFF ultra-high-framerates can become a scant 1-2ms lower lag than properly capped 240Hz G-SYNC. As long as you keep your refresh rate high, 240Hz G-SYNC has a worst-case scanout latency of 1/240sec ~= 4.1ms, and 360HGz G-SYNC has a worst-case scanout latency of 1/360sec ~= 2.7ms
The bottom line is the 3fps cap differential is the most common parrot advice becase it has worked so well for so many people. (See
Capping Differentials FAQ for some complex reading on this).
The right tool for the right job -- for example for games running frame rates far beyond refresh rates (e.g. CS:GO) may favour ultra-high-framerate VSYNC OFF for best scores, while other games running frame rates within VRR range (e.g. PUBG) may benefit far more from G-SYNC.
We are big proponents of "Right Tool For The Right Job".
forii wrote: ↑14 Apr 2020, 16:05
2)You said elmb with g-sync with frame rate cap has less lag than elmb + vsync on, that's correct (even if you said that v-sync doesnt active if fps are not above Hz of monitor).
That's right. Capped G-SYNC looks visually identical to VSYNC ON, but with less lag than VSYNC ON. Good caps looks just like perfect fps=Hz.
forii wrote: ↑14 Apr 2020, 16:05
3) What about ELMB-sync vs ELMB? Because I heard g-sync doesnt give any input lag, or it actually does?
forii wrote: ↑14 Apr 2020, 16:05
4)And... how does it work with 240hz, because I believe that does matter a lot, we are not talking about input lag of 60hz or even 144hz monitor, we talking here of 240Hz monitor with quite low input lag by itself (Asus VG259QM)
See
read my reply in the other thread
For 240Hz, scanout latency is very low. (Scanout latency is because not all pixels on a monitor refresh simultaneously).
Sometimes questions creating unnecessarily complex decision. The moral of the story is, Right Tool For Right Job -- especially if you earn money playing competitive games. For games that are running frame rates that are massively permanently beyond Hz (CS:GO is a common game) sometimes ultra-high-capped VSYNC OFF is the favourite tool of choice, for people where 1ms-2ms is actually important.
Let's cut to the chase. If you earn (1) championship money in esports, and (2) you only play CS:GO, and (3) monitors don't exist yet that capture the entire CS:GO frame rate range (e.g.
1000Hz monitor), then (4) It's often difficult to beat VSYNC OFF in competitive advantage. Most CS:GO pros use this.
Now.... CS:GO is not the only game ever invented. If you are trying to play games that that have framerates always within the G-SYNC range -- it can be a competitive advantage to use G-SYNC
with the right settings. Different games benefit more from G-SYNC than others. The visual advantage of G-SYNC (no stutters, no tearing) is good for some unoptimized engines (Example: PUBG), and the blur advantage of motion blur reduction (ELMB, ULMB, etc) is useful for very scrolly/panny/spinny games without crosshairs (Example: Rocket League).
There are some hidden competitive advantages lurking for non-CS:GO games in certain display technlogies. Some players (who play multiple games) use dramatically different settings for different games.