open wrote:I have the same monitor and I use gsync with vsync off with fps limit of 300 for overwatch. The tearing is hard to notice at 240 refreshes and 240+ fps. And when it goes lower gsync is very nice to keep it smooth. I don't really care alot about tearing. It wasn't even that bad at 144hz but the micro stutters were bad when framerates were in weird proportion to refresh.
It's certainly true that "refresh-rate-mismatch" microstutters are worse at lower refresh rates.
Mathematically, the stutter amplitude is equal to the inverse of frame rate.
Just covering ground, just to be safe. Definitions of stutters somewhat vary, but stutter, judder, microstutters, generally is defined as us by.
regular stutter: What people see at low frame rates even at perfect frametimes. e.g. 20fps, 24fps, 30fps
micro stutter: Visible minor erraticness in motion (even at high framerates, still visible but faint at 240Hz)
a stutter / single stutter / frame drop: A single stutter, usually a single frame discontinuity (frame repeat, frame drop, frame skip). Easy to see during perfect motion (e.g. TestUFO), especially in blur-reduced mode.
judder: Often similar to a form of microstutter, I've usually used this term to define 3:2 pulldown (24fps played at 60Hz)
TestUFO Animations of comparing "regular stutter" of different framerates
At 960 pixels per second and 240 frames per second, the stutter amplitude is only 4 pixels -- that's your motion blurring thickness as seen at
http://www.testufo.com (excluding GtG-added motion blurring -- that's above and beyond [url-
http://www.testufo.com/eyetracking]eye-tracking motion blur[/url]) .... But it also happens to be the microstutter amplitude too (e.g. for harmonic microstutters -- e.g. 239fps/241fps at 240Hz -- will create up to a 4-pixel "apparent" microstutter relative to fluid eye tracking position)
High-frequency stutters blend into motion blur
A great demo of stutter amplitudes can be seen at TestUFO count=6 for
http://www.testufo.com/framerates#count=6
At higher framerates, the "stutter" is so high frequency that it blends into human-perceived motion blur.
Motion Blur Reduction
The Black Frames animation configured to a large number of UFOs (for
240Hz users, for
120/144Hz users, and for
60Hz users) is also extremely educational too, on the motion-blur-versus-stutter relationship, in the exercise of black frame insertion motion blur reduction technique in a software-based manner. (Hardware-based
motion blur reduction techniques such as strobe backlights -- can do a far better job -- since they can reduce persistence to sub-frame cycles via backlight strobing)
GSYNC and VSYNC OFF
Now with variable refresh rate, like GSYNC (for non GSYNC/FreeSync users, see
this stutter-elimination GSYNC animation simulation) you don't have to worry about erratic stutters/microstutter effects until you hit really high framerates (240fps+). By the time you begin exceeding GSYNC's maximum (240fps) and use VSYNC OFF, the microstutters are already extremely small and you can keep gaining lag reducing advantages. If doing things this way (GSYNC + VSYNC OFF), I'd suggest raising your framerate cap higher (e.g. >500) since our tests still show continued VSYNC OFF lag-improving advantages all the way to >1000fps. (That said, if CPU/GPU/mouse starts behaving weird at high framerates -- game engine limitations, CPU starvation -- then that's another story. It also helps if you adjust your multicore setting in Source Engine games).
While for most use cases, GSYNC + VSYNC OFF really isn't a super-big use for most gamers, and our testing concluded it has relatively little use -- I do think it's not useless. It is still useful in certain cases (A) It keeps microstutters away until you're already a really high framerate where the microstutters is now a very small amplitude, and (B) It lets you access the ultralow lag of ultra-high-framerate VSYNC OFF (e.g. fps_max 1000fps)