To G-sync or not to G-sync...

[Muahdib20]

My basic setup is a Ryzen 1600 and Gtx 1080ti. I'm trying to decide if G-sync is going to be worth the extra money. I will be getting a 27" 1440p, 144 or higher hertz refresh rate monitor. Also trying to decide if I want to go IPS or TN as well, but mostly focused on the G-sync conundrum.

It really comes down to how noticeable is tearing and such going to be without G-sync at high refresh rates. Is it going to be super noticeable or just a blip here and there. Another issue I know that may come up is as new games come out my 1080ti will not be able to keep up the high fps that you need for a high refresh rate, so G-sync might be worth the cost now.

I play FPS games for the most part but I'm not young so don't have the lightning reflexes I used to have so that is making me consider an IPS over TN. Once again I wonder how noticeable a little slower response time is going to be.

Your thoughts are greatly appreciated.

 

[Armenius]

Tearing will always be an issue if a game doesn't match the refresh rate of the monitor exactly at all times. You need to think more in the area of not needing to care about framerates anymore. Unless I notice a stutter issue I just crank everything up to the max in every game and I'm off to the races. And the experience is glorious. My PC specs are in my signature.

IPS vs. TN is really more about whether or not you can deal with the glow that is inherent in IPS. For the most part color and contrast is the same between the panel types if they have the same color depth. IPS panels these days are very fast and can almost match the motion clarity of a TN at high refresh rates.

 

[zone74]

G-Sync is first and foremost about smooth gameplay.
It eliminates the requirement of having to keep the framerate matched to the refresh rate or one of its divisors.
It does this while also preventing screen tearing in a very low-latency way.
I would not buy a high refresh rate display without it. There is no hardware which can guarantee a locked 144 FPS in every game.

IPS panels are much higher quality than TN, and worth the extra cost in my opinion.
"IPS Glow" is so ridiculously overblown that it's difficult to read most of these complaints as anything more than trolls who couldn't afford the better panel, and are justifying the TN purchase to themselves.
It exists, but the number of times I've noticed it when using the monitor, or been bothered by it are a single digit percentage - compared to all of the issues that TN panels have which are a constant annoyance to me.

Response time is not an issue if you get one of the Asus or Acer displays.
Overdrive is extremely well tuned on their IPS panels so that it's very uniform across all transitions.
Motion blur is largely defined by framerate/refresh rate on these displays, not response time.

 

[Armenius]

G-Sync is first and foremost about smooth gameplay.
It eliminates the requirement of having to keep the framerate matched to the refresh rate or one of its divisors.
It does this while also preventing screen tearing in a very low-latency way.
I would not buy a high refresh rate display without it. There is no hardware which can guarantee a locked 144 FPS in every game.

IPS panels are much higher quality than TN, and worth the extra cost in my opinion.
"IPS Glow" is so ridiculously overblown that it's difficult to read most of these complaints as anything more than trolls who couldn't afford the better panel, and are justifying the TN purchase to themselves.
It exists, but the number of times I've noticed it when using the monitor, or been bothered by it are a single digit percentage - compared to all of the issues that TN panels have which are a constant annoyance to me.

Response time is not an issue if you get one of the Asus or Acer displays.
Overdrive is extremely well tuned on their IPS panels so that it's very uniform across all transitions.
Motion blur is largely defined by framerate/refresh rate on these displays, not response time.

In that same vein I would say complaints about gamma shift are overblown. But we are two individuals who can only speak from our own experiences.

The quality of the panel is going to come down to both the manufacturer of said panel and how it is assembled into the final monitor by the monitor manufacturers, not the panel type itself.

 

[aeliusg]

G-sync and variable sync in general is great as a fire and forget kind of feature that really makes the focus more about gameplay than trying to tune for a framerate target. You will be able to plop down highest settings and not have to worry about it for a lot longer than you previously would be able to even with the top performing card in any given generation. I'd say that it's not worth getting an LCD at all without it now. IPS and TN is a wash. IPS will still look better next to the TN, but nowadays TNs don't look bad at all. Then again, IPS isn't even slow anymore. If you don't want to save some money by picking up a Dell https://hardforum.com/threads/bestbuy-dell-27-inch-gsync-449.1942245/#post-1043174888 you may as well go IPS.

 

[Odellus]

if you looked at my PG278Q and QX2710 side-by-side, you wouldn't be able to tell which was which in terms of color. tftcentral's results back this up. TN has overtaken IPS; IPS has disgusting IPS glow that completely ruins dark scenes and has more motion blur. TN has poor viewing angles and slightly worse contrast ratio, but colors are basically the same (TN panels are generally more accurate for sRGB, actually, because they aren't wide gamut so they don't oversaturate colors with non-wide gamut aware applications) and if you like oversaturated colors, then just increase the saturation in your GPU's control panel.

the only time the poor viewing angles are noticeable is when the entire screen is filled with the same color which really just doesn't happen almost ever unless you're watching cartoons or something. games are more enjoyable due to the faster response times and any dark scenes, be it in games or movies, are actually watchable because you don't have huge white splotches in all four corners of the screen 24/7.

buy VA or OLED for tv/movies and slow games, buy TN for serious competitive gaming and/or fast games, buy IPS for professional color-critical work. simple.

 

[zone74]

if you looked at my PG278Q and QX2710 side-by-side, you wouldn't be able to tell which was which in terms of color. tftcentral's results back this up. TN has overtaken IPS; IPS has disgusting IPS glow that completely ruins dark scenes and has more motion blur. TN has poor viewing angles and slightly worse contrast ratio, but colors are basically the same (TN panels are generally more accurate for sRGB, actually, because they aren't wide gamut so they don't oversaturate colors with non-wide gamut aware applications) and if you like oversaturated colors, then just increase the saturation in your GPU's control panel.

the only time the poor viewing angles are noticeable is when the entire screen is filled with the same color which really just doesn't happen almost ever unless you're watching cartoons or something. games are more enjoyable due to the faster response times and any dark scenes, be it in games or movies, are actually watchable because you don't have huge white splotches in all four corners of the screen 24/7.

buy VA or OLED for tv/movies and slow games, buy TN for serious competitive gaming and/or fast games, buy IPS for professional color-critical work. simple.

This is what I was talking about with the hyperbolic talk surrounding IPS Glow.
Odellus is very anti-IPS for some reason.

Reviews such as those from TFT Central use a meter that sits flat on the display to measure color accuracy.
This does not take viewing angle into account at all.

This is their viewing angle comparison for the PG278Q (TN) and PG279Q (IPS)

I've had people on here suggest that this image was an example of IPS Glow, when it's caused by - and was intended to show - the effects of matte display coatings vs displays that have AR-coated Glass laminated to the LCD panel.

I consider IPS Glow to be negligible on my PG348Q. Here it is at a wide angle in a dark room (though not with the darkest of images) compared to the VA panel used in the HP Omen X35 (taken from this review)

TN panels are even worse, with the image changing noticeably as you move your head around. (same thing happens with VA) IPS provides a much more stable image.
3ms average response time for TN vs 5ms average response time for these new gaming IPS monitors is basically negligible, since image persistence on a flicker-free display will be above 5ms until you reach 200 FPS+Hz.

 

[Daffan]

G-Sync is first and foremost about smooth gameplay.
It eliminates the requirement of having to keep the framerate matched to the refresh rate or one of its divisors.
It does this while also preventing screen tearing in a very low-latency way.
I would not buy a high refresh rate display without it. There is no hardware which can guarantee a locked 144 FPS in every game.

IPS panels are much higher quality than TN, and worth the extra cost in my opinion.
"IPS Glow" is so ridiculously overblown that it's difficult to read most of these complaints as anything more than trolls who couldn't afford the better panel, and are justifying the TN purchase to themselves.
It exists, but the number of times I've noticed it when using the monitor, or been bothered by it are a single digit percentage - compared to all of the issues that TN panels have which are a constant annoyance to me.

Response time is not an issue if you get one of the Asus or Acer displays.
Overdrive is extremely well tuned on their IPS panels so that it's very uniform across all transitions.
Motion blur is largely defined by framerate/refresh rate on these displays, not response time.

IPS Ultrawides are disgusting in dark scenes. My last 2 LG and 1 Crossover have been baaaaaaaad. Like someone is shining yellow and white torches on the screen.

also the AUO 27" panel sometimes has yellow doodoo stains in the corners. Otherwise the 165hz ones are excellent.

 

[zone74]

IPS Ultrawides are disgusting in dark scenes. My last 2 LG and 1 Crossover have been baaaaaaaad. Like someone is shining yellow and white torches on the screen.
also the AUO 27" panel sometimes has yellow doodoo stains in the corners. Otherwise the 165hz ones are excellent.

Unfortunately TFT Central are not consistent at all with their IPS Glow images, and they're all very over-exposed which makes it look much worse than you see in person.
It seems like ASUS displays have a blue/purple tint while others can be yellow/brown - which I agree looks really bad.

Acer X34:

ASUS PG279Q:

ASUS PG348Q:

Again: my panel looks nothing like that PG348Q image in person, since the photo is so over-exposed, but it's not tinted brown.
And IPS Glow is still the least objectionable change with viewing angle, compared to TN's inversion or VA's gamma shifting/desaturation - at least if it's the more neutral or blue/purple variant. (I've never seen a panel that goes brown in the corners)

 

[YongKang Ma]

1. You'll have tearing if your FPS doesn't match monitor refresh rate. That's why adaptive sync tech is such a goodness since it allows your monitor refresh rate to adapt to your GPU output. Therefore, eliminating tearing.

2. IPS panels are extremely good right now. To my knowledge, they can achieve a GTG (gray to gray) response time of ~5 ms and lower. That's comparable to TN but you'll get much better color accuracy and better viewing angles.

3. Don't worry about FPS numbers too much. Anything upward of 60 is a bonus. You can turn down the settings to high and enjoy a sizeable FPS boost without sacrificing much visual fidelity. I would even urge you to customize your settings so that the game looks/performs the best for you.

 

[Mokkat]

Any 1440p 144hz TNs besides the Dell ones with a lighter AG coating? A glossy or at least semi-glossy 1440p 144hz TN with Freesync would probably be my monitor of choice

 

[Odellus]

Any 1440p 144hz TNs besides the Dell ones with a lighter AG coating? A glossy or at least semi-glossy 1440p 144hz TN with Freesync would probably be my monitor of choice

there aren't any, unfortunately, at least not from the mainstream brands. i came from a full glossy monitor and i can tell you that you get used to the AG, i don't even notice it anymore.

 

[Chief Blur Buster]

2. IPS panels are extremely good right now. To my knowledge, they can achieve a GTG (gray to gray) response time of ~5 ms and lower. That's comparable to TN but you'll get much better color accuracy and better viewing angles.

Keep in mind if you're buying a GSYNC monitor mainly for GSYNC, or also to use its ULMB:

One consideration: Blur Reduction (aka ULMB)

5ms is still bigger than the length of the blanking interval of 120Hz modes. For good ULMB operation without the trailing double-image effect (strobe crosstalk), you need a very fast LCD that can "cram GtG into VBI" -- basically finish GtG in the time interval between refresh cycles. Alas, currently, only certain TN LCDs manage to go nearly completely strobe-crosstalk-free (at least for the center 3/4ths of the display)

That's why I'm fairly excited about future 0.5ms GtG panels: Even better ULMB / MBR modes.

Obviously, 2ms vs 1ms vs 0.5ms matters a hoot of nothing for 60Hz non-strobed when GtG blur is overwhelmingly hidden by persistence blurring (MPRT). However, pixel response suddenly becomes a world of importance for various pulsed operation modes (OLED, LCD, DLP, etc). For LCDs in particular, the worse the GtG, the more strobe crosstalk (double-image-problem) there is during motion blur-reduction modes (MBR). Some people don't really mind, while others are really bothered by it.

Or when needing to do blurfree sample-and-hold (like a "strobeless ULMB" or "strobeless LightBoost" -- which requires ultra-high refresh rates aka 480Hz and up) -- GtG is also important for that mode of operation too.

OLED can't come soon enough, but in the meantime, GtG performance improvements must keep marching -- even to sub-1ms-levels -- for improved strobe backlights -- and strobing at higher refresh rates.

 

[HiCZoK]

I am wondering. There is now way with most setups that modern games will run at 144 fps. NO WAY.
So no vsync for 144hz users. is tearing less noticable than on 60hz?

To what framerate does 144hz monitor locks if vsync is turned on but fps is lower ?

 

[Desert Fish]

To what framerate does 144hz monitor locks if vsync is turned on but fps is lower ?

Always an integer divisor, so 72, 48, etc.

 

[YongKang Ma]

Keep in mind if you're buying a GSYNC monitor mainly for GSYNC, or also to use its ULMB:

One consideration: Blur Reduction (aka ULMB)

5ms is still bigger than the length of the blanking interval of 120Hz modes. For good ULMB operation without the trailing double-image effect (strobe crosstalk), you need a very fast LCD that can "cram GtG into VBI" -- basically finish GtG in the time interval between refresh cycles. Alas, currently, only certain TN LCDs manage to go nearly completely strobe-crosstalk-free (at least for the center 3/4ths of the display)

That's why I'm fairly excited about future 0.5ms GtG panels: Even better ULMB / MBR modes.

Obviously, 2ms vs 1ms vs 0.5ms matters a hoot of nothing for 60Hz non-strobed when GtG blur is overwhelmingly hidden by persistence blurring (MPRT). However, pixel response suddenly becomes a world of importance for various pulsed operation modes (OLED, LCD, DLP, etc). For LCDs in particular, the worse the GtG, the more strobe crosstalk (double-image-problem) there is during motion blur-reduction modes (MBR). Some people don't really mind, while others are really bothered by it.

Or when needing to do blurfree sample-and-hold (like a "strobeless ULMB" or "strobeless LightBoost" -- which requires ultra-high refresh rates aka 480Hz and up) -- GtG is also important for that mode of operation too.

OLED can't come soon enough, but in the meantime, GtG performance improvements must keep marching -- even to sub-1ms-levels -- for improved strobe backlights -- and strobing at higher refresh rates.

Thanks for all this useful info!

 

[HiCZoK]

Always an integer divisor, so 72, 48, etc.

So in worst scenario it's 48fps and locked with riva tuner for better framepacing. but refresh rate is still 144... At least this is how I imagine scenario for more demanding games

 

[zone74]

So in worst scenario it's 48fps and locked with riva tuner for better framepacing. but refresh rate is still 144... At least this is how I imagine scenario for more demanding games

There's little benefit to a higher refresh rate unless the framerate matches.
And you need V-Sync to keep things smooth, not a framerate limiter.
Framerate limiters stutter because they are not synced to the refresh rate.

 

[Chief Blur Buster]

There's little benefit to a higher refresh rate unless the framerate matches.

Our tests show that higher Hz at the same framerate, has much less input lag.

30fps at 240Hz has far less input lag than 30fps at 60Hz on the same monitor.

This applies to both fixed refresh and VRR (both FreeSync and G-SYNC).

And you need V-Sync to keep things smooth, not a framerate limiter.
Framerate limiters stutter because they are not synced to the refresh rate.

Actually, there's now ways to use a frame rate limiter without adding stutters.

VRR+framerate limiter = "near lagless VSYNC ON"

Lag tests showed that, too.

People who love VSYNC ON but hate the lag, can use the frame rate limiter trick to get perfect-looking VSYNC ON looking motion at a specific frame rate on a variable refresh rate monitor. Frame rate is the refresh rate. Good for greatly reduced emulator lag.

And one bonus, during VRR operation, the frame drops aren't visible as a stutter.... That is the variable refresh rate advantage, modulating minor variances in refresh rate to minor variances in frame rate.

Let's consider an odd situation: A frame rate cap of 57 with the game keeping up means your monitor is essentially looking like perfect VSYNC ON 57Hz...and if frame rates slow down to 55fps, it still looks like perfect VSYNC ON 55Hz but without the VSYNC ON lag.

If you are trying to reduce VSYNC ON lag, then there is no choice but to use creative tricks that also includes a frame rate limiter plus sometimes a different sync technology, to gain exactly the perfect stutterfree VSYNC ON look.

Also, several people have also successfully combined RTSS (only one precise enough to do so) with a frame rate limit 0.01Hz below refresh rate, to reduce input lag relative to using VSYNC ON, and getting at least one framebuffer lag savings while using VSYNC ON.

Go to www.vsynctester.com (Or even our TestUFO, it now has an exact-refresh test available in one screen) then configure RTSS frame rate cap to 0.01Hz below that displayed number, and half of your VSYNC ON lag suddenly disappears.

Some users have gone as low as 0.006Hz below, but there are limits to RTSS precision and refresh-rate-measuring precision, and besides, a precision 0.01Hz difference is 1 tiny microstutter every 100 seconds. The big bonus -- perfect VSYNC ON look without most of the lag! A smidgen more than VSYNC OFF, but far less than ordinary default VSYNC ON.

The frame rate cap in an ultra-accurate capping utility, combined with doing VSYNC ON too, is also zero-stutter in many games (except for less than one per minute induced by cap-mismatch to Hz). Tests have shown that. NVCP/NVInspector/RivaTuner, are not accurate enough to do stutterfree frame rate capping, but RTSS is, once you use an ultra-accurate refresh rate measurement utility. Note, this precision is only necessary for fixed-Hz, as VRR monitors are very forgiving of imperfect frame rate capping, and in-game frame rate capping is the recommended method for VRR.

Also, many ULMB/LightBoost fans tend to prefer VSYNC ON (due to strobing-amplified micro-stutters) so getting lag-reduced blur reduction Is a big boon for some people unusually sensitive to motion blur & stutters but also hates input lag. Some of us get nausea from motIon blur (a few even get headaches from that, and get more relief from ULMB than the typical ergonomic fixes of PWM-free or low blue light), so the science of optimizing a nearly lagless VSYNC ON became a science among a few Blur Busters fans, and we found the RTSS trick.

VSYNC OFF is still the sage advice for CS:GO but several of us have creatively come up with lag-saving VSYNC ON tricks many don't realize exists.

You know, not everyone plays competitively, and we need to also kill unnecessary VSYNC ON lag too. "VSYNC ON look-and-feel" is better in emulators, low-framerate games, sensitivi tearing/stutter, better strobing, if we could only get rid of VSYNC ON lag -- and we successfully have removed up to 75% of VSYNC ON lag via various tricks. Some emulators, specifically, suddenly get as much as 30-50ms less lag (more than two 60Hz refresh cycle savings, thanks to a trifecta of benefits -- VSYNC ON buffer savings + faster frame delivery + faster frame scanout) when using a 240Hz VRR monitor + 60fps frame rate capping. That huge input lag savings (while still having the perfect stutterfree VSYNC ON look) is very useful to many people who are stuck with 60fps software.

TL;DR: Creative careful use of frame rate caps can still preserve the stutter free "VSYNC ON" look-and-feel while removing the majority of VSYNC ON input lag.
* GSYNC/FreeSync: Any frame rate cap sufficiently below max Hz = "Near-lagless VSYNC ON for variable-refresh-rate monitors" (and frame rates can gently dip harmlessly without visible stutter)
* Fixed-Hz: VSYNC ON + RTSS ultra-precision cap ~0.01fps below exact-measured Hz = "Near-lagless VSYNC ON for fixed-refresh monitors" (assuming frame rates always keeps up)

 

[HiCZoK]

There's little benefit to a higher refresh rate unless the framerate matches.
And you need V-Sync to keep things smooth, not a framerate limiter.
Framerate limiters stutter because they are not synced to the refresh rate.

This makes sense... Even on my 60hz display I have turn on vsync and limit fps to 60 with riva tuner to makes games be super smooth.

So the question is : Do i get poor quality screen with gsync (like that s2417dg) and gsync takes care of smoothness and lower fps.
Or do I go for better quality display like some of those samsung qled screens... but these only have freesync which don't work with nvidia

 

[Chief Blur Buster]

This makes sense... Even on my 60hz display I have turn on vsync and limit fps to 60 with riva tuner to makes games be super smooth.

That is a good way to reduce lag while keeping smoothness -- simultaneously using VSYNC ON *and* a precision frame rate capping like RTSS (RivaTuner).

But instead of using "60", you should measure your exact refresh rate and then change your RivaTuner/RTSS cap to 0.01fps less than refresh rate. That keeps your smoothness, but removes most of the VSYNC ON lag.

It's very important it be fractionally below -- to remove most of the VSYNC ON lag. Most computers set 60Hz are actually something like "59.941774Hz" or "60.01256Hz". So measure your refresh rate exactly, then subtract 0.01, and punch that into RTSS/RivaTuner as your frame rate cap (the only one accurate enough to do this, so far). So if you benchmark "59.941774Hz" for your precision refresh rate measurement utility you would enter "59.93" into RTSS/RivaTuner to avoid buffers from piling up into an increased VSYNC ON input lag.

If you hate stutters, you can also go the VRR route, too.

 

[HiCZoK]

Yep so I probably want gsync... but I don't want to abandon sweat 3000:1 contrast... With my 1060 I am never getting stable 144fps in all games

 

[Absalom]

Why has no one mentioned Fast-Sync? Fast-sync + G-Sync is pretty much the way to go these days, at least for gaming. Even moreso if you're already sporting a Pascal gpu. It's a bit moot to convince someone invested in an AMD gpu to not go Freesync, although I still think they're missing out on some things.

None of the methods mentioned above offer the same overall properties as decoupling the renderer from the display synchronization (i.e. Fast-Sync). This tech combination pretty much offers almost all the "pros" of VSYNC OFF and VSYNC ON - with little to none of the "cons". Usually the cons are "Why doesn't Fast-Sync work with this random/old game?" or running the renderer full-tilt generates too much heat or noise for the system in question.

The reason CS:GO players prefer VSYNC OFF is because the game renderer will sample input (usually it's the mouse here that matters) at the highest possible rate. This reflex action is often faster than even the player's eyes can process, so it's less about the refresh rate and more about input response. Having a game renderer run full-tilt leaves little in the way of misinterpreting a pro player's "flick shots."

That said, if you watch a ton of movies, then G-Sync is probably a poor investment. When it comes to watching movies (film, blu-ray, etc.), I prefer turning G-Sync off and continue to use tried and true methods such as ReClock, madVR, custom refresh rates, etc. to help sync the video framerate perfectly with the display.

Then there's the whole IPS vs TN debate, which is more a state of the industry. Also how much refresh rate is overkill? I'm not sure myself. All I know is that I can't go back to anything less than 100Hz, much less TN.

 

[Daffan]

That is a good way to reduce lag while keeping smoothness -- simultaneously using VSYNC ON *and* a precision frame rate capping like RTSS (RivaTuner).

But instead of using "60", you should measure your exact refresh rate and then change your RivaTuner/RTSS cap to 0.01fps less than refresh rate. That keeps your smoothness, but removes most of the VSYNC ON lag.

It's very important it be fractionally below -- to remove most of the VSYNC ON lag. Most computers set 60Hz are actually something like "59.941774Hz" or "60.01256Hz". So measure your refresh rate exactly, then subtract 0.01, and punch that into RTSS/RivaTuner as your frame rate cap (the only one accurate enough to do this, so far). So if you benchmark "59.941774Hz" for your precision refresh rate measurement utility you would enter "59.93" into RTSS/RivaTuner to avoid buffers from piling up into an increased VSYNC ON input lag.

If you hate stutters, you can also go the VRR route, too.

Thanks I'l try this.

I had to go into Rivatuner 7.0.0 config for default profile and add some decimal stuff apparently.

[Framerate]
Limit=5998
LimitDenominator=1000

I benchmarked 59.999060hz so I put in 5998 (Which should be 59.98hz apparently)

Let me know If i did anything wrong lol.

 

[IdiotInCharge]

Always an integer divisor, so 72, 48, etc.

So in worst scenario it's 48fps and locked with riva tuner for better framepacing. but refresh rate is still 144... At least this is how I imagine scenario for more demanding games

Keep in mind (just fyi) that we're talking about instantaneous framerate here, which applies to a single frame. If the next frame is rendered ahead of the refresh rate, then your instantaneous framerate goes back up.

Trying to describe this situation properly is why we use frametimes instead of framerates when we want more than a quick look at performance.

 

[Chief Blur Buster]

Thanks I'l try this.

I had to go into Rivatuner 7.0.0 config for default profile and add some decimal stuff apparently.

[Framerate]
Limit=5998
LimitDenominator=1000

I benchmarked 59.999060hz so I put in 5998 (Which should be 59.98hz apparently)

Let me know If i did anything wrong lol.

That's pretty much correct.

 

[p_monks33]

I really like GSYNC, its nice to play games that I know my GPU can't push at 60FPS, and it is still a smooth experience.

 

[zone74]

Our tests show that higher Hz at the same framerate, has much less input lag.
30fps at 240Hz has far less input lag than 30fps at 60Hz on the same monitor.
This applies to both fixed refresh and VRR (both FreeSync and G-SYNC).

Yes, it helps with latency. What I meant was that it does little for motion handling, and it also has the potential to introduce frame-pacing issues. (using a limiter like RTSS is no guarantee against that)
Unfortunately you picked an example that would require 1/8th Refresh V-Sync which doesn't seem to exist, but say 60 FPS at 240Hz has 4 potential windows for a frame to be displayed without being dropped, while 60 FPS at 60Hz or 1/4 Refresh V-Sync guarantees good frame-pacing.
If you're shorting the buffers with a framerate limiter, it should hopefully present frames with good pacing, but it's not guaranteed like V-Sync.

Actually, there's now ways to use a frame rate limiter without adding stutters.
VRR+framerate limiter = "near lagless VSYNC ON"
Lag tests showed that, too.

Yes, that's one of the reasons why I highly recommend VRR displays instead of fixed refresh rate displays.

TL;DR: Creative careful use of frame rate caps can still preserve the stutter free "VSYNC ON" look-and-feel while removing the majority of VSYNC ON input lag.
* Fixed-Hz: VSYNC ON + RTSS ultra-precision cap ~0.01fps below exact-measured Hz = "Near-lagless VSYNC ON for fixed-refresh monitors" (assuming frame rates always keeps up)

I've tried this but it does not work for me. It's all very smooth for a while and then I get a frame-pacing nightmare where it looks like the game is running at half the framerate for 10-20 seconds, back to smooth for a minute or two, and then back to the frame-pacing nightmare. This is not evident on RTSS' framerate or frametime graph, only the display. (FCAT would probably capture it)
Only V-Sync remains consistently smooth at all times on a fixed refresh rate display in my experience. The problem with that is latency - which is why I don't recommend fixed refresh rate monitors now that VRR is available.

 

[Desert Fish]

That is a good way to reduce lag while keeping smoothness -- simultaneously using VSYNC ON *and* a precision frame rate capping like RTSS (RivaTuner).

But instead of using "60", you should measure your exact refresh rate and then change your RivaTuner/RTSS cap to 0.01fps less than refresh rate. That keeps your smoothness, but removes most of the VSYNC ON lag.

It's very important it be fractionally below -- to remove most of the VSYNC ON lag. Most computers set 60Hz are actually something like "59.941774Hz" or "60.01256Hz". So measure your refresh rate exactly, then subtract 0.01, and punch that into RTSS/RivaTuner as your frame rate cap (the only one accurate enough to do this, so far). So if you benchmark "59.941774Hz" for your precision refresh rate measurement utility you would enter "59.93" into RTSS/RivaTuner to avoid buffers from piling up into an increased VSYNC ON input lag.

I generally don't notice any input lag with V-sync. But I was wondering, have you measured how RTSS compares to setting maximum pre-rendered frames to 1?

 

[Chief Blur Buster]

I've tried this but it does not work for me. It's all very smooth for a while and then I get a frame-pacing nightmare where it looks like the game is running at half the framerate for 10-20 seconds, back to smooth for a minute or two, and then back to the frame-pacing nightmare. This is not evident on RTSS' framerate or frametime graph, only the display. (FCAT would probably capture it)

Which games are you playing? This could be an interesting data point of research -- specific games, specific drivers, specific brand (AMD versus NVIDIA) that does it but not others.

Another potential issue is that the cap is not accurate enough; try testing 0.1Hz below rather than 0.01Hz below. If the framepacing nightmare stops, then try 0.05Hz, then 0.025Hz, etc. Refresh rate calculators don't always perfectly line up with RTSS precision, and different cards may produce different precision.

Tests were done with a 1080Ti with latest WHQL drivers (at the time).

I generally don't notice any input lag with V-sync. But I was wondering, have you measured how RTSS compares to setting maximum pre-rendered frames to 1?

There's an old forum thread at Blur Busters containing lots of input lag experimentation with all the frame-cappers (RTSS, NVInspector, in-game, etc) as well as various pre-rendered frame settings.

The TL;DR is that after weeks of forum experimentation, we all found in-game capping had the least lag (for lowest-lag FreeSync / GSYNC, cap a few frames per second below Hz), while RTSS capping was the most mathematically precise (for ultra-low lag fixed-Hz VSYNC ON use). Some (not all) of these findings were written into the GSYNC-101 article series.

However, this is not always reliable in 100% of all games -- sometimes they're just stubbornly stuttery.

 

[ncjoe]

why the heck is it so complicated to run gsync or vsync ?????
have to do this , do that ..........should be just plug it in and it automaticly
sets it to perform the best with video card you have.....

You'd think after being out all these years NVidia and amd would have
ease of use built in to the cards/ monitors...

 

[IdiotInCharge]

why the heck is it so complicated to run gsync or vsync ?????
have to do this , do that ..........should be just plug it in and it automaticly
sets it to perform the best with video card you have.....

You'd think after being out all these years NVidia and amd would have
ease of use built in to the cards/ monitors...

Nvidia erred toward customization. Most G-Sync monitors also ship with ULMB, which cannot be used at the same time as G-Sync. Further, there are different technologies that can be used by games depending on each game's implementation.

In the driver, you can set a G-Sync tag to show when G-Sync is active so that you know that it's working.

Also, in most latest games, the biggest issue is that they want to run in a borderless window, and G-Sync requires games to run in full screen mode. Fixing that is usually enough.

 

[Absalom]

In the driver, you can set a G-Sync tag to show when G-Sync is active so that you know that it's working.

Some G-Sync monitors have LEDs that change color to notify you when G-Sync is active/inactive.

Also, in most latest games, the biggest issue is that they want to run in a borderless window, and G-Sync requires games to run in full screen mode. Fixing that is usually enough.

There's an option in the Nv Control Panel to enable G-Sync to work in both fullscreen and windowed. Unfortunately, it's a global operation and would be far more useful as a per-application setting.

 

[aeliusg]

why the heck is it so complicated to run gsync or vsync ?????
have to do this , do that ..........should be just plug it in and it automaticly
sets it to perform the best with video card you have.....

You'd think after being out all these years NVidia and amd would have
ease of use built in to the cards/ monitors...

It's not. These guys are trying to cut down to the very minimum of input lag and other visual-kinesthetic feedback discrepancies possible in their given configuration. For the typical user, that is probably >99% of people, G-Sync is a set it and forget it feature.

 

[zone74]

why the heck is it so complicated to run gsync or vsync ?????
have to do this , do that ..........should be just plug it in and it automaticly
sets it to perform the best with video card you have.....
You'd think after being out all these years NVidia and amd would have
ease of use built in to the cards/ monitors...

You can just plug it in and have it work.
The issue is that if the framerate reaches or would exceed the refresh rate, it switches from G-Sync to V-Sync which either adds latency or screen tearing depending on whether you have V-Sync set on or off. (it's recommended to be enabled globally in the NVIDIA Control Panel)
If you use RTSS to limit the framerate 3 FPS lower than the maximum refresh rate, you prevent it ever leaving the G-Sync range.
You can set RTSS to launch at system startup with a global framerate limiter and never have to think about it again.

Some G-Sync monitors have LEDs that change color to notify you when G-Sync is active/inactive.

Yes, but if you enable Windowed Mode G-Sync it will always be "on" even if it's not currently being used.
I use my monitor's "FPS Counter" display to confirm whether G-Sync is working. If it's fixed at 100Hz (it maximum) then it's not being used.

There's an option in the Nv Control Panel to enable G-Sync to work in both fullscreen and windowed. Unfortunately, it's a global operation and would be far more useful as a per-application setting.

In NVIDIA Profile Inspector you can set "GSYNC Application State" to Disallow, which prevents G-Sync from being used with that application.
Do not touch anything else related to G-Sync on that profile, or else the screen may flicker if G-Sync is actually switched on/off per-application (LED turns white), rather than staying "enabled" (red LED) but remaining fixed at the maximum refresh rate.
I created a global "_DISALLOW_GSYNC" profile which is set like this, and I add programs to it when I encounter one that does not work well with Windowed-Mode G-Sync. (some applications drop the refresh rate to 1Hz when idle)

Which games are you playing? This could be an interesting data point of research -- specific games, specific drivers, specific brand (AMD versus NVIDIA) that does it but not others.

That was The Evil Within on a GTX 960. idTech 5 is really sensitive to timing errors - though it works well with G-Sync.
Perhaps spending more time adjusting the refresh rate would help, but I did give it some time with no success.
That said, I don't generally use that system/display for gaming now, I only tried it out of curiosity after reading about it on the Blur Busters site.

 

[Chief Blur Buster]

why the heck is it so complicated to run gsync or vsync ?????

GSYNC and FreeSync and VSYNC-ON is usually plug and play (with the corresponding graphics cards, obviously).
Car owner equivalent: Just drive-and-go, use quick-oil-change retailers.

But we're tweakers. We just know it just simply doesn't get lowest-theoretically-possibility latency without tweaks. Much like monitors don't produce perfect colors until you calibrate it with your favourite Contrast/Brightness/Saturation setting. Likewise, we are just doing "input lag calibration".
Car owner equivalent: Just like opening your car's hood / engine computer / nitrogen for tire inflation / expensive oils and tuning your car to perform better.

It's not. These guys are trying to cut down to the very minimum of input lag and other visual-kinesthetic feedback discrepancies possible in their given configuration. For the typical user, that is probably >99% of people, G-Sync is a set it and forget it feature.

Correct.

That was The Evil Within on a GTX 960. idTech 5 is really sensitive to timing errors - though it works well with G-Sync.
Perhaps spending more time adjusting the refresh rate would help, but I did give it some time with no success.
That said, I don't generally use that system/display for gaming now, I only tried it out of curiosity after reading about it on the Blur Busters site.

Fascinating, we haven't tested that game. Perhaps it's the engine, but more tests are needed. Some engines may need 0.1Hz headroom while other engines are perfectly happy at 0.006Hz difference between framecap & refreshrate. The forum high-speed video test showed a very clear input-lag-reduced VSYNC ON. But apparently, this tweak failed in your particular instance -- whether the game engine, driver, or differential.

One issue I've found is that frame-pacing nightmares can happen if a game has a flawed VSYNC trigger. The VSYNC is a pulse in the display signal too, and lasts for around 3 to 5% -- e.g. about 1-2ms at 60Hz. If a frame finishes rendering in that time period, or if a frame begins rendering inside that time period AND then finishes rendering during the NEXT VSYNC time period, it might falsely cause an unnecessary "wait for next VSYNC" (double VSYNC effect / unnecessary frame rate halving), or an unexpected "render-twice" in a single refresh cycle.

<FOR DEVELOPERS>
Programmers -- if you do VSYNC state polling, then the proper VSYNC frame-pacing logic is to check for the leading edge of the VSYNC pulse (the moment the VSYNC signal pulse is entered, not exited -- e.g. the moment that RasterStatus.INVBlank changes from FALSE->TRUE) rather than arbitrarily checking if VSYNC is TRUE. There has been some engines that poll this flag before presenting the frame (as an attempt to reduce VSYNC ON input lag, or for a different reason). But this can be tricky because the flag remains TRUE for up to about a couple of milliseconds per refresh cycle, before returning back to FALSE. If the software developer does not focus on a leading or trailing edge of the VSYNC pulse, games can get framepacing nightmares in certain scenarios (60-30-60 situations). Things like unexpected race conditions against an external frame-rate capping utility, or with coincidential frametimes, or when rendertimes fall below VSYNC time. This is usually not a problem with most engines which just simply presents the frame (without any VSYNC checks or conscious lag-reducing techniques by the game developer) which is more reliable but creates lag. Now I wonder if idTech 5 is doing a flawed VSYNC poll while it's in VSYNC ON mode... (Or if it's the drivers, or if it's a bad differential)
</FOR DEVELOPERS>

Eventually, tools will need to be made to make it easier for software developers to reduce VSYNC ON input lag. It's a game of precision, but powerful-GPU users should have the choice to easily remove VSYNC ON input lag. VSYNC ON should not be hobbled in unnecessary latency, when smooth low-lag VSYNC ON techniques is indeed possible, but it's not always easy for game developers.

Monitors and graphics drivers don't ship configured to their best ability. Much like everyday-driver cars often don't ship at their fastest performance without opening your car's hood to do a little adjustment of the engine fluids / readjust tire inflation / car engine computer / etc and other stuff. (Or if you're not into cars -- same thing goes for bikes & transit mode-selection & scuba equipment & boats & whatever, replace with tuning of favourite "X" thingy)

Most of it has generally been plug-and-play. But best performance often just requires tweaks. But a display tweak potentially causes side effects (perfect in a game, bad in a different game) so monitors & drivers are often shipped at their lowest common denominators (at our chagrin sometimes), simply to make sure we are all able to just launch-and-run any game without fussing.