Gsync vs Freesync smoothness

[Hadookaan]

OK so I noticed on some games that use a g sync panel are more butter smooth, and free-sync is like slightly stutters/micro stutters in comparison, its still smoother then without freesync enabled, but not quite to the same as gysnc is like having vsync on, without the lag, when frame rates fluctuate, this was tested with Nvidia gpu 1070, the drivers used are the first freesyc compatible to be released, and currently on 430.53 all have been uninstalled with DDU, same issue, so have not tried freesync with amd card as I do not own one. the monitors I used to test this are:

XB271HU 27in 144/165hz gsync ips acer predator 1440p

Alienware 25in 240hz AW2518HF TN 1080p freesync

LG 27GK750F-B 27in TN 240Hz 1080p FreeSync

I did notice it happen with Pubg and other epic engine games, and Battlefield games, I will try others games with other engines, and update this, soon, I am not sure if this is Nvidia doing this through driver to make gsync look better, or if this is the same issue with a amd card! maybe when Navi AMD cards come out and they are good then I will have a chance to test it!

[Chief Blur Buster]

It's very important that displays gain submillisecond sync accuracy between frametimes and resfreshtimes where possible. Ideally, FreeSync is theoretically accurate down to one single horizontal scanrate unit (Typically just a few microseconds granularity) but this doesn't seem to be always happening.

[jorimt]

OK so I noticed on some games that use a g sync panel are more butter smooth, and free-sync is like slightly stutters/micro stutters in comparison, its still smoother then without freesync enabled, but not quite to the same as gysnc is like having vsync on, without the lag [...] have not tried freesync with amd card as I do not own one.

G-SYNC monitors contain a dedicated hardware module for VRR, whereas "G-SYNC Compatible" FreeSync monitors use driver software only. So, yes, it's very possible that G-SYNC Compatible mode on a FreeSync display isn't currently (and may be or may never be) as accurate/smooth as hardware G-SYNC.

I can't directly speak for true FreeSync on an AMD GPU as I've never tried it either, but as I understand it, I'm pretty sure other than the dynamic/variable overdrive differences between the two VRR formats, and FreeSync models that don't support LFC, that FreeSync on a FreeSync monitor (with LFC support) w/AMD GPU is otherwise equal VRR performance-wise to G-SYNC on a G-SYNC monitor w/Nvidia GPU.

[Chief Blur Buster]

Theoretically, there is enough software accuracy to framepace FreeSync as well as GSYNC -- it's just that the programmers of FreeSync drivers don't realize that I can see sub-1-millisecond microstuttering -- as in 0.5ms divergences between gametime and refreshtime (refreh-cycle-emitting-photons). On 60Hz FreeSync, a 1ms microstutter can easily be hidden in the 1/60sec (16.7ms) motion blur, but at 240Hz, a 1ms-error microstutter is a 25% change to a 4ms refresh cycle (1/240sec).

Power management sometimes inject 1ms-2ms errors into Present() timings according to my Tearline Jedi experiments, so I advise turning off power management, this may improve FreeSync framepacing. I find that I am able to get "tens-millisecond-league" precision most of the time in Tearline Jedi, provided I follow a few general best-practices. And framepacing compensation to prioritize gametime sync over rendertime sync. Since the varying rendertimes can diverge gametimes away from refreshtimes (pixels-to-photons) -- creating VRR microstutter.

Regardless....If driver programmers from AMD/NVIDIA are reading this.....framepace those FreeSync refresh cycles properly!

Users of FreeSync: Disable GPU power management to improve FreeSync fluidity slightly.

[jorimt]

Theoretically, there is enough software accuracy to framepace FreeSync as well as GSYNC

[...]

Regardless....If driver programmers from AMD/NVIDIA are reading this.....framepace those FreeSync refresh cycles properly!

Right, though the OP stated that he's only tried G-SYNC on a G-SYNC monitor w/module, and G-SYNC Compatible on FreeSync monitors, all only with an Nvidia GPU. He hasn't tried true FreeSync on a FreeSync monitor with an AMD GPU.

So I was simply giving AMD FreeSync the benefit of the doubt in this regard and assumed it's a G-SYNC Compatible mode-specific issue on FreeSync monitors currently, as, again, I've never experienced FreeSync on an AMD GPU either.

Of course, if you've experienced this issue yourself on a FreeSync monitor using FreeSync on an AMD GPU, that's another story, and, yes, something that AMD (and Nvidia with their G-SYNC Compatible mode) certainly should look into improving.

[1000WATT]

It seems to me wrong to compare stuttering between monitors tn 1ms \ ips 4ms.
On a monitor of 240 Hz tn 1 ms, any stuttering is more pronounced compared to 165 Hz ips 4 ms.
(I miss amd + freesync, vrr worked in windowed mode at opengl (osu!). )

[Hadookaan]

It seems to me wrong to compare stuttering between monitors tn 1ms \ ips 4ms.
On a monitor of 240 Hz tn 1 ms, any stuttering is more pronounced compared to 165 Hz ips 4 ms.
(I miss amd + freesync, vrr worked in windowed mode at opengl (osu!). )

I thought that it might be due to the 4ms vs the 1 & 2ms but its not and hes now I know/ ie found out, for instance when I have Gsync on the free sync monitor and lock the fps, to lets say 120fps and the the refresh will be 120hz, and its just as smooth as the Gsync panel unlocked, its when the Freesync panels fps/hz fluctuate I notice its not as smooth, and see these slight micro/stutter's
so I think maybe chief is probably right track, by saying its out of sync slightly 0.5ms that's causing this! not sure if there is a way to fix it, only way is to keep a steady fps without is dropping, by using in-game fps limiter or Riva Tuner Statistics

Chief you said turn off power management is that the Gpu p/m as I have it on prefer max performance on that or the Cpu which is on max performance also under power options.

[1000WATT]

My opinion does not contradict Chief and jorimt. I also believe that there should be a difference.

I think What the Chief said relates to gpu amd, the transitions p state are just awful. Avoid switching the p-state by any means.

Chief can you add bbcode "/ spoiler" for me? People reading my messages will do so at their own peril and risk.

[Synoxia]

Theoretically, there is enough software accuracy to framepace FreeSync as well as GSYNC -- it's just that the programmers of FreeSync drivers don't realize that I can see sub-1-millisecond microstuttering -- as in 0.5ms divergences between gametime and refreshtime (refreh-cycle-emitting-photons). On 60Hz FreeSync, a 1ms microstutter can easily be hidden in the 1/60sec (16.7ms) motion blur, but at 240Hz, a 1ms-error microstutter is a 25% change to a 4ms refresh cycle (1/240sec).

Power management sometimes inject 1ms-2ms errors into Present() timings according to my Tearline Jedi experiments, so I advise turning off power management, this may improve FreeSync framepacing. I find that I am able to get "tens-millisecond-league" precision most of the time in Tearline Jedi, provided I follow a few general best-practices. And framepacing compensation to prioritize gametime sync over rendertime sync. Since the varying rendertimes can diverge gametimes away from refreshtimes (pixels-to-photons) -- creating VRR microstutter.

Regardless....If driver programmers from AMD/NVIDIA are reading this.....framepace those FreeSync refresh cycles properly!

Users of FreeSync: Disable GPU power management to improve FreeSync fluidity slightly.

I hope this won't be considered a necro post but i musk ask!
Do you mean that current hardware g-sync modules are simply smoother than freesync?
What if using a Nvidia gpu with a Freesync monitor, should one disable power management? (Prefer maximum performance in NVCPL?)
Thank you