[ H ]: Freesync 2 (HDR) vs Gsync

[BFG10K]

Blind testing found the two a wash, and nobody thought Gsync was worth $200 more.

https://www.hardocp.com/article/2018/03/30/amd_radeon_freesync_2_vs_nvidia_gsync

 

[jpiniero]

Umm, doesn't the frame rate have to be lowish to see any kind of effect on the *Sync? I guess that kind of kills the point of spending the extra money on the G-Sync monitor but I think it would have been better if they had included a third test with sync disabled.

 

[Mr Evil]

I can understand that it would be difficult to find two similar monitors, but why didn't they use two GPUs that cost the same? That would have brought the performance closer together, and made the monitors the main difference.

p.s. That's a small fortune in silicon they have hanging on their wall.

 

[tamz_msc]

I think the fact that none of the respondents thought that the 200$ difference was worth paying for is the real takeaway from this test.

 

[Konan]

Isn't this the same place that did the "Feels Per Second" (FPS) with the AMD Vega?

 

[ZGR]

A

Umm, doesn't the frame rate have to be lowish to see any kind of effect on the *Sync? I guess that kind of kills the point of spending the extra money on the G-Sync monitor but I think it would have been better if they had included a third test with sync disabled.

No. It is great for 144hz and above too!

 

[BFG10K]

No. It is great for 144hz and above too!

Only if the monitor can actually go higher than 144Hz.

But my personal opinion is that no form of *sync is needed on high refresh rate displays. A standard double-buffered output looks perfectly smooth to me.

 

[ThatBuzzkiller]

Only if the monitor can actually go higher than 144Hz.

But my personal opinion is that no form of *sync is needed on high refresh rate displays. A standard double-buffered output looks perfectly smooth to me.

I'm not sure if adaptive refresh becomes useful anymore either if a user can guarantee a minimum of frametime of 8.3ms (120Hz) with the display's maximum refresh rate being 240Hz (4.1ms) which pretty much translates to a frametime variance of 4.2ms. A refresh rate of 30Hz to 60Hz translates to a frametime variance of 16.6ms ...

The discontinuities in frametime delivery becomes nearly 4x less noticeable in the former compared to the latter ...

Adaptive refresh technology is far more compelling at lower refresh rates to cover the larger gaps in frametime delivery and it is why believe display standards should work towards extending the lower bounds of the refresh rate range to as low as 20Hz since it is arguably low end systems who will benefit the most from this feature!

 

[ZGR]

If I turn off Gsync on my 180hz, I see tearing. I tried 120hz ULMB, and that is tear city.

Is tearing a big deal at 144hz and above? Not as much at 60hz, but it is still very noticeable.

At this point, Freesync needs to be built into everything with Nvidia supporting it.

 

[PhonakV30]

Nvidia G-sync Does not support HDR ??? I thought they do

 

[Muhammed]

Nvidia G-sync Does not support HDR ??? I thought they do

They do, it's just that no GSync monitor support HDR as of yet, GSync monitors with HDR will appear once the BFGD is released. FreeSync 2 monitors with HDR is just starting to appear last month.

 

[bystander36]

If I turn off Gsync on my 180hz, I see tearing. I tried 120hz ULMB, and that is tear city.

Is tearing a big deal at 144hz and above? Not as much at 60hz, but it is still very noticeable.

At this point, Freesync needs to be built into everything with Nvidia supporting it.

At 180hz and high FPS, v-sync is nearly as good and hardly noticeable. And for the majority of people, even without any sync at 180hz, tearing is barely noticeable with a lot of scrutiny and game dependent.

But when two forms of variable sync is used at high FPS, the difference is simply too small. Testing at the lower range would be more interesting at the least.

 

[Jaskalas]

Umm, doesn't the frame rate have to be lowish to see any kind of effect on the *Sync?

First, on my 144hz display, I definitely appreciate having my *Sync active. Screen tearing is a blight on image quality.
Second, my $200 GPU is never going to see the upper end of that frame rate even at 1080p. So it'll pretty much always be "lowish" and within the bounds of the *Sync's active frame rate.
In my experience, *Sync is necessary to achieve better than 30/60 fps.

 

[jpiniero]

Second, my $200 GPU is never going to see the upper end of that frame rate even at 1080p. .

But [H] would, given they used a 1080 Ti and Vega 64.

 

[thetuna]

p.s. That's a small fortune in silicon they have hanging on their wall.

It's not too expensive to buy failed wafers, I have a bunch for decorations as well.
https://www.ebay.com/sch/i.html?_nkw=silicon+wafer

 

[crisium]

Tearing drives me nuts even on my 240hz panel, so I am glad I have G Sync. And it's much more flexible than V Sync with the bonus of less input lag so why not?

Never saw the point in downplaying it like some users do here. Though I understand some are not as sensitive to framerates, tearing, and input lag.

Admittedly some oversell it though. If you average 40-45 frames in VRR, it's not magically as smooth as 60hz V Sync. It's about as smooth as 43hz V Sync, but with less lag, and minor framerate variance is far less troublesome.

 

[Headfoot]

Started using gsync 100% of the time recently. Both freesync and gsync are fantastic features. It's not magic but man it feels pretty close sometimes. On my older monitors I would have to set my settings to never drop below 60fps minimum or I would notice it immediately and it would feel crappy.

I was recently playing total warhammer II everything on ultra at 1440p with a 20 vs 80 units battle and it would regularly dip down to 40ish FPS even on my 1080ti OC. But there was no stuttering on zoom-in/zoom-out which I used to get all the time when the FPS jumped up or down. Really nice

 

[Dribble]

Tbh it was a bit of a mess, far too many differences between the systems:
-2 completely different gpu's (performance wise) as 1080Ti is faster then Vega 64.
-2 completely different displays HDR aside, one being samsung VA panel, one being an asus TN panel.
-gsync vs freesync
-HDR vs SDR

Yet tested on only 1 game:
-which was running at high fps therefore reducing the impact of variable sync.
-which works well with HDR. HDR to work well requires a good game implementation and many games don't have that looking faded and actually worse then SDR in many peoples eyes. The overblown colours used in SDR actually make the game pop, where as using HDR to be more photo realistic can look more real but equally more boring.

Finally, it was a subjectively scored with far too small a number of inputs (10) for any sort of accuracy.

So not sure what it was trying to compare - variable sync quality of gsync vs freesync, VA vs TN, HDR vs SDR, samsung panels vs asus, 1080Ti vs Vega 64. They could have picked any of those but by picking them all at once it really didn't compare anything other then complete system A vs complete system B.

 

[zlatan]

Nvidia G-sync Does not support HDR ??? I thought they do

G-Sync allows HDR, but not support it directly. FreeSync 2 is not just allows it, but also enables the game to query the monitor characteristics. After that it adapt the output to the contrast and gamut of the display. This is a huge advantage for the gamers.

 

[Despoiler]

Only if the monitor can actually go higher than 144Hz.

But my personal opinion is that no form of *sync is needed on high refresh rate displays. A standard double-buffered output looks perfectly smooth to me.

I agree with this and it's also my experience. I have a 120hz 1080p display that can do black frame insertion/strobing. Black frames make an LCD look like a CRT. It removes LCD persistence, which started really screwing me up several years ago. I would lose my character or cursor in complex scenes when fighting in say a MOBA. Black frame insertion was a savior and I simply will not consider a monitor without it. Keeping with 1080p resolution with a good GPU keeps the FPS @ 120 or more. *Sync technologies have no use in my personal display strategy.

As far as the article, it's good to see media validation of AMD's point about these technologies. If they are more or less equal that should be common knowledge based on the testing.

 

[Muhammed]

G-Sync allows HDR, but not support it directly.

GSync supports HDR fully, it's just that GSync displays with HDR is not available in the market. They will be next month:

Update March 22, 2018: Nvidia have again confirmed that their 27-inch G-Sync HDR panels will start to arrive on the shelves in Q1 of this year - and that means we'll have the first batch of 16:9 panels by the end of next month.

https://www.pcgamesn.com/nvidia-gsync-hdr-release-date

 

[zlatan]

GSync supports HDR fully...

Than please help me where can I find the the tone mapping library for G-Sync.

 

[Muhammed]

Than please help me where can I find the the tone mapping library for G-Sync.

Why would any one need that? NVIDIA has a separate module that handles everything for the display. Through a combination of driver and module interference they achieve the same thing. Each module is closely integrated with it's specific display and NVIDIA handpicks these displays thoroughly. In fact this process is much more straightforward than AMD's FreeSync 2 which needs a specific API and the support of the game/ video player to handle everything.

 

[zlatan]

Why would any one need that? NVIDIA has a separate module that handles everything for the display. Through a combination of driver and module interference they achieve the same thing. Each module is closely integrated with it's specific display and NVIDIA handpicks these displays thoroughly. In fact this process is much more straightforward than AMD's FreeSync 2 which needs a specific API and the support of the game/ video player to handle everything.

Every HDR display has a built-in tone mapping module, but it don't know how to apply the tone map properly. I have a ROG Swift PG27UQ for testing the HDR, and the results are very bad. This is why I want to know the monitor characteristics, to allow a usable tone mapping before the HDR transport, and not lose a lot of detail, because of the built-in hardware. I can do this for Freesync 2, and I want to do this for G-Sync as well, but I can't, so my G-Sync custumers will have worse image quality and I can't help them.

 

[Muhammed]

Every HDR display has a built-in tone mapping module, but it don't know how to apply the tone map properly. I have a ROG Swift PG27UQ for testing the HDR, and the results are very bad. This is why I want to know the monitor characteristics, to allow a usable tone mapping before the HDR transport, and not lose a lot of detail

Consoles and current PC games don't need that when dealing with HDTVs or HDR monitors, they work flawlessly. So what changed?

so my G-Sync custumers will have worse image quality and I can't help them

What sort of difference are we talking about?

because of the built-in hardware. I can do this for Freesync 2

So FreeSync 2 displays has a built in hardware just like GSync now? Is it proprietary as well?