[Anthos]

Hi all,

I currently have an AW2721D that got last year and had it paired with a used 780Ti. My initial plan was at some point to get a 3xxx series GPU but obscene pricing with stock kinda made that not happen. And then early Feb this year my 780Ti.. exploded.. (literally). So then I had to dust off my 610 GT (lol) to be able to use my pc at all (can’t even run native res [lol x2]). Anyway by that point decided to wait until the new gen is out since the prices were still quite high. Aaaand then this thing happened where Nvidia decided to have the whole 4080 fiasco so I ‘ve been thinking the possibility of getting an RDNA 3 (or maybe RDNA 2 if they follow the same route) GPU. Now I’ve been out of the game for quite some time so I am a bit confused on some things. I’ve been trying to read on some stuff and I feel like I am getting contradicting info. To put it simply if I was to pair an AMD GPU with a Gsync ultimate monitor am I leaving something at the table that I wouldn’t with an Nvidia GPU? Some people say that you lose the variable overdrive option, others say differently. And even though I read a fair number of articles about variable overdrive I am still quite confused about it (isn’t the whole point of gsync/freesync that it matches the refresh rate of the screen with the FPS? am I misunderstanding something?). I just feel like I am going around in circles. Heeelp…

p.s (also want to say that the reason I got the AW2721D was not for the Gsync ultimate module but mainly for the AG coating being mild, but since the module exists anyway might as well make the most of it).

[PCM2]

Hi Anthos,

It’s a good question and one which many people are confused by, given the amount of misinformation surrounding this topic. With relatively recent G-SYNC modules, they can be accessed ‘directly’ (with an Nvidia GPU, using G-SYNC) or via Adaptive-Sync for AMD GPUs (‘AMD FreeSync’). You gain all module features and should have a very similar VRR experience with an AMD GPU to what you’d get using an Nvidia GPU, which includes variable overdrive. We actually specifically test and comment on this in our reviews, by the way (some models we’ve looked at recently with G-SYNC modules include the AW3423DW, AG254FG and PG32UQX).

Whilst this is the experience I’ve had in my testing and from feedback I’ve received from others this is usually the case, I should caveat this by saying that the whole testing and certification process is centred around Nvidia GPUs. I’ve seen a small number of reports of this sort of thing on the AW3821DW for example, where people sometimes have it losing signal and VRR seemingly shutting off for no apparent reason with their AMD GPU. Most people seem to run without issue with AMD GPUs. It’s not clear what the exact cause is so all I would say is that some systems and GPU drivers on the AMD side could potentially have some issues which will make the experience a bit flaky.

We also explain the practical purpose of variable overdrive in our reviews and summarise this in our responsiveness article. Basically, the level of pixel overdrive (or grey to grey acceleration) required varies depending on refresh rate. As refresh rate is increased, the pixel response requirements for a good performance increase and the monitor should use use (and can get away with) higher levels of pixel overdrive. So if you have that certain level of pixel overdrive that works well at 144Hz, for example, that same level of pixel overdrive could provide much more noticeable overshoot at 60Hz. The overshoot persists for longer at 60Hz and is therefore much easier to perceive. The monitor needs to tone things down as refresh rate decreases – and that’s exactly what variable overdrive does. And it’s why for monitors which don’t have effective variable overdrive, the user will often have to select a different overdrive setting for lower vs. higher refresh rates if they want the best experience.

[Anthos]

Many thanks for your reply, I greatly appreciate the depth of explanation.

So pretty much the gist as that with both GPUs one receives a similar experience with a V2 module.. but… sometimes with AMD ones hiccups might be expected and you can’t be sure until you try I guess. Seems good enough considering the alternative of definitely having features being incompatible by default.

For the variable overdrive I always feel like there’s a thing or two that I can’t wrap my head around. However I just read your responsiveness article and I think it might have cleared some things up. Then I did some googling and I think I got re-confused.

https://www.reddit.com/r/nvidia/comments/sm4ruv/is_native_gsync_variable_overdrive_supposed_to/

The reason why I started to get info in the first place was because I was having a hard time understanding that if monitors with modules can change the overdrive on the fly then why is there a setting of fast, faster, extreme in the monitor settings? I am confused because that is a manual setting for something that is done automatically if you can understand what I mean. Is that just an “extra” setting for how “aggressively” it applies the variable overdrive? Or is it something different altogether?

I feel like I am being extra dumb but for whatever reason the more I read into it the less I seem to understand about it.

[PCM2]

Overdrive settings are included by monitor manufacturers to suit different preferences and tolerances people have to slower than optimal pixel responses (gives conventional trailing) vs. the overshoot that can replace that. No amount of pixel overdrive tuning can make a natively slow panel ‘fast’ (without obvious overshoot) and it can’t make a reasonably fast panel like that used for the AW2721D ‘visually perfect’ in terms of pixel responsiveness – say, equivalent to an OLED monitor like the AW3423DW. There is still a compromise that needs to be made in terms of either some slower than optimal pixel responses or overshoot (usually a mixture of both). Furthermore, monitor manufacturers often like to include an ‘extreme’ highly aggressive overdrive setting with insane level of overshoot simply so they can give misleading response time specifications for the monitor. So the overdrive setting set by the user dictates a baseline which pushes things one way or the other (slower responses vs. faster responses with more overshoot) regardless of whether variable overdrive is used on top.

[Anthos]

Thanks, that clears things up I think (i ll refrain from googling anything ever again just in case)

Only one more thing that I wanted to ask. With the AW2721D I’ve been using the novideo_srgb tool to clamp the colours to the sRGB spectrum. Is the AMD way visually equal when compared with that tool? Or is one more effective at it than the other? Many thanks.

[PCM2]

Overgoogle syndrome is never pretty. 😉

The Nvidia tool has a few ‘Advanced’ options discussed in the sRGB emulation article which you can’t achieve on the AMD side, such as the ability to target alternative colour spaces including Adobe RGB and DCI-P3. But as noted in the article, they both use the same EDID information so using the defaults on the Nvidia tool (clamps to sRGB) will be equivalent to using the AMD driver tweak (clamps to sRGB using the same data).

[Anthos]

Yeah, I’ve been using that tool ever since it initially popped around. Nifty little tool, I don’t think I would have gotten the monitor if not for it. I still can’t understand how that is not a basic function in the GPU’s control panel though. Oh well. Many thanks, I appreciate all your info. 🙂

[Author]

Posts