Introduction to FreeSync and Adaptive Sync

The first time anyone talked about adaptive refresh rates for monitors – specifically applying the technique to gaming – was when NVIDIA demoed G-SYNC back in October 2013. The idea seemed so logical that I had to wonder why no one had tried to do it before. Certainly there are hurdles to overcome, e.g. what to do when the frame rate is too low, or too high; getting a panel that can handle adaptive refresh rates; supporting the feature in the graphics drivers. Still, it was an idea that made a lot of sense.

The impetus behind adaptive refresh is to overcome visual artifacts and stutter cause by the normal way of updating the screen. Briefly, the display is updated with new content from the graphics card at set intervals, typically 60 times per second. While that’s fine for normal applications, when it comes to games there are often cases where a new frame isn’t ready in time, causing a stall or stutter in rendering. Alternatively, the screen can be updated as soon as a new frame is ready, but that often results in tearing – where one part of the screen has the previous frame on top and the bottom part has the next frame (or frames in some cases).

Neither input lag/stutter nor image tearing are desirable, so NVIDIA set about creating a solution: G-SYNC. Perhaps the most difficult aspect for NVIDIA wasn’t creating the core technology but rather getting display partners to create and sell what would ultimately be a niche product – G-SYNC requires an NVIDIA GPU, so that rules out a large chunk of the market. Not surprisingly, the result was that G-SYNC took a bit of time to reach the market as a mature solution, with the first displays that supported the feature requiring modification by the end user.

Over the past year we’ve seen more G-SYNC displays ship that no longer require user modification, which is great, but pricing of the displays so far has been quite high. At present the least expensive G-SYNC displays are 1080p144 models that start at $450; similar displays without G-SYNC cost about $200 less. Higher spec displays like the 1440p144 ASUS ROG Swift cost $759 compared to other WQHD displays (albeit not 120/144Hz capable) that start at less than $400. And finally, 4Kp60 displays without G-SYNC cost $400-$500 whereas the 4Kp60 Acer XB280HK will set you back $750.

When AMD demonstrated their alternative adaptive refresh rate technology and cleverly called it FreeSync, it was a clear jab at the added cost of G-SYNC displays. As with G-SYNC, it has taken some time from the initial announcement to actual shipping hardware, but AMD has worked with the VESA group to implement FreeSync as an open standard that’s now part of DisplayPort 1.2a, and they aren’t getting any royalties from the technology. That’s the “Free” part of FreeSync, and while it doesn’t necessarily guarantee that FreeSync enabled displays will cost the same as non-FreeSync displays, the initial pricing looks quite promising.

There may be some additional costs associated with making a FreeSync display, though mostly these costs come in the way of using higher quality components. The major scaler companies – Realtek, Novatek, and MStar – have all built FreeSync (DisplayPort Adaptive Sync) into their latest products, and since most displays require a scaler anyway there’s no significant price increase. But if you compare a FreeSync 1440p144 display to a “normal” 1440p60 display of similar quality, the support for higher refresh rates inherently increases the price. So let’s look at what’s officially announced right now before we continue.

FreeSync Displays

There are four FreeSync displays launching today, one each from Acer and BenQ, and two from LG. Besides the displays launching today, seven additional displays should show up in the coming weeks (months?). Here’s the current list of FreeSync compatible displays, with pricing where it has been disclosed.

The four displays launching today cover two primary options. For those that want higher refresh rates, Acer and BenQ have TN-based 40-144Hz displays. Both are 27” WQHD displays, so it’s quite probable that they’re using the same panel, perhaps even the same panel that we’ve seen in the ASUS ROG Swift. The two LG displays meanwhile venture out into new territory as far as adaptive refresh rates are concerned. LG has both a smaller 29” and a larger 34” 2560x1080 (UW-UXGA) display, and both sport IPS panels (technically AU Optronics' AHVA, but it's basically the same as IPS).

The other upcoming displays all appear to be using TN panels, though it's possible Samsung might offer PLS. The UE590 appears to be TN for certain, with 170/160 degree viewing angles according to DigitalTrends. The UE850 on the other hand is targeted more at imaging professionals, so PLS might be present; we'll update if we can get any confirmation of panel type.

One of the big benefits with FreeSync is going to be support for multiple video inputs – the G-SYNC displays so far are all limited to a single DisplayPort connection. The LG displays come with DisplayPort, HDMI, and DVI-D inputs (along with audio in/out), and the Acer is similarly equipped. Neither one has any USB ports, though the BenQ does have a built-in USB hub with ports on the side.

Our testing was conducted on the 34UM67, and let me just say that it’s quite the sight sitting on my desk. I’ve been bouncing between the ASUS ROG Swift and Acer XB280HK for the past several months, and both displays have their pros and cons. I like the high resolution of the Acer at times, but I have to admit that my aging eyes often struggle when running it at 4K and I have to resort to DPI scaling (which introduces other problems). The ASUS on the other hand is great with its high refresh rates, and the resolution is more readable without scaling. The big problem with both displays is that they’re TN panels, and having come from using a 30” IPS display for the past eight years that’s a pretty painful compromise.

Plopping the relatively gigantic 34UM67 on my desk is in many ways like seeing a good friend again after a long hiatus. “Dear IPS (AHVA), I’ve missed having you on my desktop. Please don’t leave me again!” For the old and decrepit folks like me, dropping to 2560x1080 on a 34” display also means reading text at 100% zoom is not a problem. But when you’re only a couple feet away, the relatively low DPI does make the pixels much more visible to the naked eye. It even has built-in speakers (though they’re not going to compete with any standalone speakers in terms of audio quality).

The launch price of $649 is pretty impressive; we’ve looked at a few other 21:9 displays in the past, and while the resolution doesn’t match LG’s 34UM95, the price is actually $50 less than the LG 34UM65’s original $699 MSRP (though it’s now being sold at $599). So at most, it looks like putting in the new technology to make a FreeSync display costs $50, and probably less than that. Anyway, we’ll have a full review of the LG 34UM67 in the coming weeks, but for now let’s return to the FreeSync discussion.

Pricing vs. G-SYNC

It certainly appears that AMD and their partners are serious about pricing FreeSync aggressively, though there aren’t direct comparisons available for some of the models. The least expensive FreeSync displays start at just $449, which matches the least expensive G-SYNC display (AOC G2460PG) on price but with generally better specs (29” 2560x1080 and IPS at 75Hz vs. 24” 1920x1080 TN at 144Hz). Looking at direct comparisons, the Acer XG270HU and BenQ XL2730Z are WQHD 144Hz panels, which pits them against the $759 ASUS ROG Swift that we recently reviewed, giving FreeSync a $160 to $260 advantage. As AMD puts it, that’s almost enough for another GPU (depending on which Radeon you’re using, of course).

Based on pricing alone, FreeSync looks poised to give G-SYNC some much needed competition. And it’s not just about the price, as there are other advantages to FreeSync that we’ll cover more on the next page. But for a moment let’s focus just on the AMD FreeSync vs. NVIDIA G-SYNC ecosystems.

Right now NVIDIA enjoys a performance advantage over AMD in terms of GPUs, and along with that they currently carry a price premium, particularly at the high end. While the R9 290X and GTX 970 are pretty evenly matched, the GTX 980 tends to lead by a decent amount in most games. Any users willing to spend $200 extra per GPU to buy a GTX 980 instead of an R9 290X might also be willing to pay $200 more for a G-SYNC compatible display. After all, it’s the only game in town for NVIDIA users right now.

AMD and other companies can support FreeSync, but until – unless! – NVIDIA supports the standard, users will be forced to choose between AMD + FreeSync or NVIDIA + G-SYNC. That’s unfortunate for any users that routinely switch between AMD and NVIDIA GPUs, though the number of people outside of hardware reviewers that regularly go back and forth is miniscule. Ideally we’d see one standard win out and the other fade away (i.e. Betamax, HD-DVD, etc.), but with a one year lead and plenty of money invested it’s unlikely NVIDIA will abandon G-SYNC any time soon.

Prices meanwhile are bound to change, as up to now there has been no competition for NVIDIA’s G-SYNC monitors. With FreeSync finally available, we expect prices for G-SYNC displays will start to come down, and in fact we’re already seeing $40-$125 off the original MSRP for most of the G-SYNC displays. Will that be enough to keep NVIDIA’s proprietary G-SYNC technology viable? Most likely, as both FreeSync and G-SYNC are gamer focused more than anything; if a gamer prefers NVIDIA, FreeSync isn’t likely to get them to switch sides. But if you don’t have any GPU preference, you’re in the market for a new gaming PC, and you’re planning on buying a new monitor to go with it, R9 290X + FreeSync could save a couple hundred dollars compared to GTX 970 + G-SYNC.

There's something else to consider with the above list of monitors as well: four currently shipping FreeSync displays exist on the official day of launch, and Samsung alone has five more FreeSync displays scheduled for release in the near future. Eleven FreeSync displays in the near term might not seem like a huge deal, but compare that with G-SYNC: even with a one year lead (more or less), NVIDIA currently only lists six displays with G-SYNC support, and the upcoming Acer XB270HU makes for seven. AMD also claims there will be 20 FreeSync compatible displays shipping by the end of the year. In terms of numbers, then, DP Adaptive Sync (and by extension FreeSync) look to be winning this war.

FreeSync Features

In many ways FreeSync and G-SYNC are comparable. Both refresh the display as soon as a new frame is available, at least within their normal range of refresh rates. There are differences in how this is accomplished, however.

G-SYNC uses a proprietary module that replaces the normal scaler hardware in a display. Besides cost factors, this means that any company looking to make a G-SYNC display has to buy that module from NVIDIA. Of course the reason NVIDIA went with a proprietary module was because adaptive sync didn’t exist when they started working on G-SYNC, so they had to create their own protocol. Basically, the G-SYNC module controls all the regular core features of the display like the OSD, but it’s not as full featured as a “normal” scaler.

In contrast, as part of the DisplayPort 1.2a standard, Adaptive Sync (which is what AMD uses to enable FreeSync) will likely become part of many future displays. The major scaler companies (Realtek, Novatek, and MStar) have all announced support for Adaptive Sync, and it appears most of the changes required to support the standard could be accomplished via firmware updates. That means even if a display vendor doesn’t have a vested interest in making a FreeSync branded display, we could see future displays that still work with FreeSync.

Having FreeSync integrated into most scalers has other benefits as well. All the normal OSD controls are available, and the displays can support multiple inputs – though FreeSync of course requires the use of DisplayPort as Adaptive Sync doesn’t work with DVI, HDMI, or VGA (DSUB). AMD mentions in one of their slides that G-SYNC also lacks support for audio input over DisplayPort, and there’s mention of color processing as well, though this is somewhat misleading. NVIDIA's G-SYNC module supports color LUTs (Look Up Tables), but they don't support multiple color options like the "Warm, Cool, Movie, User, etc." modes that many displays have; NVIDIA states that the focus is on properly producing sRGB content, and so far the G-SYNC displays we've looked at have done quite well in this regard. We’ll look at the “Performance Penalty” aspect as well on the next page.

One other feature that differentiates FreeSync from G-SYNC is how things are handled when the frame rate is outside of the dynamic refresh range. With G-SYNC enabled, the system will behave as though VSYNC is enabled when frame rates are either above or below the dynamic range; NVIDIA's goal was to have no tearing, ever. That means if you drop below 30FPS, you can get the stutter associated with VSYNC while going above 60Hz/144Hz (depending on the display) is not possible – the frame rate is capped. Admittedly, neither situation is a huge problem, but AMD provides an alternative with FreeSync.

Instead of always behaving as though VSYNC is on, FreeSync can revert to either VSYNC off or VSYNC on behavior if your frame rates are too high/low. With VSYNC off, you could still get image tearing but at higher frame rates there would be a reduction in input latency. Again, this isn't necessarily a big flaw with G-SYNC – and I’d assume NVIDIA could probably rework the drivers to change the behavior if needed – but having choice is never a bad thing.

There’s another aspect to consider with FreeSync that might be interesting: as an open standard, it could potentially find its way into notebooks sooner than G-SYNC. We have yet to see any shipping G-SYNC enabled laptops, and it’s unlikely most notebooks manufacturers would be willing to pay $200 or even $100 extra to get a G-SYNC module into a notebook, and there's the question of power requirements. Then again, earlier this year there was an inadvertent leak of some alpha drivers that allowed G-SYNC to function on the ASUS G751j notebook without a G-SYNC module, so it’s clear NVIDIA is investigating other options.

While NVIDIA may do G-SYNC without a module for notebooks, there are still other questions. With many notebooks using a form of dynamic switchable graphics (Optimus and Enduro), support for Adaptive Sync by the Intel processor graphics could certainly help. NVIDIA might work with Intel to make G-SYNC work (though it’s worth pointing out that the ASUS G751 doesn’t support Optimus so it’s not a problem with that notebook), and AMD might be able to convince Intel to adopt DP Adaptive Sync, but to date neither has happened. There’s no clear direction yet but there’s definitely a market for adaptive refresh in laptops, as many are unable to reach 60+ FPS at high quality settings.

FreeSync vs. G-SYNC Performance

One item that piqued our interest during AMD’s presentation was a claim that there’s a performance hit with G-SYNC but none with FreeSync. NVIDIA has said as much in the past, though they also noted at the time that they were "working on eliminating the polling entirely" so things may have changed, but even so the difference was generally quite small – less than 3%, or basically not something you would notice without capturing frame rates. AMD did some testing however and presented the following two slides:

It’s probably safe to say that AMD is splitting hairs when they show a 1.5% performance drop in one specific scenario compared to a 0.2% performance gain, but we wanted to see if we could corroborate their findings. Having tested plenty of games, we already know that most games – even those with built-in benchmarks that tend to be very consistent – will have minor differences between benchmark runs. So we picked three games with deterministic benchmarks and ran with and without G-SYNC/FreeSync three times. The games we selected are Alien Isolation, The Talos Principle, and Tomb Raider. Here are the average and minimum frame rates from three runs:

Except for a glitch with testing Alien Isolation using a custom resolution, our results basically don’t show much of a difference between enabling/disabling G-SYNC/FreeSync – and that’s what we want to see. While NVIDIA showed a performance drop with Alien Isolation using G-SYNC, we weren’t able to reproduce that in our testing; in fact, we even showed a measurable 2.5% performance increase with G-SYNC and Tomb Raider. But again let’s be clear: 2.5% is not something you’ll notice in practice. FreeSync meanwhile shows results that are well within the margin of error.

What about that custom resolution problem on G-SYNC? We used the ASUS ROG Swift with the GTX 970, and we thought it might be useful to run the same resolution as the LG 34UM67 (2560x1080). Unfortunately, that didn’t work so well with Alien Isolation – the frame rates plummeted with G-SYNC enabled for some reason. Tomb Raider had a similar issue at first, but when we created additional custom resolutions with multiple refresh rates (60/85/100/120/144 Hz) the problem went away; we couldn't ever get Alien Isolation to run well with G-SYNC using our custome resolution, however. We’ve notified NVIDIA of the glitch, but note that when we tested Alien Isolation at the native WQHD setting the performance was virtually identical so this only seems to affect performance with custom resolutions and it is also game specific.

For those interested in a more detailed graph of the frame rates of the three runs (six total per game and setting, three with and three without G-SYNC/FreeSync), we’ve created a gallery of the frame rates over time. There’s so much overlap that mostly the top line is visible, but that just proves the point: there’s little difference other than the usual minor variations between benchmark runs. And in one of the games, Tomb Raider, even using the same settings shows a fair amount of variation between runs, though the average FPS is pretty consistent.

Closing Thoughts

It took a while to get here, but if the proof is in the eating of the pudding, FreeSync tastes just as good as G-SYNC when it comes to adaptive refresh rates. Within the supported refresh rate range, I found nothing to complain about. Perhaps more importantly, while you’re not getting a “free” monitor upgrade, the current prices of the FreeSync displays are very close to what you’d pay for an equivalent display that doesn’t have adaptive sync. That’s great news, and with the major scaler manufacturers on board with adaptive sync the price disparity should only shrink over time.

The short summary is that FreeSync works just as you’d expect, and at least in our limited testing so far there have been no problems. Which isn’t to say that FreeSync will work with every possible AMD setup right now. As noted last month, the initial FreeSync driver that AMD provided (Catalyst 15.3 Beta 1) only allows FreeSync to work with single GPU configurations. Another driver should be coming next month that will support FreeSync with CrossFire setups.

Besides needing a driver and FreeSync display, you also need a GPU that uses AMD’s GCN 1.1 or later architecture. The list at present consists of the R7 260/260X, R9 285, R9 290/290X/295X2 discrete GPUs, as well as the Kaveri APUs – A6-7400K, A8-7600/7650K, and A10-7700K/7800/7850K. First generation GCN 1.0 cards (HD 7950/7970 or R9 280/280X and similar) are not supported.

All is not sunshine and roses, however. Part of the problem with reviewing something like FreeSync is that we're inherently tied to the hardware we receive, in this case the LG 34UM67 display. Armed with an R9 290X and running at the native resolution, the vast majority of games will run at 48FPS or above even at maximum detail settings, though of course there are exceptions. This means they look and feel smooth. But what happens with more demanding games or with lower performance GPUs? If you're running without VSYNC, you'd get tearing below 48FPS, while with VSYNC you'd get stuttering.

Neither is ideal, but how much this impacts your experience will depend on the game and individual. G-SYNC handles dropping below the minimum FPS more gracefully than FreeSync, though if you're routinely falling below the minimum FreeSync refresh rate we'd argue that you should lower the settings. Mostly what you get with FreeSync/G-SYNC is the ability to have smooth gaming at 40-60 FPS and not just 60+ FPS.

Other sites are reporting ghosting on FreeSync displays, but that's not inherent to the technology. Rather, it's a display specific problem (just as the amount of ghosting on normal LCDs is display specific). Using higher quality panels and hardware designed to reduce/eliminate ghosting is the solution. The FreeSync displays so far appear to not have the same level of anti-ghosting as the currently available G-SYNC panels, which is unfortunate if true. (Note that we've only looked at the LG 34UM67, so we can't report on all the FreeSync displays.) Again, ghosting shouldn't be a FreeSync issue so much as a panel/scaler/firmware problem, so we'll hold off on further commentary until we get to the monitor reviews.

One final topic to address is something that has become more noticeable to me over the past few months. While G-SYNC/FreeSync can make a big difference when frame rates are in the 40~75 FPS range, as you go beyond that point the benefits are a lot less clear. Take the 144Hz ASUS ROG Swift as an example. Even with G-SYNC disabled, the 144Hz refresh rate makes tearing rather difficult to spot, at least in my experience. Considering pixel response times for LCDs are not instantaneous and combine that with the way our human eyes and brain process the world and for all the hype I still think having high refresh rates with VSYNC disabled gets you 98% of the way to the goal of smooth gaming with no noticeable visual artifacts (at least for those of us without superhuman eyesight).

Overall, I’m impressed with what AMD has delivered so far with FreeSync. AMD gamers in particular will want to keep an eye on the new and upcoming FreeSync displays. They may not be the “must have” upgrade right now, but if you’re in the market and the price premium is less than $50, why not get FreeSync? On the other hand, for NVIDIA users things just got more complicated. Assuming you haven’t already jumped on the G-SYNC train, there’s now this question of whether or not NVIDIA will support non-G-SYNC displays that implement DisplayPort’s Adaptive Sync technology. I have little doubt that NVIDIA can support FreeSync panels, but whether they will support them is far less certain. Given the current price premium on G-SYNC displays, it’s probably a good time to sit back and wait a few months to see how things develop.

There is one G-SYNC display that I’m still waiting to see, however: Acer’s 27” 1440p144 IPS (AHVA) XB270HU. It was teased at CES and it could very well be the holy grail of displays. It’s scheduled to launch next month, and official pricing is $799 (with some pre-orders now online at higher prices). We might see a FreeSync variant of the XB270HU as well in the coming months, if not from Acer than likely from some other manufacturer. For those that work with images and movies as well as playing games, IPS/AHVA displays with G-SYNC or FreeSync support are definitely needed.

Wrapping up, if you haven’t upgraded your display in a while, now is a good time to take stock of the various options. IPS and other wide viewing angle displays have come down quite a bit in pricing, and there are overclockable 27” and 30” IPS displays that don’t cost much at all. Unfortunately, if you want a guaranteed high refresh rate, there’s a good chance you’re going to have to settle for TN. The new UltraWide LG displays with 75Hz IPS panels at least deliver a moderate improvement though, and they now come with FreeSync as an added bonus.

Considering a good display can last 5+ years, making a larger investment isn’t a bad idea, but by the same token rushing into a new display isn’t advisable either as you don't want to end up stuck with a "lemon" or a dead technology. Take some time, read the reviews, and then find the display that you will be happy to use for the next half decade. At least by then we should have a better idea of which display technologies will stick around.