Samsung C34F791

The video below shows this text test, a mixed desktop background and dark desktop background from a variety of viewing angles. You can see a loss of perceived contrast and general lightening of the image from more extreme viewing angles, but vertically in particular performance is stronger than TN (Twisted Nematic) panels as there is no colour inversion or more extreme colour shifts. The final third of the video highlights the ‘VA glow’ mentioned previously, which blooms out from extreme viewing angles but is not never as obvious as ‘IPS glow’ regardless of viewing angle.

Responsiveness

Input lag

Using a sensitive camera and small tool called SMTT 2.0, we compared the latency of the C34F791 to a range of monitors of known latency. Over 30 repeat readings were taken to help enhance accuracy. Using this method, we measured 16.86ms (over 1.5 frames @100Hz) of input lag. The ‘Game Mode’ and ‘Response Time’ setting did not have a significant impact on this value. Note that this value is influenced both by the element of input lag you see (pixel responsiveness) and the element you feel feel (signal delay). It indicates a moderate signal delay that sensitive users might notice, but most users should find fine. We don’t have the equipment to accurately measure input lag whilst FreeSync is active, unfortunately.

Perceived blur (pursuit photography)

In this article we explore the factors affecting monitor responsiveness. One of the key concepts explored here is ‘perceived blur’, influenced by both what the monitor is doing (i.e. pixel responsiveness) and by eye movement as you track motion on the screen. We also introduce a photography technique called ‘pursuit photography’ which captures motion on a monitor in a way that reflects not only what the screen is doing (pixel responsiveness) but also blur due to eye movement. And this second factor is often the most significant on modern monitors. The following images are pursuit photographs taken using the UFO Motion Test for ghosting, with the test running at its default speed of 960 pixels per second. This is a good practical speed for such photography and reflects both elements of perceived blur nicely.

The monitor was tested at all ‘Response Time’ settings, at both 60Hz and 100Hz. Note that the UFO in this test moves across the screen from left to right at a frame rate matching the refresh rate of the screen. All three rows of the test are analysed (i.e. a dark, medium and light cyan background) so that a decent range of transitions between various shades are shown. We’ve also included a reference shot at both 60Hz and 100Hz, specifically the Dell S2417DG running the test with the medium cyan background. This shows how things should look where pixel response time isn’t really a limiting factor and only eye movement is really affecting perceived blur. Any interlacing patterns shown here are simply moiré from the camera rather than any artifacts being shown by the monitor itself.



At 60Hz the object itself (UFO) appears soft and blurry, reflecting a high level of perceived blur due to eye (camera) movement. Using the ‘Standard’ setting for ‘Response Time’, there was a fair degree of trailing behind the object for the dark background (first image), which is trailing from slower than optimal pixel response times. The medium cyan background showed much fainter trailing, but also a little bit of overshoot (a light ‘halo’ trailing which is quite faint in this example). For the light background there was a further reduction in trailing, although the overshoot was a bit more pronounced now. Bumping up to ‘Faster’ had very little effect in this test. The ‘Faster’ setting gave strong overshoot for the dark background, as you can see an inky trail behind the UFO. The overshoot was also just slightly increased for the medium background and increased fairly significant for the light background, with an obvious halo.

At 100Hz, the object appeared more sharply focussed and narrower, owing to a significant reduction in perceived blur from eye (camera) movement. Using the ‘Standard’ setting gave a fairly bold and deep trail behind the UFO where the dark background was used. This indicates slower than optimal pixel transitions – it may appear bolder than the trailing at 60Hz, but if you look carefully at the image you can see a fringe of this bold trailing near the rear of the object at 60Hz as well. The trail is simply more pronounced at 100Hz due to the natural increase in refresh rate and frame rate for the test. The trailing was fainter with the medium cyan background and fainter still for the light cyan background. For the light background there was also a bit of overshoot visible. The ‘Faster’ setting provided only subtle changes, with perhaps a very slight decrease in trailing for the medium background and a touch more overshoot for the light background. The ‘Fastest’ setting introduced some overshoot for the dark background, giving it a dirty appearance, and introduced a little for the medium background as well. The overshoot levels were further increased for the light background.

From this analysis, it would be tempting to conclude that ‘Faster’ is the optimal setting. However; whilst observing a broader range of pixel transitions, we noticed that overshoot became quite a bit more widespread with this setting, particularly where lighter shades were involved in the transition. We found this slightly distracting at times and didn’t feel that the reduction in conventional trailing was at all noticeable. We therefore settled for the ‘Standard’ setting, giving the least amount of overshoot without any real drawbacks elsewhere.

Interlace pattern artifacts

On some monitors artifacts can be observed when viewing content moving on the screen, particularly for high refresh rate screens. These artifacts can appear as a mesh or interlaced or woven lines of an alternating shade that is lighter or darker than intended; hence the name ‘interlace pattern artifacts’. Such artifacts were not present on this model during motion. There were very faint static interlacing patterns in places if you looked closely, with the horizontal bands alternating between a very slightly lighter and darker version of the intended shade. These aren’t worth worrying about in our view, though, as they were too faint to notice when gaming and aren’t something most users would notice during general use anyway.

Responsiveness in games and movies

The monitor gave a competent 100Hz experience on Battlefield 1 (BF1), where the frame rate kept pace with the refresh rate. The increased refresh rate provided an improvement to the ‘connected feel’ and that feeling or precision and ‘snappiness’ when interacting with the character. This was something that no 60Hz monitor can capture, regardless of how low its input lag may be. This is due to the fact the monitor is outputting 1.67 times as much information every second. This increased refresh rate and frame rate also provided significant benefits in reducing perceived blur, with the game world appearing sharper and more detailed during movement compared to a 60Hz display. With both factors combined, the playability of this game and our ability to be competitive in multiplayer was greatly enhanced compared to gaming on a 60Hz model. Most of the pixel transitions were also sufficiently fast to provide a really decent 100Hz experience, too. This was a far cry from older VA models, even those with high refresh rates such as the Acer Z35. Rather than obvious widespread weaknesses in pixel responsiveness, giving a smeary and at times quite messy experience overall, most pixel transitions were performed at a good speed. There was some extra trailing in places for ‘common transitions’ between light and medium shades, but this was light and powdery in nature rather than heavy and smeary. It didn’t really catch our eye or affect the playability of the game. There was also a little bit of overshoot in places, particularly where the lightest shades were involved, but this was quite light and not obtrusive to our eyes.

There were some weaknesses apparent where darker transitions were involved, however. This is something of an inevitability with VA panels, but the degree of weakness here does vary. For this model, things were actually rather good for a VA panel even where the weakest transitions were present. There were no obvious extended ‘smoke-like’ smeary trailing and no obvious look of wet ink being smeared across the screen. There was some ‘break-up’ trailing where dark shades with a slight hue (such as rich brown earth in the darkness or dark rocks with a slight blue-grey hue to them) appeared with a more colourful trail behind them than you might expect. There was also a touch of overshoot added to the mix for some of these transitions involving dark shades, although it was difficult to distinguish between this and ‘break-up’ trailing on this model. Rather than confuse you with further descriptions of these weaknesses, we’ll give you a video-based run-through instead. This video is actually for the ASUS MX34VQ, which is a touch more responsive than the Samsung using its optimal pixel overdrive setting. The weaknesses highlighted here as they would manifest in a video are generally quite similar, however.

On Dirt Rally the ‘connected feel’ was again excellent on this 100Hz display, where the frame rate kept up. Whilst we didn’t feel this was as beneficial to our gaming experience as on FPS titles like BF1, that is probably more down to our skill (or lack of) on this game. For more competitive racing fans, having a good ‘connected feel’ between the controller and the car in the game could certainly be of great benefit. The perceived blur was also greatly reduced compared to the more restricted frame rates offered by a 60Hz monitor, with the racing environment, tracks and competitors all appearing sharper and more ‘focussed’ whilst racing. The same weaknesses observed on BF1 were apparent here, although most transitions were again very snappy and provided a solid 100Hz experience. Some of the more obvious weaknesses became apparent when driving at night, as this introduced some of the slowest ‘high contrast’ transitions involving particularly dark shades. Even then, it was more of a case of ‘slightly extended trailing’ rather than the obvious smearing that many VA models suffer from. Even those with high refresh rates, such as the Acer Z35. There was particularly obvious smearing on the Acer for some of the pixel transitions between very dark greys and slightly lighter greys, such as mountains or trees silhouetting the slightly brighter night sky. On the Samsung things were far more constrained in such scenarios. And there was no obvious overshoot, either.

To finish off, we tested our Blu-ray test titles. There were no obvious weaknesses observed from issues with the monitor’s pixel responsiveness, such as trailing or overshoot. The 24 -30fps or so at which these titles run is really the main limiting factor in terms of their fluidity and the requirements in terms of pixel responsiveness is significantly lower than running higher frame rate content such as games. The 100Hz refresh rate didn’t provide any benefit given the low frame rate, either.

FreeSync – the technology and activating it

AMD FreeSync is a variable refresh rate technology supported by certain AMD GPUs (including the GPUs listed here, as well as future models). As with Nvidia G-SYNC, an alternative variable refresh rate technology for Nvidia users, the technology is designed to allow the monitor to dynamically adjust its refresh rate to match the frame rate of the content. In doing so, there is an elimination of the stuttering (VSync on) or tearing and juddering (VSync off) that would generally occur if the refresh rate and frame rate was mismatched. In addition, there is reduced input lag compared to running with VSync enabled – although aren’t able to accurately measure this. Unlike Nvidia G-SYNC, which uses a proprietary ‘G-SYNC board’ integrated into the monitor, AMD FreeSync makes use of VESA Adaptive-Sync. This functionality is integrated into the DP 1.2a or HDMI 1.4a (and above) port controllers of some monitors, including the C34F791.

Our test system uses a Club3D Radeon R9 290 royalAce FreeSync-compatible GPU. This GPU uses DP 1.2a, where FreeSync worked as expected, but lacks HDMI 2.0 so we were unable to properly test functionality via HDMI. FreeSync was automatically activated when we first hooked the display up to the GPU. That should generally be what happens, provided you also have ‘FreeSync’ enabled in the ‘System’ section of the monitor’s OSD. There are two options to select that will enable the technology here; ‘Standard Engine’ and ‘Ultimate Engine’ as we explore shortly. To make sure everything is also set up correctly in the graphics driver, open ‘AMD Radeon Settings’ and click ‘Display’. You should then ensure that the first slider, ‘AMD FreeSync’, is set to ‘On’. If you hover over this, it will also report the variable refresh rate display supported by the display. Note that the image below is for a different monitor and is just used here for reference.

The variable refresh rate range supported by the monitor is 48 – 100Hz with ‘Ultimate Engine’ selected (80 – 100Hz with ‘Standard Engine’ selected). Above the refresh rate ceiling for FreeSync (100Hz, corresponding to 100fps in a game), the GPU respects your selection of ‘VSync on’ or ‘VSync off’ in the graphics driver. With ‘VSync on’ the frame rate will not rise beyond 100fps, at which point VSync activates, bringing with it the usual latency penalty when it does so. With ‘VSync off’ the frame rate is free to rise above 100fps, but the monitor will simply stay at 100Hz and therefore tearing and juddering will occur. VSync is configured in the ‘Gaming’ section of ‘Radeon Settings’, where it is referred to as ‘Wait for Vertical Refresh’. This can either be set globally under ‘Global Settings’, or independently for each game title. The default is ‘Off, unless application specifies’ which means that VSync will only activate if it is selected in the game itself (if there is such an option, which there usually is). The in-game option may be referred to as ‘sync every frame’ or something along those lines rather than ‘VSync’ or ‘Vertical Sync’. Most users will probably wish to enable VSync when using FreeSync to ensure that they don’t get any tearing. One of the final two options in the list, shown in the image below, would therefore be preferable.

Below the floor of FreeSync operation (48Hz), the monitor employs AMD LFC (Low Frame Rate Compensation) when FreeSync is active, regardless of the VSync setting used. This is only the case when ‘Ultimate Engine’ is selected, as LFC requires the refresh rate ceiling to be at least 2x the refresh rate floor of operation (48 – 100Hz satisfies this). The documentation linked to is outdated in that it states the maximum must be at least 2.5x the minimum, which is no longer the case. With LFC, the refresh rate will try to stick to multiples of the frame rate when the frame rate dips below 48fps. So if for example the game was running at 35fps, the monitor would run at 70Hz. This is effective in combatting the tearing or stuttering that would occur if the monitor simply stayed at a static refresh rate matching its floor or ceiling (i.e. 48Hz or 100Hz). If ‘Standard Engine’ is used, the monitor respects your selection of VSync on or VSync off below the 80fps (80Hz) floor when using this option, so you get stuttering or tearing, respectively.

The monitor will say ‘FreeSync’ at the bottom of the ‘Information’ section of the OSD if it’s connected up to a compatible GPU and the option is enabled in the OSD. There is no indication of the monitor that FreeSync is actually active and being used, only that it has been ‘enabled’. It should be obvious from the behaviour in a game when it is or isn’t being used, though, and provided everything is set up as above there shouldn’t be any issues. The final thing to bear in mind that FreeSync can only remove stuttering or juddering from frame and refresh rate mismatches, it can’t compensate for other interruptions to smooth gameplay such as network latency or insufficient system memory.

FreeSync – the experience

We tested a range of game titles including; Battlefield 1 (BF1), Elder Scrolls Online (ESO), Star Wars Battlefront (SWBF), Hitman, Mass Effect Andromeda and various indie titles using FreeSync. For the most part the technology worked as we would expect on these titles, but we did observe some flickering at certain frame rates. On these titles the screen would flicker like crazy between 45 – 53fps in-game (45 – 53Hz), but was fine either above or below this level. Some titles may flicker at other frame rates with FreeSync active as well, and we are aware of users reporting this on some titles. We completely understand the frustration with this issue, but it isn’t restricted to the CF791 and is actually a broader issue. Similar observations have been made on the ASUS MX34VQ and for some titles various 16:9 FreeSync models as well. This is an issue that’s up to AMD to address with an improved driver really. For reference, we were using AMD Crimson ReLive Edition 17.2.1 and as mentioned earlier had the Samsung monitor driver installed on a system running Windows 10.

For frame rates and refresh rates outside of the range affected by flickering, we found FreeSync did its thing and performed much as we would hope for given our past experiences with the technology. We will simply be focussing on BF1 for this analysis as the end result was the same on all of the titles we tested. Furthermore, this title has the flexibility with graphics settings to be run at a full range of different frame rates on our Radeon 290-based test system. With our usual settings employed the frame rate fluctuated quite readily between around 50 -100fps, although generally stayed around 70 – 80fps. With FreeSync disabled and the monitor running a static 100Hz, these fluctuations resulted in tearing and juddering (VSync off) or stuttering (VSync on). To us these imperfections were obvious, although sensitivity to such things does vary. With FreeSync enabled the game just ‘flowed’ better, not interrupted by such nuisances.

As frame rate dropped away from 100fps the level of perceived blur increased and ‘connected feel’ worsened. This was the case even with FreeSync enabled; it can’t compensate for reduced frame rates and make them feel like higher frame rates. It can simply smooth out the experience by removing the tearing and stuttering that would otherwise occur. As frame rates approached the ‘flicker zone’ on BF1 and therefore dropped comfortably (or uncomfortably) below 60fps, it wasn’t really a nice place to be. Not just because of the looming threat that flickering might kick in, but also because of the massive drop in the ‘connected feel’ and considerable increase in perceived blur. We also observed a bit more overshoot as frame rate dropped, which is something we commonly observe on FreeSync models. This is an advantage of G-SYNC – the pixel overdrive is tuned for different refresh rates individually and ‘slackened off’ at lower refresh rates where such rapid pixel responses aren’t required and overshoot can become more of an eyesore. Having said that, although the overshoot was somewhat more noticeable at lower vs. higher refresh rates, it was not really particularly obtrusive or widespread really.

The 34” 3440 x 1440 curved ‘UltraWide’ experience

This article takes a detailed look at the 3440 x 1440 resolution and the sort of experience it provides on the desktop, when playing games and when watching movies. We won’t be repeating too many of the fundamentals of this here and instead would urge you to read that article. One notable difference with the Samsung, though, is that it has a fairly steep curve to the screen at 1500R. This is twice as steep as this model’s predecessor, the S34E790C. It’s also slightly steeper than the curvature of models like the ASUS MX34VQ, and AOC AG352UCG. When we first started using the screen, we’d just come from a range of flat screens and certainly noticed the curve. The steepness of the curve was quite evident when looking at the taskbar on the desktop, for example. Even then, the experience did not feel unnatural and we didn’t feel things looked odd or out of place. Designers and other users who require geometrically correct shapes will of course have to be very cautious with this, but for general use it wasn’t an issue. We also found our eyes quickly adjusted to this and before long we felt right at home using the 1500R curved screen. We can’t exactly back up Samsung’s claims that the viewing experience is made more comfortable by this sort of curvature as it creates a more uniform viewing distance between the centre and edges of the screen. But that’s just because we don’t really find flat screens inherently uncomfortable. And we certainly didn’t find this one uncomfortable – and perhaps reading text at the edges of the screen was a bit easier. The images below are just for illustrative purposes and make the curve appear quite exaggerated and more noticeable than it is when you actually use the monitor.

For gaming we found that the curve gave an extra feeling of depth and immersion, complementing the 21:9 aspect ratio nicely. The enhanced Field of View (FOV) afforded to most game titles (including BF1) was also appreciated, which is something covered by our 3440 x 1440 experience article. Compared to the marginally less steep curve of models like the ASUS MX34VQ (1800R), the 1500R curve made a subtle but noticeable difference in this respect. Just as we’ve found with increasing curvature from the original 3000R screens to around 2000R, the difference is quite subtle once you get used to it, but something we quite liked. We know users don’t always like change and some see the curve as a ‘gimmick’, but generally once those users see it for themselves and actually use the monitor for a bit they learn to rather like the curve. With that said, again note that the images below (of Mass Effect Andromeda) exaggerate the curve and in no way reflect how the monitor looks in person in terms of image quality either. They are just for illustrative purposes.

Interpolation and upscaling

This monitor has an integrated scaler and provides two options in the OSD (On Screen Display) which affect the handling of non-native resolutions; ‘Auto’ and ‘Wide’. The ‘Auto’ options maintains the aspect ratio of the resolution you have selected, so 2560 x 1440 (WQHD) is displayed with large black borders at the side. The image itself appears undistorted and is displayed with 1:1 pixel mapping, in this case just as a 27” 2560 x 1440 display. This is therefore a useful fall-back for titles that don’t support 21:9. For 1920 x 1080 (Full HD), the monitor again maintains the 16:9 aspect ratio and displays the image using interpolation, again using up 2560 x 1440 pixels on the screen. Due to the interpolation process, there is a degree of softening compared to viewing 1920 x 1080 natively on a 27” screen. It isn’t too severe really and certainly doesn’t give the image as much of a soft look as some screens do when using interpolation. Some users may also like to use ‘MagicUpscale’ to increase sharpness, although we found even the weaker of the two settings (‘Mode 2’) overly sharp and artificial-looking.

The ‘Wide’ option ensures that all pixels on the screen are illuminated, ignoring the aspect ratio of the source resolution. The aforementioned 16:9 aspect ratio resolutions therefore appear stretched in an obvious way, because that is exactly what happens to the image. The monitor can also display 2560 x 1080 using interpolation, using the full screen space and keeping the 21:9 aspect ratio. It does this regardless of whether ‘Auto’ or ‘Wide’ is selected. This uses considerably less GPU horsepower than 3440 x 1440, so might be useful for some of the more demanding game titles. The interpolation process again gives a bit of softening compared to a 34” native 2560 x 1080 display, but it’s actually rather good overall. It’s certainly one of the better models we’ve seen for displaying 2560 x 1080 non-natively and we feel this could be a useful option for some users.

Conclusion

The Samsung C34F791 builds on some of the strengths of the S34E790C, featuring a newer VA panel with similar strengths in contrast but various changes and improvements such as a steeper curve, more generous colour gamut and superior refresh rate. Out of the box the CF791 offered an impressive image setup, with a well-balanced white point and good gamma handling. Gamma did not track perfectly at the ‘2.2’ curve that is often desired, but it came close by tracking at ‘2.1’ on average. And with the generous colour gamut afforded by the Quantum Dot backlight, vibrancy was not in short supply. Colours were rich and lively with strong variety. Colour consistency was also pleasing given the panel type, perhaps enhanced slightly by the curvature, although a little saturation was still lost towards the edges and bottom of the screen. Overall, the image was a touch more vibrant than on other recent UltraWides we’ve tested, such as the ASUS MX34VQ and AOC AG352UCG. But this was not a massive difference really, just a slight edge in a department many users will likely admire. Fortunately it didn’t impart the unnatural look of an untamed wide gamut (Adobe RGB+) model displaying sRGB content. Things still looked natural overall rather than completely overblown.

The contrast performance was strong, as you’d hope for given the panel type. Although we were unable to quite achieve contrast matching the specified 3000:1, we did get plenty of readings around 2500:1 and a bit above. And that gave a distinctly ‘VA’ look to the image with respectable depth and strong distinction between bright and dark shades. As is usual for a VA model there was a degree of ‘black crush’, although this was as low as we’ve seen on a model of this size. The overall atmosphere created in dark scenes when gaming or watching movies was very good, far superior to what any non-VA model could produce. The screen surface didn’t mar lighter shades with obtrusive graininess, either. There was just a touch of light graininess rather than an unsightly smear of heavy grain. In this respect the monitor was similar to the other VA UltraWides we’ve looked at recently and indeed quite similar to its predecessor.

One area in particular where it is unreasonable to expect perfection from the VA panel type is responsiveness. Although perfection was not achieved, which should come as no shock, the monitor did provide a competent experience that made very good use of its maximum supported refresh rate of 100Hz. Overall, pixel responses were good enough to drive 100fps @ 100Hz comfortably, with most weaknesses resulting in a bit of extra ‘powdery’ trailing rather than the extensive smeary trailing that some VA models produce. The monitor was free from obtrusive overshoot, at least in the ‘Standard’ setting which we actually favour for this reason. FreeSync also worked to combat the tearing and stuttering associated with mismatches between frame rate and refresh rate. As we often observe with the technology, overshoot was a bit more noticeable as refresh rate decreased. But was never too much of an eyesore. The more obvious visual distractions came from some unmistakable flickering, which affected the lower end of the FreeSync range on the titles we tested (45 – 53Hz or 45 – 53fps in a game with FreeSync active). This aside, we always found the experience on this monitor most enjoyable at higher frame rates and feel FreeSync is a nice bonus rather than a defining feature of the monitor for us. Indeed, our greatest enjoyment during multiplayer gaming came from our more powerful Nvidia GPU. This could could deliver a solid 100fps without requiring any drastic image-quality destroying graphical tweaks on titles like Battlefield 1.

It would also be improper not mentioning the curve of the monitor. This is the strongest curve we’ve experienced on any monitor, somewhat steeper than the 1800R – 2000R of competing UltraWides and significantly steeper than the 3000R curve of this model’s predecessor. We felt this provided a bit of extra depth to the experience, but was not actually as noticeable or distracting as you might think. Especially given how steep it looks in some images. Once you’re sitting in front of the monitor using it, it is something you soon adapt to and it becomes a nice part of the experience rather than this terrifying and odd feature. Overall we feel this is a well-refined monitor that builds on the strengths of its predecessor to deliver a convincing and immersive 100Hz gaming experience.

The bottom line; a monitor that delivers a rich and engaging gaming experience, making surprisingly good use of its 100Hz refresh rate.

 

Donations are also greatly appreciated.