Dreame S100 vs Samsung R85H / R86H

The Dreame S100 is a relatively bright television. In movie mode, the brightness we measured reached around 600 nits. This is a really good result, although not the highest when considering the competition at this price point. In fullscreen scenes, it is evident that the television can produce an image that impresses with its brightness. Unfortunately, when more complex scenes appeared on the screen, the local dimming algorithm in this model primarily focused on maintaining the deepest blacks. However, this came at the expense of HDR performance – and in a drastic manner. In such scenes, the brightness of bright elements dropped below 100 nits, which is a frankly tragic result, far from any HDR effect. One can improve this aspect, but not without a "however." It is simply necessary to change the local dimming setting in the television's menu, but then the blacks in such scenes simply become grey. It's a pity that Dreame couldn't better refine the local dimming algorithms, and the S100 is another example that the sheer number of local dimming zones does not always translate to a better image. As a consolation, it is worth adding that the Dreame S100 is advertised as a "QLED" type television, although in reality it uses a PFS LED filter. As a result, the colours are appropriately saturated, and the colour gamut coverage of DCI-P3 we measured exceeded 94%.

When it comes to brightness, the R85H performs really respectably. In most of our test scenes the TV reaches around 800 to 900 nits, and with larger bright elements it can exceed 1000 nits. These are not, of course, levels on a par with the pricier Mini-LEDs, but they are more than enough for watching films and TV shows in HDR, and the effect can already be quite noticeable. The problem reappears in more demanding scenes where there is a small bright element on a very dark background. Here the TV can significantly limit its brightness so as not to light up too large a portion of the screen. In our measurements such small highlights fell even to around 200 to 250 nits, so it is clear that a small number of zones combined with the particular behaviour of the algorithm has consequences here. Fortunately, this is not something that disturbs viewing all the time. In typical mixed HDR scenes the R85H looks good and can deliver appropriate image dynamics. It's just that with very small points of light you shouldn't expect an effect as strong as on TVs with many more zones or on OLEDs.

We tested the Dreame S100 in 'movie' mode and we definitely recommend using this setting for watching all types of content. Interestingly, within the deeper settings of the television, we also find a dedicated 'Filmmaker' mode, which is now available from almost every manufacturer. However, in the case of the Dreame S100, we do not recommend using it – it blocks all settings related to the picture and does not allow for any configuration in the menu - even motion smoothing. So, how do the factory colours of the S100 fare? Not too badly, but also not the best. The television itself was quite heavily tinted with blue components, which is clearly visible in our white balance measurements. This made the picture artificially brightened, and the colours had a slight bluish tint. This is confirmed by the Color Checker palette, which showed how the colour samples tend to drift towards blue, generating quite significant colour errors. The picture in HDR mode looked definitely better, where the colour errors were not as prominent, although they still fell far from perfection. Here, the white balance was also shifted too much towards blue, causing considerable errors in colour reproduction.

In Filmmaker mode our R85H sample performed acceptably in SDR, but a few things stood out immediately in the measurements. In the white balance red increased with rising brightness while blue decreased. As a result the brighter parts of the picture had a slightly too warm tint. Gamma only held to the target value at the start. It then gradually fell, which left most of the image a little too bright. The combination of these two effects also showed up in the colours, with errors growing particularly in the brighter and more saturated samples. It did not look bad, but for Filmmaker mode we had expected a bit more accuracy. Interestingly, in HDR our sample fared noticeably better. The white balance was much more even and, over most of the range, the individual components stayed close together. Only in the very brightest areas did red start to be lacking, causing white to shift subtly. The EOTF curve, by contrast, looked very good and followed the target luminance levels across almost the entire range. That does not mean HDR colours were perfect. The Colour Checker still showed some deviations, especially for some of the more saturated hues, but the errors were clearly smaller than in SDR.

The Dreame S100 has proven to be a device very amenable to professional calibration. In its menu, we found a plethora of advanced options that allowed us to effectively correct factory errors. Thanks to advanced white balance correction – both in SDR and HDR content – we completely eliminated the excess blue components. The image immediately became much more natural and pleasant to the eye. After fully tuning the colours, the errors in the Color Checker test dropped below Delta E values of 2-3. For the average viewer, this means that any colour distortions have become nearly impossible to perceive with the naked eye. This is truly a pleasant surprise, as new brands in the television market often encounter significant issues with the correct implementation and functioning of sliders in the calibration menu. Here, everything worked as it should. The only thing we must criticise is the way the television manages its brightness in HDR. Due to the specific nature of the local dimming algorithms, anomalies can be seen on the EOTF curve we measured. Where high highlights should appear, there is a slight boosting of the curve above the reference line. In practice, this causes an unnatural brightening of some of the brightest elements in films or games. Despite this minor flaw, professional calibration brings about a huge, very positive change in this model.

After calibration the R85H shows it has genuinely substantial capabilities in colour reproduction. It performed best in SDR, where the white balance could be brought almost perfectly into line and the gamma stays very close to the target value across most of the range. This also translated into very low errors on the Colour Checker, so there is really little to pick fault with. HDR also shows a clear improvement, although some things can no longer be corrected as well. At the highest brightness levels the character of the backlight begins to make itself felt and the individual components of the white balance start to drift apart. The same applies to the most saturated colours, which do not always land exactly where they should. Despite these limitations the overall image is much more accurate than with the factory settings, and the biggest difference is definitely seen in SDR. There is, however, one thing calibration cannot fix. With heavily saturated elements the colour can slightly bleed onto adjacent bright areas of the image. It is easiest to notice this on white, which can pick up a faint tint of a neighbouring colour. It does not appear in every scene, but when the material is demanding the effect can be visible. This is more a characteristic of the way the RGB backlight works than a matter of picture settings.

For scaling and digital processing in the Dreame S100, the technology that the manufacturer proudly calls Dreamind™ Pro AI is responsible. However, it must be clearly stated that this is not a proprietary processor, but merely a marketing name for a chip made by MediaTek – more precisely, the MT5889 model. Unfortunately, we are talking about a rather outdated unit, from a time when MediaTek lagged significantly behind the competition in terms of performance. This unfortunately has a direct impact on the image as that is what this chip also deals with. The upscaling in the S100 is quite mediocre, and the processor has a tendency to heavily and artificially sharpen edges (fortunately, this can be somewhat tempered with the sharpness slider in the menu). In the options, we also find the "Contour Smoothing" feature, which theoretically is supposed to improve tonal transitions. In practice, it offers almost no improvement – setting it to any level rarely eliminates visible banding of colours, but it can unnecessarily blur elements of the image that should remain sharp.

A separate issue is the phenomenon of overscan (cropping the edges of the frame), which is noticeable here practically all the time. Although there is an option to set the screen to display the full frame in low-resolution materials, this requires activating the previously mentioned Filmmaker mode. And as you know, this mode completely blocks the adjustment of any other sliders in the picture menu. In our test tables, we marked that the television offers an option to disable overscan, as technically such a function exists in the system – however, in everyday use, when maintaining sensible image settings, you will not be able to take advantage of it.

In the R85H image processing is handled by the Micro RGB AI Processor. The name is new, but when it comes to upscaling lower-quality material the TV’s behaviour is very similar to what Samsung has shown before. And that’s not a bad thing, because the manufacturer has dealt with this aspect quite well for years. Lower-resolution content is displayed cleanly, edges do not become excessively soft, and the TV does not try to artificially extract details at any cost. The result is simply well balanced, and with regular broadcasts or older recordings the R85H performs as you would expect from a Samsung in this class. There remains only a minor drawback in the form of overscan, which cannot be completely turned off at the lowest resolutions. A separate matter is the additional processing of weaker sources. The R85H can smooth posterisation and somewhat tidy up heavily compressed footage, but it is worth emphasising that it does this relatively aggressively. At stronger noise reduction settings it begins to remove not only compression artefacts but also fine details and the natural film grain. It is therefore best used only when the material truly needs it.

When it comes to motion fluidity, the Dreame S100 performs really well. It features a panel with a refresh rate of 144 Hz, although in reality, the vast majority of people using it in a traditional way will still be limited to 120 Hz (as this is the maximum supported by today's equipment apart from PCs). Thanks to the high refresh rate, the image in motion remains really sharp, largely thanks to the well-functioning MEMC smoother. In the menu, we find predefined modes that allow the television to efficiently generate artificial frames, which effectively improves the fluidity of dynamic sports scenes, but also eliminates image stuttering in film and series materials recorded at 24 frames.

The R85H has a 144 Hz panel, so motion is really smooth here. Just bear in mind that the extra hertz above 120 are mainly used when connected to a PC, so in this category we don't treat them as any particular advantage. For films and TV, what's more important is that the panel is fast enough to provide access to an advanced motion-smoothing feature, "Motion Clarity", which lets you set motion anywhere from more film-like to very smooth. The panel's response time is a bit worse. Because it's a VA panel, fast-moving objects can leave a slight trail. On bright backgrounds this isn't very noticeable, but the darker the scene, the more pronounced the smearing becomes. This is a typical trait of this type of panel and it's easiest to notice in dark shots.

The Dreame S100 seems almost like the perfect screen for a gamer. The television is equipped with 2 HDMI 2.1 ports with a bandwidth of 40 Gbps, which means that functions like automatic game mode (ALLM) and variable refresh rate (VRR) ranging from 48 to 144 Hz work seamlessly on consoles. It also features a well-designed gaming panel (Game Bar), allowing users to continuously monitor important parameters—from active picture modes to the number of frames rendered by the console, and even the option to add a crosshair in the middle of the screen. All of this looks great on paper; however, the biggest problem with the S100 is the implementation of HDR mode in games. The television does not have the HGiG format implemented correctly. As a result, an unsuspecting user may incorrectly configure the console when first using the screen. Poor configuration can lead to an image that is too dark or too washed out (as a guideline: we suggest setting the brightness in the Xbox console menu to around 600 nits, and in the PS5 to 12 clicks). The Dolby Vision format also proves problematic, as it often would not activate at all between the XBOX and Dreame S100. And when it was finally managed to be activated, playing was not possible due to a very high input lag often exceeding 100ms. If the manufacturer can improve the HDR issue in games via an update, the S100 will be truly excellent hardware for gamers. For now, however, compromises need to be accepted.

Samsung has been steadily expanding its gaming features for several years, and the R85H really doesn't miss anything important here. The TV supports 4K at 120 Hz, VRR and ALLM, so we can easily take advantage of the PlayStation 5 or Xbox Series X. HGiG also works correctly, so with HDR properly set on the console, games retain the correct brightness and we don't unnecessarily lose detail in the brightest parts of the picture. The only major omission is Dolby Vision Gaming, which Samsung does not support on any of its TVs. The Game Bar is also very useful — a handy menu you can bring up without leaving the game. We can quickly check the frame rate, VRR status or change key picture settings. Samsung also adds a few less obvious features, such as the ability to change the aspect ratio and Game Motion Plus, which can further smooth games running at 30 or 60 frames per second. Of course this comes at the cost of higher input lag (around 55 ms), but for slower-paced titles it may be an interesting addition for some people.