tft display graph factory
Display screen is everywhere nowadays. Do you still remember the TVs or computer monitors 20 years ago? They were quadrate, huge and heavy. Now let’s look at the flat, thin and light screen in front of you, have you ever wondered why is there such a big difference?
Actually, the monitors 20 year ago were CRT (Cathode Ray Tube) displays, which requires a large space to run the inner component. And now the screen here in your presence is the LCD (Liquid Crystal Display) screen.
As mentioned above, LCD is the abbreviation of Liquid Crystal Display. It’s a new display technology making use of the optical-electrical characteristic of liquid crystal.
Liquid crystal is a state of substance that has both the characteristics of liquid and solid crystal. It don’t emit light itself, but it can let the light pass perfectly in specific direction. Meanwhile, liquid crystal molecule will rotate under the influence of a electric field, and then the light goes through it will rotate too. That said, liquid crystal can be a switch of light, which is the key in display technology.
STN LCD: STN is for Super-twisted Nematic. The liquid crystal in STN LCD rotate more angles than that in TN LCD, and have a different electrical feature, allowing STN LCD to display more information. There are many improved version of STN LCD like DSTN LCD (double layer) and CSTN LCD (color). This LCD is used in many early phones, computers and outdoor devices.
TFT LCD: TFT is for Thin Film Transistor. It’s the latest generation of LCD technology and has been applied in all the displaying scenario including electronic devices, motor cars, industrial machines, etc. When you see the word ‘transistor’, you may realize there’s integrated circuits in TFT LCD. That’s correct and the secret that TFT LCD has the advantage of high resolution and full color display.
In a simple way, we can divide TFT LCD into three parts, from bottom to top they are: light system, circuit system and light and color control system.In manufacturing process, we’ll start from inner light and color control system and then stretch out to whole module.
It’s accustomed to divide TFT LCD manufacturing process into three main part: array, cell and module. The former two steps are about the production of light and color control system, which contains TFT, CF (color filter) and LC (liquid crystal), named a cell. And the last step is the assembly of cell, circuit and light system.
Now let’s turn to the production of TFT and CF. Here is a common method called PR (photoresist) method. The whole process of PR method will be demonstrated in TFT production.
In Part 1, we were able to obtain and display values from KX022-1020 accelerometer on the TFT LCD panel. In Part 2, we’ll show how to control the TFT monitor while reading the contents of the program!
As before, we’ll be usingSainSmart ST7735RTFT monitor. It’s a compact LCD display that can be used both with Arduino and Raspberry Pi. The monitor has a built-in microSD card slot, so it’s possible to store and load images, in addition to reading and writing data. In this tutorial, we will only try to display values on the TFT monitor.
Once you made the change, compile the unzipped “TFT 18” directory with zip again, add it as a library in Arduino (or Arduino Create) Add Library, or place it under the “libraries” directory in Arduino’s installed directory and load the library.
for (uint16_t x=0; x < tft.width; x+=6) { tft.drawRect(tft.width/2 -x/2, tft.height/2 -x/2 , x, x, color); } } void testfillrects(uint16_t color1, uint16_t color2) { tft.fillScreen(BLACK); for (uint16_t x=tft.width-1; x > 6; x-=6) {
for(uint16_tx=0;x
Next, let’s display the accelerometer values on the TFT monitor! In the case of the Sensor Evaluation Kit, basically, it’s not necessary to change the wiring on the TFT monitor side. All that is needed is to insert KX022-1020 accelerometer to the Sensor Shield.
for (uint16_t x=0; x < tft.width; x+=6) { tft.drawRect(tft.width/2 -x/2, tft.height/2 -x/2 , x, x, color); } } void testfillrects(uint16_t color1, uint16_t color2) { tft.fillScreen(BLACK); for (uint16_t x=tft.width-1; x > 6; x-=6) {
for(uint16_tx=0;x
When it goes to the edge of the 120px display, the graph is cleared with drawrect. The numbers on the upper part are updated with drawrect for each frame in the same way.
This concludes the tutorial on how to display and graph accelerometer values using TFT LCD monitor! There are quite a few side projects we can consider developing. For example, we can combine this TFT monitor and Arduino Pro Mini to make a wristwatch featuring small games, etc. It is also possible to make a data logger using, for instance, different sensors from the Sensor Evaluation Kit.
We covered the basics of accelerometer previously inUsing Arduino with Parts and Sensors – Accelerometer Part 1andUsing Arduino with Parts and Sensors – Accelerometer Part 2. Today we’ll be testing KX022-1020 accelerometer using TFT liquid crystal panel. We’ll discuss how to control the TFT LCD in more detail in the next article. In addition, we’ll further exploreArduino Create. For more information about Arduino Create, please refer back tothisarticle.
Now the sample program is working fine, let’s try to display the values on a 1.8 inch TFT LCD monitor. Although this TFT liquid crystal monitor has a resolution slightly smaller than 126 x 160 px, it’ll be quite useful when displaying numbers or letters with Arduino etc.
In the past, we used 7-segment LED to display numerical values only. But this time, I tried to display the graph along with the values obtained from the accelerometer.
When using the TFT monitor, the connection method and the library used in the program may be different depending on the specification of each TFT monitor. The TFT monitor used in this tutorial is a monitorSainSmart ST7735R. In addition to Arduino, the monitor is also compatible with Raspberry.
In order to use the monitor to run the program in Arduino, we’ll have to modify the downloaded library a little bit.We’ll go over how to control the TFT LCD in more detail in the next article. Once everything is set, you will be able to output numerical values in the monitor as shown in the video below:
In the next part, we’ll create a simple device using the same accelerometer and TFT monitor. We’ll show how to create graphs and display the values obtained from the accelerometer on the TFT monitor.
Develop and maintain gorgeous user interfaces and menu structures for your GTT using this immersive Windows environment tool. Design Software for the GTT Rev2.0 line, supports the GTT29A, GTT35A, GTT38A, GTT43A, GTT50A, GTT52A, and GTT70A. Version 1.10.0.9818 brings new tools, such as Axis Ticks and Two Colour Bargraphs, updates the Actions tab with options such as Set Bargraphs Colour and Set Gauge Test Colour, optimizations found in the updated Settings windows that allow even more loading options, and a number of bug fixes.
Luminance is an important quality characteristic of a TFT-LCD display. Consumers prefer to have a TFT-LCD display with enough luminance. Several control factors in manufacturing a TFT-LCD display may affect the luminance of a TFT-LCD display. Tuning the control factors using a trial and error method incurs significant overhead in time and resources such as the material and the workforce. Therefore, a more efficient method is required for tuning the control factors in manufacturing TFT-LCD displays. To fine-tune the control factors, we need to perform experiments to find the luminance of TFT-LCD displays for different settings of the control factors. An experiment to find the luminance for a specific parameter setting of the control factors requires about 1.5 months for the case company. To reduce the time in fine-tuning the control factors, we should not perform too many experiments in a new method. In this paper, we first used the traditional Taguchi method to find a control factors’ setting to enhance the luminance of TFT-LCD displays. However, because the Taguchi method allows only a few fixed levels for each control factor, it usually cannot find the global optimal setting for the control factors. Therefore, this paper proposes to improve the Taguchi method by incorporating the Taguchi method with the neural network and the genetic algorithm. In the proposed method, we used the data collected from the Taguchi method to train a neural network to predict the luminance of a TFT-LCD display for a given control factors’ setting. Then, we used a genetic algorithm to search for the global optimal control factors’ setting using the predicted luminance of a control factors’ setting as the fitness value of the setting.
The case company is a TFT-LCD display manufacturer in Taiwan. In performing the optical film lamination step in the liquid crystal module assembly process (LCM), the case company observed color streaks, as shown in Figure 1, on some of their TFT-LCD displays. The color streaks reduced the luminance of the TFT-LCD displays. Therefore, to enhance the luminance of TFT-LCD displays, the optical film lamination step needs to be optimized. The optical film lamination step consists of five stages: the raw material inspection, the soft-to-hard lamination, the circuit test, the protection film lamination, and the clean process, as shown in Figure 2. According to a failure analysis performed by the case company, the color streaks on a TFT-LCD display were mainly due to improper setting of control factors in the soft-to-hard lamination stage. Therefore, this paper focuses on optimizing the soft-to-hard lamination stage to improve the luminance of TFT-LCD displays.
For the quality improvement purpose, engineers of the case company selected five positions, as shown in Figure 3, on a TFT-LCD substrate to measure their corresponding values of luminance. The luminance of a TFT-LCD display is the average of the five values of luminance of the selected positions. The case company manufactured about 5000 TFT-LCD displays per day. Therefore, the TFT-LCD displays’ average luminance at a specific date was calculated by averaging the luminance of all the TFT-LCD displays made at that particular date. Figure 4 shows the trend chart for the TFT-LCD displays’ average luminance. The average of the average luminance at all dates in a selected time interval is defined as the baseline, the red line shown in Figure 4. Before fine-tuning the control factors in the sot-to-hard lamination stage, the baseline was 17.03, which was less than the required minimum luminance of 25. Thus, to enhance the TFT-LCD display’s luminance, one needs to find a setting of the control factors such that the baseline is larger than the required minimum luminance of 25. Therefore, we propose in this paper a new method that combines the Taguchi method, a neural network, and a genetic algorithm to fine-tune the setting of control factors in the soft-to-hard lamination stage to enhance the TFT-LCD displays’ luminance.
This guide is about DWIN HMI Touch Screen TFT LCD Display. HMI Means Human-Machine Interface. DWIN is specialized in making HMI Touch screen displays that are compatible with all microcontrollers like Arduino, STM32, PIC, and 8051 families of Microcontrollers.
This is a Getting Started tutorial with 7-inch DWIN HMI TFT LCD Display. We will see the architecture, features, board design, components, and specifications. We will also learn about the TTL & RS232 interfaces. Using the DGUS software you can create UI and with SD Card you can load the firmware on display memory.
One of the method to load the firmware to the T5L DWIN LCD Display is by using the SD Card. An SD Card of up to 16GB can be used to download the firmware files. We can easily insert the Micro SD card into the SD Card slot on the backside.
Under the Download, section Go to tools and Download the DGUS-DWIN Graphic utilized software i.e DGUS software and XR21X Driver. This driver is for the HDL662B board.
After copying the file, remove the SD Card from your computer and insert it into the SD Card slot of DWIN LCD Display. Then power the display using the USB Cable. The firmware downloading process will start automatically.
The next part of this tutorial includes creating UI and interfacing DWIN LCD Display with Arduino. For that you can follow the DWIN LCD Arduino Interfacing Guide.
What makes the PG32UQX particularly impressive is that it also delivers a range of top-end, cutting edge monitor technologies all from one screen. Perhaps most interesting is the use of a Mini LED backlight, offering 1152 direct dimming zones behind the panel and designed to offer the best HDR performance you can get from an LCD desktop gaming monitor. This offers far better local dimming capabilities than the wide range of edge lit displays that flood the market, and is also a step beyond the older FALD (Full Array Local Dimming) methods seen in previous generations like the 27″ 4K 144Hz models we mentioned earlier. For instance the Asus ROG Swift PG27UQ – their flagship model at the time in August 2018 had “only” 384 zones. So it’s a big step up here. With the usage of a Mini LED backlight comes support for the top tier of VESA’s DisplayHDR certification, with this screen offering HDR 1400 support. This includes, as the name suggests, a 1400+ nit peak brightness! A Quantum Dot panel coating provides a wide colour gamut with Asus quoting 98% DCI-P3 and a very wide 160% sRGB coverage, and the panel also supports the necessary 10-bit colour depth for HDR content.
If you enjoy our work and want to say thanks, donations to the site are very welcome. If you would like to get early access to future reviews please consider becoming a TFT Central supporter.
Check Availability and Pricing – Affiliate LinksTFTCentral is a participant in the Amazon Services LLC Associates Programme, an affiliate advertising programme designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.co.uk, Amazon.de, Amazon.ca and other Amazon stores worldwide. We also participate in a similar scheme for Overclockers.co.uk.
The PG32UQX comes in familiar ROG Swift display style. It has a 3-side “borderless” panel along the top and sides, with a thin 1.5mm plastic edge, and then a 6.5mm black panel border before the image starts. The bottom has a thicker black plastic bezel which is a bit thicker towards the middle than the edges, up to around 26mm maximum.
The stand attaches in to the middle of the screen but can be removed to reveal a VESA 100mm mounting option if you wish. Some of the lighting features are built in to the stand so you would lose those of course if you removed the provided stand. In the top of the stand is another smaller ROG logo which glows red if you have turned on the ‘Light in Motion’ ROG Light Signature in the OSD menu. The stand has a dark grey plastic finish and comes already attached to the screen. The foot is a dark grey aluminium base and needs to be screwed on to the bottom of the stand which is quick and easy. This foot provides a strong and sturdy base for the display although it has a bit of an annoying section at the back which sticks out quite far. This means you need quite a deep desk as the screen has to sit further forward than other displays to accommodate that extra foot. This is the same on other ROG screens we’ve tested and a bit of an ongoing annoyance in the stand design.
The materials were of a very good standard and the build quality felt solid. There is a good range of ergonomic adjustments offered by the stand. Tilt offers a wide range and is smooth to move, if a little stiff to operate. There is a somewhat limited 70mm height adjustment which smooth and reasonably easy to move. At the lowest setting the bottom edge of the screen (in the middle section of the display) is 70mm from the top of the desk, and 140mm when at maximum height extension. Side to side swivel is provided and is smooth and easy to use. Rotation to portrait mode is not included here though, not that we really miss that. There is pretty much no wobble at all from the screen when you re-position it, and the stand is providing a very sturdy and stable base.
The PG32UQX offers a good range of connectivity with 1x DisplayPort 1.4 (with DSC support) and 3 x HDMI 2.0 offered for video connections. There is no HDMI 2.1 offered, and that has yet to be used on any available desktop monitor. Although it’s interesting to note that this screen uses a Native G-sync hardware module and until recently the “v2” module has been limited to a single DP 1.4 and single HDMI 2.0. NVIDIA haven’t yet added HDMI 2.1 support, but it’s interesting to see a couple of extra HDMI 2.0 ports are now available. We refer to this as the v2.3 module (v2.2 having DisplayPort 1.4 and 2x HDMI 2.0) to make that distinction which is talked about more later in the review, along with further discussion about the lack of HDMI 2.1 and whether that’s a big problem or not. There are also 2x USB 3.1 downstream ports and a USB upstream port on the back here along side the video connections. In the photo above you’ve also got a good view of the OSD controller dial and buttons which we will talk about in a moment.
On the front bottom bezel Asus have added a 2″ size LiveDash OLED display so that users can conveniently stream system stats, custom GIFs, or game clan logos to highlight allegiance. This can be controlled via the OSD menu to show an FPS counter, your GameVisual settings, Input Source, Timer or stopwatch. You can also control the brightness of the graphic or turn it off altogether if you’d rather. It’s quite a nice idea but for some reason Asus have insisted on a very annoying and distracting ROG logo that flashes across the screen fairly often! This is really distracting when you’re trying to use the screen for anything, and just made us want to turn the LiveDash off completely. Asus need to sort that out or at least add a user control to disable that logo flash! You don’t need a constant reminder that you’re using a top-end and expensive ROG gaming screen – your bank statements or your better half will constantly remind you we’re sure!
The OSD menu is controlled primarily through a dial controller on the bottom edge of the screen. This has a clickable button in the middle to select settings, and then there is one other button on either side of the dial for other selections and options. This is an alternative to the small joysticks that have become common on displays. It works fairly well, it’s quick and easy to use and has a smooth feel to the movement. It takes a day or two to get used to it when you’re used to other OSD controls but it’s decent enough.
One thing we should mention is that there is a small internal fan in this screen which is needed to ensure it stays cool enough, especially during high refresh HDR gaming. It helps keep the G-sync module and FALD backlight cooler. We have seen small fans used in screens before, including on the first generation 4K 144Hz displays like the Asus ROG Swift PG27UQ. You can hear the fan on the PG32UQX if you listen closely for it, and if you have an otherwise silent PC setup you will notice it. It is fairly quiet and certainly we found that it was not audible beyond our normal couple of PC case and graphics card fans which are fairly quiet. We would have course preferred it without the fan, and it’s likely to bother silent PC enthusiasts but it wasn’t a major issue we felt. Perhaps a little noisier than other G-sync module screens like the LG 38GL950G for instance, but nothing major.
It should be noted that unlike some other Native G-sync screens the 10-bit colour depth is available even at 4K 144Hz thanks to the inclusion of Display Stream Compression (DSC) on the DP 1.4 connection. DSC provides a visually lossless compression which means that unlike older 4K 144Hz models you don’t need to sacrifice colour depth or chroma levels to reach the maximum refresh rate. You need a compatible DSC graphics card though of course. The Asus reviewers guide confirms that the screen can do true 10-bit at 4K 120Hz thanks to DSC, but for the very top 4K 144Hz it has to switch to 8-bit+FRC due to DSC limitations apparently. Visually that should be pretty impossible to see but we mention it for completeness and for those particularly technical readers. The main take-home here is that you don’t need to drop to 8-bit colour depth which is great if your content or game is HDR and supports 10-bit. More importantly you don’t need to drop from 4:4:4 full RGB chroma here, which does cause some visual loss in clarity but was a necessary step in earlier generation of 4K 144Hz displays before DSC was adopted.
This section tests the full range of luminance (the brightness of the screen) possible from the backlight, while changing the monitors brightness setting in the OSD menu. This allows us to measure the maximum and minimum adjustment ranges, as well as identify the recommended setting to reach a target of 120 cd/m2 for comfortable day to day use in normal lighting conditions. Some users have specific requirements for a very bright display, while others like a much darker display for night time viewing or in low ambient light conditions. At each brightness level we also measure the contrast ratio produced by the screen when comparing a small white sample vs. a black sample (not unrealistic full screen white vs. full screen black tests). The contrast ratio should remain stable across the adjustment range so we also check that.
Graphics card settings were left at default with no ICC profile or calibration active. Tests were made using an X-rite i1 Display Pro Plus colorimeter paired with the Calman Ultimate software for very high levels of accuracy.
We have plotted the luminance trend on the graph above. The screen behaves as it should in this regard, with a reduction in the luminance output of the screen controlled by the reduction in the OSD brightness setting in a linear relationship. The average contrast ratio of the screen was measured at 1283:1 out of the box which was excellent for an IPS-type panel and a fair bit beyond the manufacturer 1000:1 specification even.
Performance is measured and evaluated with a high degree of accuracy using a range of testing devices and software. The results are carefully selected to provide the most useful and relevant information that can help evaluate the display while filtering out the wide range of information and figures that will be unnecessary. For measurement we use a UPRtek MK550T spectroradiometer which is particularly good for colour gamut and spectrum measurements. We also use an X-rite i1 Pro 2 Spectrophotometer and a X-rite i1 Display Pro Plus for measurements in various ways. Various software packages are incorporated including Portrait Displays Calman Ultimate package. We measure the screen at default settings (with all ICC profiles deactivated and factory settings used), and any other modes that are of interest such as sRGB emulation presets. We then calibrate and profile the screen before re-measuring the calibrated state.
Greyscale dE– this graph tracks the accuracy of each greyscale shade measured from 0 (black) to 100 (white). The accuracy of each grey shade will be impacted by the colour temperature and gamma of the display. The lower the dE the better, with differences of <1 being imperceptible (marked by the green line on the graph), and differences between 1 and 3 being small (below the yellow line). Anything over dE 3 needs correcting and causes more obvious differences in appearance relative to what should be shown. In the table beneath the graph we provide the average dE across all grey shades, as well as the white point dE (important when considering using the screen for lots of white background and office content), and the max greyscale dE as well.
RGB Balance and colour temperature – the RGB balance graph shows the relative balance between red, green and blue primaries at each grey shade, from 0 (black) to 100 (white). Ideally all 3 lines should be flat at the 100% level which would represent a balanced 6500k average colour temperature for all grey shades. This is the target colour temperature for desktop monitors, popular colour spaces like sRGB and ‘Display DCI-P3’ and is also the temperature of daylight. It is the most common colour temperature for displays, also sometimes referred to as D65. Where the RGB lines deviate from this 100% flat level the image may become too warm or cool. Beneath this RGB balance graph we provide the average correlated colour temperature for all grey shades measured, along with its percentage deviance from the 6500k target. We also provide the white point colour temperature and its deviance from 6500k, as this is particularly important when viewing lots of white background and office content.
Gamma– we aim for 2.2 gamma which is the default for computer monitors in SDR mode. Testing of some modes might be based on a different gamma but we will state that in the commentary if applicable. A graph is provided tracking the 2.2 gamma across different grey shades and ideally the grey line representing the monitor measurements should be horizontal and flat at the 2.2 level. Depending on where the gamma is too low or too high, it can have an impact on the image in certain ways. You can see our gamma explanation graph to help understand that more. Beneath the gamma graph we include the average overall gamma achieved along with the average for dark shades (0 black to 50 grey) and for lighter shades (50 grey to 100 white).
dE colour accuracy– a wide range of colours are tested and the colour accuracy dE measured. We compare these produced colours to the sRGB reference space, and if applicable when measuring a wide gamut screen we also provide the accuracy relative to a specific wide gamut reference such as DCI-P3. An average dE and maximum dE is provided along with an overall screen rating. The lower the dE the better, with differences of <1 being imperceptible (marked by the green area on the graph), and differences between 1 and 3 being small (yellow areas). Anything over dE 3 needs correcting and causes more obvious differences in appearance relative to what should be shown. dE 2000 is used for improved accuracy and providing a better representation of what you would see as a user, compared with older dE methods like dE 1994, as it takes into account the human eye’s perceptual sensitivity to different colours.
Check Availability and Pricing – Affiliate LinksTFTCentral is a participant in the Amazon Services LLC Associates Programme, an affiliate advertising programme designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.co.uk, Amazon.de, Amazon.ca and other Amazon stores worldwide. We also participate in a similar scheme for Overclockers.co.uk.
So given this is specifically configured on the screen we should take that in to consideration when measuring the default display accuracy as well. By the way, if you disable this sRGB gamma option in the OSD then the alternative is a BT.1886 gamma which as explained by Portrait Displays is“the recommended gamma function for flat panel displays used in HDTV studio production”.Note that this same ‘SDR YCbCr sRGB gamma’ option is also available for HDMI inputs in the OSD menu.
Initially out of the box the screen was set in the ‘Racing mode’ preset with the ‘Display SDR input’ gamut mode set to sRGB as opposed to the full native wide gamut of the backlight. This is a working sRGB emulation mode which restricts the gamut to that smaller colour space nicely. This is likely to be more appropriate for general uses and SDR content. The display was set with a moderate 55% brightness which was still too bright for general use, and as with most displays will need to turn it down. Colours looked like a standard gamut screen as expected and colour temperature felt right. We went ahead and measured the default state with our testing equipment.
We also wanted to test the performance in the native wide gamut mode. Without changing anything else in the OSD menu we switched the ‘Display SDR input’ to wide gamut . Immediately you could tell that the colours looked different with far more vivid reds and greens in particular. For SDR content this leads to oversaturated and inaccurate colours so unless you specifically like that look, we would probably only recommend switching to this mode for multimedia and games where HDR is supported along with the wide colour gamut, or perhaps also if you’re working with wide gamut content. For normal SDR content the wide gamut colours looked very garish here.
If you have access to a calibration tool you can of course calibrate this screen for colour-aware applications to other reference spaces, but we thought we would also compare the default performance of the screen relative to Adobe RGB here. Adobe RGB is the common colour space for photo editing and so might be something some people want to work with away from their gaming and normal usage. There is a very good 99.9% coverage of Adobe RGB here, but it does extend quite a long way beyond that reference space in red shades, giving a 121.2% relative coverage. Overall colour accuracy relative to the Adobe RGB colour gamut is good with a 2.1 average, but some higher errors particularly in red (5.0 max) where the produced shade is quite a lot deeper than the intended Adobe RGB red primary. Note that we have compared the wide gamut mode here against Adobe RGB as it’s probably the most common defined colour space usage beyond normal SDR/sRGB content. DCI-P3 is used for video/cinema content creation but that is fairly niche. HDR content requires Rec.2020, or as much of that space as the screen can manage and no display can cover 100% Rec.2020 fully. The PG32UQX does well with 82% coverage though here. We will look at HDR mode in more detail later and consider the colours relative to the Rec.2020 colour space then.
The OSD settings were adjusted as shown in the table above, as guided during the calibration process and measurements. These OSD changes allowed us to obtain an optimal hardware starting point and setup before software level changes would be made at the graphics card level. These adjustments before profiling the screen would help preserve tonal values and limit banding issues. After this we let the software carry out the further corrections and create an ICC profile.
You can use our settings and try our calibrated ICC profile if you wish, which are available in our ICC profile database. Keep in mind that results will vary from one screen to another and from one computer / graphics card to another.
Check Availability and Pricing – Affiliate LinksTFTCentral is a participant in the Amazon Services LLC Associates Programme, an affiliate advertising programme designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.co.uk, Amazon.de, Amazon.ca and other Amazon stores worldwide. We also participate in a similar scheme for Overclockers.co.uk.
We wanted to test here how uniform the brightness was across the screen, as well as identify any leakage from the backlight in dark lighting conditions. Measurements of the luminance were taken at 35 points across the panel on a pure white background. The measurements for luminance were taken using BasICColor’s calibration software package, combined with an X-rite i1 Display Pro colorimeter with a central point on the screen calibrated to 120 cd/m2. The below uniformity diagram shows the difference, as a percentage, between the measurement recorded at each point on the screen, as compared with the central reference point.
Spectral distribution graph showing calibrated mode at 6500k (left) and with low blue light filter at preferred level 3 setting (right). Blue peak is at 452 nW/nm
The blue light peak is at 452 nW/nm and there are settings for a Low Blue Light filter available in the OSD menu on this display with levels 1 to 4 available. These make the image progressively warmer. Level 3 measured at 5359k and looks reasonable to use for viewing later at night or when working with a lot of white background content. The level 4 mode looks very yellow in appearance and has a white point of 4460k which is a bit too aggressive we think.
Check Availability and Pricing – Affiliate LinksTFTCentral is a participant in the Amazon Services LLC Associates Programme, an affiliate advertising programme designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.co.uk, Amazon.de, Amazon.ca and other Amazon stores worldwide. We also participate in a similar scheme for Overclockers.co.uk.
The screen can accept an HDR10 input signal and supports the DisplayHDR 1400 standard which is currently the top tier of VESA’s certification scheme. This requires backlight local dimming, a peak brightness of at least 1400 nits, wide colour gamut >90% DCI-P3 and 10-bit colour depth support. The reason this screen can reach this HDR1400 tier, particularly related to the high peak luminance requirement, is because of its impressive Full Array Local Dimming (FALD) Mini LED backlight. This is a direct backlight unit that sits behind the LCD panel and is split in to 1152 independent zones. This allows the screen to be brightened or dimmed for HDR content in the most effective way currently available from any desktop LCD monitor.
The 1152 local dimming zones of this backlight easily surpass even previous generations of flagship LCD monitor backlights including the 384 zone FALD featured on the older 27″ PG27UQ display. With more zones the screen is capable of brightening and dimming content in smaller zone areas for more finite control over the content. This also really helps reduce blooming and improve the overall appearance of the screen in HDR content, particularly in darker scenes. It is also of course miles beyond what you can achieve with edge-lit local dimming backlights of lower cost screens (typically HDR 600 certified displays) where there are very few dimming zones (often 8, 16 or 32 if you’re lucky) and those zones are much larger. The result is a dramatically enhanced HDR experience for gaming and video. We will talk about real-life experience in a moment, let’s look first at the setup of the HDR mode:
The Quantum Dot coating helps the backlight offer a wide colour gamut that can cover 82% of the Rec.2020 space that is defined for HDR content. Basically the more of this space that a screen can cover, the better. This corresponds by the way to 112.7% relative DCI-P3 coverage that we measured earlier in the review. The colour accuracy is excellent as well in HDR mode with an average dE across all samples of only 0.7. Because the screen cannot fully cover the Rec.2020 colour space (no display can) there are some higher dE errors up to 2.6 maximum in 100% saturated colours like red, blue and green for instance. If we exclude the 100% saturated shades there then the average dE is 0.5. Excellent colour accuracy in HDR mode here.
Having both this PG32UQX top end Mini LED HDR display with us at the same time as the LG 32EP950 OLED display we just finished reviewing gave us the chance to compare the two screens side by side for HDR content as well. The PG32UQX has a more obvious pop to the bright areas because of the increased peak brightness, although sometimes we felt this was perhaps a little over the top and created a slightly unrealistic appearance sometimes. This could be down to the way the content is mastered, but also because there is typically a desire from display manufacturers to make local dimming backlights pop a bit more. It looks very impressive and you can clearly see and demonstrate the value of HDR, but 1600 nits was maybe a bit unnecessary for close up desktop use.
This night time video around Yokohama Chukagai and Minatomirai (make sure to play it in a compatible HDR browser) was a good example to compare on both screens side by side. Both looked excellent but the PG32UQX had brighter, but perhaps less realistic, looking highlights in some places. From up close even the more modest ~580 nits peak brightness of the LG 32EP950 OLED looked very good still. We didn’t see any noticeable issues with blooming or halos on the PG32UQX which was impressive though in this video. Obviously you don’t get any risk of blooming or halos on the OLED display thanks to the per-pixel dimming control.
When it comes to HDR gaming there is no contest really here between these two displays. The LG 32EP950 lacks a high refresh rate or VRR support and isn’t even aimed at gamers. The PG32UQX is custom built for HDR gaming with a G-sync module, excellent VRR performance and a high refresh rate for significantly improved motion clarity and frame rate support. The additional peak brightness also really looks great in gaming, making bright areas pop very nicely and leading to a very enjoyable experience. For gaming, the actual accuracy of the image is less important than the subjective appearance and the PG32UQX and it’s Mini LED backlight does really well here.
One thing we should note though is that while the PG32UQX is undoubtedly a better option for HDR gaming, it is not suitable for any kind of professional HDR content creation or anything that relies on high levels of accuracy. We explained this in detail in our LG 32EP950 review here, but the accuracy of content is always a problem with any kind of dynamic local dimming backlight like this. Sure, the PG32UQX can produce a very pleasing HDR experience for its intended usage and content consumption, but what it is not, is a reference HDR display or suitable for content creation. Obviously that is not the target market for this screen but it’s important to understand the limitations with true image accuracy when using any local dimming backlight.
The screen uses overdrive technology to boost pixel transitions across grey to grey changes as with nearly all modern displays. The part being used is an AU Optronics M320QAN02.6 AHVA (IPS-type) technology panel. Have a read about response time in our specs section if you need additional information about this measurement. Our thanks to NVIDIA for hooking us up with an RTX 3090 for all our testing.
TFTCentral is a participant in the Amazon Services LLC Associates Programme, an affiliate advertising programme designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.co.uk, Amazon.de, Amazon.ca and other Amazon stores worldwide. We also participate in a similar scheme for Overclockers.co.uk.
To provide the fairest view of the measured response times, and one that can be properly compared with data from the last 25 years of monitor measurements, we have first of all provided them based on the traditional response time measurement process. If you want to read more about this, and why we switched to a different method for 2021, you can see our detailed article here. While our new ‘gamma corrected’ method provides better accuracy relative to real-life experience, and overcomes some limitations with the traditional method, we felt it was useful to also provide the legacy measurements in full here. After all, with all previous displays measured in that way it’s the only sensible way to have a direct reference point for this new screen. We would normally only provide the overall average G2G figure for “legacy measurement”, but we thought the full view would be useful here given the screens gaming focus.
This 9.3ms G2G figure alone is a bit unfair relative to other screens we have tested in the past based on the traditional, although less representative of real-life, response time measurement method. We just don’t have enough data from older tested screens based on this method to provide a fair comparison. If you want to consider how this screen performs relative to other IPS gaming displays you would be better off comparing the figures from the traditional method. You will see in the first table above that if we had stuck with the old method then the average G2G is a more respectable 7.3ms G2G. Still a bit slow relative to a lot of other modern IPS panels but only be a couple of ms. It makes a bit of a difference to the “heat-map” table as well, but those slower changes that we’ve described above are still identifiable in that table as well. The good thing about the new improved method is that it really highlights any problem areas more clearly and tells a better story of the panels performance. We will talk about refresh rate compliance in the following section. For now we need to look at the actual real-life motion clarity to complete the picture here:
Above we have provided a view of the real life motion clarity at the maximum 144Hz captured with a pursuit camera in the optimal Normal OD mode on the PG32UQX. This is then compared side by side with some other recent and popular 144Hz IPS displays we have tested (well, 160Hz in the case of the LG 27GN950 but close enough). The result of those slower transitions from white/light shades to darker shades is evident here in motion tests by the subtle pale blur and trails behind the moving object. This is not overshoot, and is nowhere near as obvious and distracting in usage as major overshoot can be. This is just a bit of a pale blur to the image because some of the response times are slower. In comparison you can see the other screens have a better motion clarity overall. The response time average figures that are provided are using the traditional G2G measurement method for all four screens to keep it fair.
What makes up for this nicely is the screens other features. The 32″ screen size is a really nice jump up from common 27″ 4K displays and really allows you to make better use of that resolution and screen area. You get a super sharp and clear image, but with the larger screen size that might actually help you see more of that detail in your games. That’s a definite improvement over older 27″ 4K screens. The increased 5″ size also adds to the feeling of immersion and it seems like quite a sweet spot in size/resolution. Then there is of course the Mini LED backlight which offers HDR capability far beyond any of the gaming screens compared above. You just can’t compare something with a small number of edge lit zones to this screen when it comes to HDR gaming and content playback. Dark areas are dimmed to improve black depth dramatically, while bright areas really pop with a peak brightness of up to 1600 nits. You get a really impressive HDR experience with vivid colours, the capability to produce very bright images and very good local dimming control. You can read more about our analysis of HDR content, brightness and local dimming in our HDR section earlier in this review
The PQ32UQX can support variable refresh rates from both AMD FreeSync and NVIDIA G-sync systems. Despite being a Native G-sync screen with their added hardware module we confirmed that it can be used for FreeSync from an AMD graphics card over DisplayPort, as it includes the latest firmware that allows this. G-sync also works over HDMI if for some reason you want to connect a PC via that connection, but importantly it does not offer FreeSync over HDMI though unfortunately which could have been useful for console VRR support. The VRR range supported is between 1 and 144Hzthanks to the Native G-sync module.
We should also also make a point here that this screen features an updated Native G-sync hardware module (discussed more later on) that includes Display Stream Compression (DSC) support over DisplayPort 1.4. That means that even at the max 4K 144Hz you can run at 10-bit colour depth and full RGB chroma. You don’t need to sacrifice chroma and image quality at all like you did on earlier generation 4K 144Hz screens, like their 27″ PG27UQ for instance which was one of the first. That’s great news. If you’re running a graphics card without DSC then the max refresh rate seems to be limited to 96Hz from what we could tell although we didn’t test this extensively. If you’re going to spend $3000 on a monitor we’re pretty sure you are going to have a top end graphics card with DSC anyway.
There’s various things you need to consider when it comes to response times and gaming, particularly on a display with high refresh rate support. Gaming screens invariably give you a control for the overdrive impulse in the OSD menu which can help you tweak things, but response time performance and overshoot levels can vary depending on the active refresh rate. This behaviour is often different depending on whether the screen is a traditional Native G-sync screens (with hardware module) or whether it’s an adaptive-sync screen, and not all screens behave in the same way. We always try to test each variable in our reviews but the key considerations you need to make are:
Performance during VRR (Variable Refresh Rate) – bearing in mind that the refresh rate will fluctuate anywhere from 1Hz up to the maximum supported by the screen (e.g. 1 – 144Hz on a 144Hz display). It’s important to understand if the response times and overshoot will vary as the refresh rate changes. There may be a need to switch between different overdrive settings in some cases, depending on your usually attained frame rate/refresh rate output and graphics card capability. This can sometimes become fiddly if your refresh rates fluctuate a lot, especially between different games, so it’s always easier if you can leave a display on a single overdrive setting which is suited to the whole range. Some screens also feature “variable overdrive” which helps control the response times and overshoot depending on the active refresh rate. This is particularly apparent with traditional Native G-sync module screens.
Whether the response times can keep up with the frame rate– you will also want to consider whether the response times of the panel can consistently keep up with the frame rate. For instance a 144Hz screen sends a new frame to the display every 6.94ms, so the pixel response times need to ideally be consistently and reliably under this threshold. If they are too slow, it can lead to added smearing and blurring of the image in practice and sometimes make the higher refresh rates unusable in real life. We consider this in our analysis.
In this section for our reviews we look at the response time behaviour across the range of supported refresh rates and consider whether they are sufficient to keep up with the frame rate demands of the screen. The grey line on the graph shows the refresh rate threshold, that being the average G2G response time that the panel needs to be able to achieve to keep up properly with the refresh rate and frame rate. For instance at a 60Hz refresh rate the response times need to be consistently and reliably under 16.67ms, while at 144Hz refresh rate the response times ideally need to be under 6.94ms to keep up with the frame rate demands. If they are not then this can lead to some additional smearing and blurring on moving content as the pixels can’t keep up. For these tests we will plot the average G2G figure at a range of measured refresh rates, while operating at the optimal overdrive control.
The screen includes a Native G-sync hardware modulewhich is in fact a slightly updated revision of what we refer to as the “v2 module”. The v2 module includes DisplayPort 1.4 and HDMI 2.0 connections, an update from the v1 module which had DisplayPort 1.2 and HDMI 1.4. As a result allows support for high bandwidths of modern displays at high resolutions and refresh rates. We’ve seen the v2 module used in a fair few screens to date, including recently in our review of the 24.5″ sized Asus ROG Swift 360Hz PG259QN.
We saw the v2 module updated slightly when we tested the Dell Alienware AW3821DW and AW2721D in that the module now supports a second HDMI 2.0 port as well, with previous generations being limited to just a single HDMI 2.0. It’s not a major step change in the module as it doesn’t yet support anything like HDMI 2.1 or DisplayPort 2.0 for instance, so we referred to that as v2.2. Here on the PQ32UQX there is a third HDMI 2.0 port added (we are being spoiled now!) so let’s call this v2.3.
A lot of people will be asking though where HDMI 2.1 is? This is of course the connection of choice for modern games consoles like the PS5 and Xbox One X and would offer support for 4K 120Hz, variable refresh rates (HDMI-org VRR) and ALLM (Auto Low Latency Mode). It is also available on the top end modern graphics cards, if you can get hold of one. First let’s explain why it is not featured here…
At the moment the Native G-sync module does not support this connection type, and if/when that appears that would be (in our classification at least) the v3 module. So right now the limitation is because of the use of the G-sync v2 module. So why not drop the G-sync module and then you can add HDMI 2.1 you might ask? Well you could do that, but then that leads to a lot of other, arguably bigger problems than a missing connection port. This display has been co-created with NVIDIA as it is incredibly complex to get a FALD/Mini LED backlight working properly with high refresh rates, and particularly with variable refresh rates. Everything needs very carefully tuning, including the response times, backlight local dimming and refresh rates. Response times need variable overdrive which is a feature of the module, but the main challenge is getting the FALD to respond quickly enough, and in the correct way when VRR and high refresh rates are used. So far you won’t see any FALD gaming displays that don’t feature the Native G-sync module, it’s just not possible (or at least very easy) to get it working without the module being used to make it happen. So if you ditched the module in order to get HDMI 2.1, you’d be causing more problems.
Let’s think then about whether you even need HDMI 2.1 here? For PC gaming, which is absolutely the target audience for this screen, no you don’t need it at all. DisplayPort is the primary connection of choice for graphics cards and thanks to DSC there is no sacrifices needed here anyway to reach the full resolution, refresh rate and colour levels. So for PC gaming, it isn’t needed anyway.
We have talked in the previous section about the lack of HDMI 2.1 connection on this screen which on the face of it might be off-putting to some console users and is certainly a bit of a shame. However, Asus and NVIDIA have made some adjustments to maximise the capability of the provided HDMI 2.0 connection which covers some angles of usage. The main point is that they have added support for 4K at 120Hz even over the HDMI 2.0! To get around the bandwidth limit of this connection they have made use of 4:2:0 chroma sub-sampling to compress the colour data. This is noticeable in PC desktop use but becomes pretty hard to see in gaming, which is of course pretty much all you’d be doing from a console anyway. You can see our section about chroma sub-sampling from the older PG27UQ for more information about what it is, and the image impacts it has. On that screen 4:2:2 sub-sampling was needed for PC usage over DisplayPort as well to get to the maximum 144Hz refresh rate. It is not a major sacrifice in practice for dynamic content like games. HDMI 2.1 allowing for full 4:4:4 chroma would have been preferred, but at least this is something.
While we didn’t manage to get FreeSync over HDMI working from an AMD graphics card, we have been in touch with Asus who confirm that the screen CAN support VRR over HDMI from an Xbox Series X. This does at least mean for Xbox you should be able to use VRR. Without HDMI 2.1 the screen lacks HDMI-org VRR which is the standard used by Sony though for variable refresh rates on their latest gen console and they do not at this time support VRR over HDMI 2.0.
Read our detailed article about input lag and the various measurement techniques which are used to evaluate this aspect of a display. The screens tested are split into two measurements which are based on our overall display lag tests and half the average G2G response time, as measured by our oscilloscope. The response time element, part of the lag you can see, is split from the overall display lag and shown on the graph as the green bar. From there, the signal processing (red bar) can be provided as a good estimation of the lag you would feel from the display. We also classify each display as follows:
The total lag measured was a very impressive 2.50ms total. The pixel response times account for basically all of that, and so we can say that there appears to be basically no signal processing lag on this screen which is excellent, and to be expected thanks to the use of the Native G-sync hardware module. A solid result from this display and making it suitable for fast and competitive gaming.
If you enjoy our work and want to say thanks, donations to the site are very welcome. If you would like to get early access to future reviews please consider becoming a TFT Central supporter.
What is going to separate this screen from other emerging 4K 144Hz displays is the HDR performance. There are other 32″ 4K 144Hz monitors planned that will be much lower cost than this model including Asus’ own ROG Swift PG32UQ for instance, but they will be fundamentally different when it comes to the backlight and HDR capability. One of the main reasons for the high cost of the PG32UQX is the Mini LED backlight which performed extremely well we felt. The HDR experience was excellent, with a really well configured HDR mode, wide colour gamut and 10-bit colour depth support. What really set this screen apart was the eye-scorching >1600 nit peak brightness and the local dimming capability of the 1152 zone backlight. The contrast was excellent, the bright areas really popped and it was even very hard to find any issues with blooming or halos while viewing real content. If you want an amazing HDR experience this display will be far beyond any of the lower cost 32″ 4K 144Hz models that are planned.
In other areas we were impressed by the very good factory calibration and setup. The use of a low-glow panel was a pleasant surprise. The contrast ratio was surprisingly good for an IPS panel too, and we of course appreciated the solid all round performance that this panel technology brings for image quality and viewing angles. The design of the screen is starting to feel a little dated now, but it’s sturdy and well built and there’s quite a few extras Asus have added that a lot of people will probably like including the RGB lighting effects for instance. The foot of the stand is a bit of an annoying shape, and while the OLED dashboard display was a nice idea, the flashing ROG logo needs to be disabled as it’s just too distracting and annoying. Some people will be disappointed by the lack of HDMI 2.1 for latest gen consoles but Asus have done a reasonable job of improving support there where possibly over HDMI 2.0 to help, and console gaming isn’t really a target use for this display. At the end of the day the alternative in order to provide HDMI 2.1 was to drop the G-sync module which would have made the Mini LED backlight for gaming, and all the benefits that brings, very hard to include or use.
Check Availability and Pricing – Affiliate LinksTFTCentral is a participant in the Amazon Services LLC Associates Programme, an affiliate advertising programme designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.co.uk, Amazon.de, Amazon.ca and other Amazon stores worldwide. We also participate in a similar scheme for Overclockers.co.uk.
This has to be one of the most eagerly anticipated displays of recent times! The LG 27GL850 has attracted a lot of attention since it was announced, as an attractive alternative to the fairly wide range of 1440p IPS high refresh rate gaming monitors already available on the market. It’s been over 4 years since we saw the first screen of this type appear in the form of the Acer Predator XB270HU, followed by many popular options like the Asus ROG Swift PG279Qand Eizo Foris FS2735 for example. These models have made use of the high refresh rate IPS-type panels from AU Optronics which have offered excellent response times and performance over the years although have been criticized in some places for their quality control and sometimes high levels of IPS glow and backlight bleed. More recent high refresh rate IPS alternatives like the Gigabyte Aorus AD27QD for example have also been released, but based on an alternative panel option from Innolux.
The LG 27GL850 marks the first screen of its type to feature one of LG.Display’s new 144Hz IPS panels, from their ‘Nano IPS’ generation. Of particular interest are the wide gamut backlight (98% DCI-P3 quoted) and the advertised 1ms G2G response time, which has apparently been verified by an independent team at Underwriter Laboratories (UL). We look forward to testing that ourselves and understanding what test conditions might have applied to achieve such an impressive number, if indeed it is achieved.
Check Availability and Pricing – Affiliate LinksTFTCentral is a participant in the Amazon Services LLC Associates Programme, an affiliate advertising programme designed to provide a means for sites to earn advertising fees by advertising and linking to Amazon.com, Amazon.co.uk, Amazon.de, Amazon.ca and other Amazon stores worldwide. We also participate in a similar scheme for Overclockers.co.uk.
The 27GL850 offers a good range of modern connectivity with 1x DisplayPort 1.4 and 2 x HDMI 2.0 offered for video connections. For PC connectivity the DisplayPort is the most common option, with HDMI being available then for connecting external games consoles or Blu-ray players potentially. Both will support variable refresh rates from AMD graphics cards or games consoles. From NVIDIA cards you can only use VRR over DisplayPort.
The materials were of a good standard and the build quality felt decent. A fairly functional and modest design compared with some other gaming displays, with no RGB lights or fancy features. There was no audible noise from the screen, even when conducting specific tests which can often identify buzzing issues. The whole screen remained pretty cool even during prolonged use.
The LG 27GL850 features an LG.Display LM270WQA Nano IPS technology panel. This is the first time we’ve seen this new panel used offering a high refresh rate and 1440p resolution. It is part of LG.Display’s Nano IPS generation and so offers a wide colour gamut as well. This is rare in this part of the market, with the wide range of screens based on the AU Optronics IPS panel all being a standard sRGB gamut.
The panel offers 1.07 billion colours using an 8-bit + FRC colour depth. Some people may complain that the panel is not a native full 10-bit panel, but in reality you are going to be very hard pressed to see any real difference in practice between a good 8-bit+FRC panel and a true 10-bit panel. Even more so when you consider whether you’re going to actually be able to use the 10-bit colour depth in real use. Whether or not you can make use of this 10-bit colour depth depends on your graphics card and applications although many consumer gaming graphics card will be able to use it for modern games and HDR. You would need a high end professional grade graphics card to make use of 10-bit in pro applications like Photoshop though.
This section tests the full range of luminance (the brightness of the screen) possible from the backlight, while changing the monitors brightness setting in the OSD menu. This allows us to measure the maximum and minimum adjustment ranges, as well as identify the recommended setting to reach a target of 120 cd/m2 for comfortable day to day use in normal lighting conditions. Some users have specific requirements for a very bright display, while others like a much darker display for night time viewing or in low ambient light conditions. At each brightness level we also measure the contrast ratio produced by the screen when comparing a small white sample vs. a black sample (not unrealistic full screen white vs. full screen black tests). The contrast ratio should remain stable across the adjustment range so we also check that.
Graphics card settings were left at default with no ICC profile or calibration active. Tests were made using an X-rite i1 Display Pro colorimeter. It should be noted that we used the BasICColor calibration software here to record these measurements, and so luminance at default settings may vary a little from the LaCie Blue Eye Pro report you will see in other sections of the review.
We have plotted the luminance trend on the graph above. The screen behaves as it should in this regard, with a reduction in the luminance output of the screen controlled by the reduction in the OSD brightness setting. This is a linear relationship.
We restored our graphics card to default settings and disabled any previously active ICC profiles and gamma corrections. The screen was tested at default factory settings using our new X-rite i1 Pro 2 Spectrophotometer combined with LaCie’s Blue Eye Pro software suite. An X-rite i1 Display Pro colorimeter was also used to verify the black point and contrast ratio since the i1 Pro 2 spectrophotometer is less reliable at the darker end.
CIE Diagram – confirms the colour space covered by the monitors backlighting in a 2D view, with the black triangle representing the displays gamut, and other reference colour spaces shown for comparison. Usually shown as a comparison against the common sRGB space
dE average / maximum – we aim for as low as possible. If DeltaE >3, the color displayed is significantly different from the theoretical one, meaning that the difference will be perceptible to the viewer. If DeltaE <2, LaCie considers the calibration a success; there remains a slight difference, but it is barely undetectable. If DeltaE < 1, the color fidelity is excellent.
Note that we turned the Smart Energy Saving setting off in the OSD menu before taking these measurements. Initially out of the box the screen was set with the ‘Gamer 1’ preset mode. The display was set with a maximum 100% brightness which was too bright and uncomfortable to use for long periods. You will want to turn that down. The screen looks vivid and colourful thanks to the wide gamut backlight. We went ahead and measured the default