BenQ XL2411Z

From Display-Corner
Jump to: navigation, search

Contents

Specifications

Brand:BenQ
Model:XL2411Z
Size:24"
Resolution:1920x1080
Panel type:TN
Max. refresh rate:144 Hz
Panel:AUO M240HW01 V8
Processor:MST 8556T (MSTAR)
Backlight type:LED (white, edge)
LED driver:MP3398 (Monolithic Power)
Price (approx.):USD 300
Available since:12/2013








Reviews

BenQ XL2420Z/XL2411Z review on PC monitors.

BenQ XL2720Z review on TFT central.

Comparison with other LCDs

The BenQ XL2411Z is more or less identical to the BenQ XL2411T except that it offers a feature called Motion Blur Reduction, BenQ's alternative to 2D LightBoost. Note that the Z-series monitors still support NVIDIA's 3D LightBoost, but the Strobelight Utility, which was used to enable 2D LightBoost and also worked with AMD graphics cards, does not work anymore with the Z-series monitors. LightBoost can be still used for 2D, of course, but needs to be enabled with NVIDIA 3D VISION hard- and software.

LightBoost is only supported for 100Hz and 120Hz refresh frequencies, whereas Motion Blur Reduction can be used for frequencies up to 144Hz.

Regarding image quality, the XL2411Z is also similar if not identical to the XL2420Z. The XL2420Z has just a fancier design, comes with a remote control (not wireless though), and provides more input options (DisplayPort, 2nd HDMI, USB).

Service menu BenQ XL2411Z

How to call up the service menu

Have the monitor connected to a valid signal source. The service menu can be entered by keeping the 2nd button from the left ("down" button) pushed while powering on the monitor. Pushing the 4th button from the left ("menu" button) will then toggle the service menu on and off. In order to access the OSD menu, make the service menu disappear and then push any other but the "menu" button so that the icons above the buttons appear. Then you can push the "menu" button to enter the OSD menu. The "menu" button will react normal again after the next power-on.

Firmware

As a first in computer monitor's industry, BenQ offers firmware upgrades for its Z-series monitors. When these monitors were first released they came with a firmware that had a sub-optimal Motion Blur Reduction implementation. This brought BenQ to offer customers to send in such monitors for an upgrade or, as an alternative, allow enthusiasts to do the upgrade themselves (which, however, requires additional programming hardware). See BlurBuster's firmware upgrade instructions for a step-by-step upgrade how-to. Unfortunately, BenQ does not keep the firmware download page up to date. Currently (mid 2015), monitors are shipped with firmware version 4 which is said to implemented a better overdrive when Motion Blur Reduction is active. As far as the flicker-free mode is concerned, there is no difference between firmware version 2 and 4 though. The results shown below have all been measured with version 2.

Note that, although the XL2411Z as almost identical to the XL2411T, upgrading a BenQ XL2411T with the Z-version firmware would not work (see below, Accessing the backlight signals).

Motion Blur Reduction (MBR)

Motion Blur Reduction (MBR for short) is BenQ's alternative to 2D LightBoost. Like LightBoost, MBR pulses the backlight once per refresh interval. In contrast to LightBoost, it does so without adopting a special LC panel update timing and without using different overdrive parameters for the different pixel lines, both of which should affect settling behavior negatively. That is, LightBoost has the higher potential to get things right.

MBR features at a glance

On the other hand, BenQ offers more options regarding MBR, at least with firmware version v2. Besides letting the user activate MBR at frequencies other than 100Hz and 120Hz and without requiring specific NVIDIA graphics hardware, it is also possible to tweak the backlight pulse parameters, like the pulse width and the pulse phase with respect to VSync in great detail. This is possible either via the service menu (parameters strobe duty and strobe phase), or with BlurBuster's Strobe Utility, or via direct DDC/CI programming (Display Data Channel/Command Interface). Tweaking the pulse parameters is especially useful if the LC panel update is accelerated by tweaking the monitor timing.

Backlight pulse width

Mode  min [ms]   max [ms] 
LightBoost 100Hz 1.88 3
LightBoost 120Hz 1.4 2.25
MBR 60Hz 0.17 5
MBR 100Hz 0.1 3
MBR 120Hz 0.08 2.5
MBR 144Hz 0.07 2.1

As soon as strobed backlight is used, be it for LightBoost or for MBR, the backlight LED current is increased by a factor of about 1.8 in order to compensate, at least partly, for the time the LEDs are turned off between the pulses.

The pulse width can be controlled either in 10 steps (LightBoost) or in 30 steps (MBR), linearly within the range given in the table. In MBR mode, also the pulse phase can be modified, in steps of approximately a 100th of the refresh cycle time, where a value of 0 makes the backlight pulse onset roughly coincide with the update of the first pixel line. The maximal value is 100 which suggests that the pulse onset can be arbitrarily chosen anywhere within the refresh cycle. Unfortunately, this is not the case as the pulse phase is limited to a value that makes the pulse offset (not onset!) coincide with a time shortly before the first pixel line is updated. In other words, the update of the first pixel line occurs either before or after the backlight pulse, but cannot occur during the pulse.

Monitor timing (tweaking it)

As mentioned above, the monitor timing can be tweaked in order to speed up the update process of the LC panel. This is done by virtually making the vertical synchronization phase longer while preserving the vertical refresh rate, as described in the guide to the aforementioned BlurBuster's Strobe Utility. For 120Hz, the maximal possible total number of vertical lines is 1350, for 100Hz it is 1500 (empirical findings). These timing modes push the limits regarding pixel clock frequency and it might be worthwhile reducing the total number of pixel columns from 2080 to 2020 in order to lower the line and pixel clock frequencies. Depending on the graphics card it might be necessary to use different sync polarities and within-sync timing values (back porch, front porch, sync duration) in order to make the monitor accept the unconventional timing. It is not fully clear yet, whether at all or in how far the image quality depends on the sync timings. It should be noted though, that whenever the monitor timing deviates too much from any standard timing – which is clearly the case for the timings mentioned here –, the monitor adopts the 60Hz backlight pulse widths. This is somewhat problematic as it allows to choose pulse durations which could shorten the lifetime of the LEDs. At 120Hz, for example, the pulse width can be set to 5ms, which is the intended maximum for 60Hz though and, thus, pushes the average LED current at 120Hz way beyond the limit considered safe by the manufacturer. Moreover, the tweaked monitor timings might not work at all (i.e., black screen) if the value for the pulse phase is too high. So, know what you are doing! If, by some mishap, the monitor stays black and does not revert back to a working mode by itself anymore, the only way to disable MBR mode might be to power off the monitor, disconnect it from the power line and possibly even from the computer, let it rest for a while until fully discharged (trying to switch it on in the disconnected state helps with discharging), and then reconnect it again.

By the way, the within-sync timing does not seem to have any effect whatsoever on how long the software blocks when executing the OpenGL SwapBuffers();glFinish() function sequence or when the LC panel is updated (latency). This is the case, at least, with the tested NVidia driver and hardware but, technically, it does not have to be this way. So the behavior might be different with other driver versions or, for example, with AMD graphics cards.

MBR and AMA

AMA is BenQ's implementation of overdrive and care should be taken when changing the AMA setting before or while in MBR mode. The thing is that the internal overdrive parameters according to AMA="High" or AMA="Premium" are changed during the activation of MBR, so there are actually hidden AMA settings "MBR/High" and "MBR/Premium". However, when changing the AMA setting while already in MBR mode, the MBR-specific overdrive parameters (i.e., according to "MBR/High" or "MBR/Premium") are overwritten by the parameters normally used in non-MBR mode (i.e., according to "High" or "Premium"). In other words, the order in which to set AMA and to enable MBR mode matters, except for AMA=off which always results in overdrive being shut off.


Comparison of white-points over all gray levels for LBL=0 (LBL deactivated, shown in blue) and LBL=10 (maximal LBL, shown in red), after the white-point has been re-calibrated to D65 (CIEx,y=0.3127,0.3290) at the highest intensity level (RGB(LBL=0)=99,95,99; RGB(LBL=10)=87,83,100}. Each small dot represents the color measurement for one of 256 different gray levels, and the dots are connected according to the order of gray levels, from maximal luminance (big dot, D65 calibration) to background luminance. These trajectories in color space are basically what gamma curves are in luminance space.
Obviously, there are no differences of practical relevance; all the residual differences can be attributed to round-off noise (pixel value processing in the monitor), measurement noise, and drift effects.

Low Blue Light

With the Z-version of the XL-monitors, BenQ included the "Low Blue Light" feature (LBL for short), which can also be found in BenQ's other "Eye care" monitors. Apparently, BenQ wants customers to believe that the blue light emitted by the monitors is harmful to the eyes and that LBL does something about it. The argumentation is extremely far-fetched and all the talk about blue light hazard, also in other contexts, and the LBL feature in particular, is just a commercial gimmick. This does not mean that light of short wavelength (like ultra-violet), applied at high intensities, can not be harmful. But neither the wavelength nor the intensity of the light emitted by computer monitors or TV screens is even close to what could be considered harmful. Nevertheless, blue light is assumed to have some modulating effect on the circadian rhythm and on the attentive state and alertness. This might provoke long working or gaming hours, which is obviously less healthy than having a good night's sleep instead. Whether a difference between a rather cold color setting and a warmer color setting (i.e., using the LBL feature) is sufficient to get any relevant effect of such kind is questionable. As far as gamers are concerned, who are the primary target customers of BenQ's XL-monitors, it is also questionable whether they even would benefit from being exposed to less blue light that otherwise might keep them longer awake and more alerted.

From a technical point of view, LBL is nothing else but a shift of the white-point away from blue, so it is nothing that could not be achieved also with the normal color settings. If LBL was really about cutting down high blue intensities, one could think of ways of doing this with minimal effect on the image appearance, for example, by having LBL affect only bright colors or only color regimes where a reduced blue component might not be that obvious to the observer. The review on TFT central somewhat suggests that something like this had actually been implemented in the BenQ monitors, suggesting that LBL is more than just a simple white-point shift. They measured the light spectra with and without LBL after having re-calibrated the monitor to the 6500K white-point, and they claimed that the blue components were indeed reduced by LBL. However, also other components of the spectrum were reduced and the residual differences between both spectra, with and without LBL, can be well explained by the slightly different white-point calibrations. Getting really different spectra for one and the same white-point would only be possible if the spectra of the primary colors changed, which is obviously nothing that could be done that easily. Having different white-points for different intensities would be possible though, whether this was useful or not. But as measurements show (presented in the figure at the right), there is no practically relevant difference in the white-point shift across different gray levels between LBL on and LBL off. So all LBL does is to apply another adjustment to the RGB gains on top of the user settings.


Settling behavior

There are three backlight operating modes of interest regarding settling behavior: Flicker-free, 2D LightBoost, and Motion Blur Reduction (MBR for short). LightBoost and MBR are very similar in that they use strobed backlight (one strobe per refresh interval) instead of the steady-state backlight used in flicker-free mode. For all three modes it is advantageous to activate overdrive at a medium level, which is done by setting AMA="High" (before enabling MBR).

Unless otherwise mentioned, all measurements shown here were taken after having reset the monitor (via the monitor's OSD menu) and having set "Picture mode" to "Standard". The monitor was operated at 120Hz refresh frequency with the timing tweaked to maximize duration of the vertical sync phase (total lines = 1350). The pulse durations in the strobed backlight modes were set so as to result in a maximal average luminance (i.e., for white) that is as close as possible to 100 cd/m2 (LightBoost=10% results in pw=1.4ms and 117 cd/m2; MBR-Duty=7 results in pw=1.1ms and 107 cd/m2). Note that LightBoost seems to be too dark in comparison to MBR, according to the measured  cd/m2-values and given that the LightBoost pulse width is approx. 30% longer than the MBR pulse width (107·1.4/1.1≈136). But this is just because the default Contrast setting for LightBoost is 37%, whereas it is 50% for MBR.

The pulse phase was set to 0 (affecting MBR mode only), which appears to provide a good balance of settling errors at the top and the bottom of the screen.

Photodiode with rubber sleeve put directly on the screen for measuring the stripe.

The measurements have been taken with a photo diode PDA36A (Thorlabs), the gain of which was set to 60dB which results in a bandwidth of 37.5kHz and a minimal rise/fall time (10%-90%) of about 9µs. The photodiode was placed at about 4.5cm from the screen surface at a straight angle. Ambient light was kept from the measurement by a rubber sleeve of 3cm diameter which also limited the maximal incident angle to about ±20°.

The vertical extent of the measured screen area was not only limited by the rubber sleeve but also by the stimulus being just a small horizontal stripe covering 5% of the screen height. Note that the LC cells are updated sequentially from the top of the screen to the bottom, which results in different delays for the luminance curves depending on the vertical measurement position. By limiting the measurement to only 5% of the full vertical screen size, the smear effect introduced by averaging over differently delayed luminance signals becomes close to irrelevant. For example, for a refresh frequency of 120Hz the screen is updated within around 8ms, so if the true luminance would change instantly, the measured rise time would be 5%·8ms = 0.4ms, which is negligible here.

For more detailed information on the measurement method and the presentation of the results, see Flicker-free settling and LightBoost settling.

Settling curves

Figure 1 and Figure 2 show the luminance signals for the horizontal stripe being switched between gray levels 0%-100%, 25%-75%, and 45%-55% respectively. Note that these percent values refer to pixel values and not to luminance values, so the monitor's respective gamma setting has to be taken into account when interpreting the absolute luminance levels in the plots.


Figure 1: Settling curves for three different step sizes without overdrive (AMA=off, left) and with overdrive (AMA=high, right). The curves are shifted and scaled individually according to the respective step size.
The vertical gray lines mark the time when the OpenGL command sequence SwapBuffers();glFinish(); returns control to the program, which is about when the 1st line of a frame is sent out to the monitor. Measurements have been taken at the vertical center of the screen.


Figure 2: Settling curves for three different step sizes with LightBoost at 10% (left) and with Motion Blur Reduction at StrobeDuty=7, StrobePhase=0, and AMA="MBR/High" (i.e., AMA="high" before having activated MBR).


Figure 2 actually shows the best or close to best case as it refers to the settling situation at the middle of the screen. Figure 3 shows how it looks like at the top and the bottom of the screen while in MBR mode. The situation at the top of the screen could be improved by changing the strobe phase, but this would deteriorate the situation at the bottom of the screen. StrobePhase=0 was chosen for being a good compromise between approving the situation at the bottom of the screen without screwing up the situation at the top of the screen too much. When optimizing StrobePhase we face two problems. One is that we can shift the backlight pulse only in one way (when going from StrobePhase=0) unless we are willing to make a big step in the opposite direction, namely a step of one pulse width. The second problem is that we gain relatively little and lose relatively much when increasing StrobePhase. This is just because of the different positions (in time) at which the pulse starts and ends relative to the settling curves. So shifting the pulse forward in time has more (negative) impact at the top of the screen. On the other hand, exactly because a pulse shift with respect to the settling curve has a smaller impact at the bottom of the screen, also the settling errors do not depend so strongly on the vertical position and thus extend over a wider (i.e., higher) screen region at the bottom of the screen. So one could argue that although the improvement of the local settling error at the bottom is comparatively small given the loss at the top, it is still worthwhile doing so as the screen region over which the settling error has to be accumulated is larger at the bottom than it is at the top. But this becomes a matter of preference then, that is, how bad a local but higher settling error is when compared to a lower but spatially more extensive settling error.


Figure 3: Settling curves for three different step sizes with Motion Blur Reduction (StrobeDuty=7, StrobePhase=0, AMA="MBR/High") at the top of the screen (left) and the bottom of the screen (right) – directly comparable to the left panel in Figure 2.
Note that the %-values refer to the programmed pixel values. Obviously, there is quite some gamma shift across screen locations, because if there wasn't the settled levels should be equal.
Notice how the integrals (i.e., the pulse energies) of the first high pulse (right panel) and the last high pulse (left panel) are somewhat off target because the backlight strobe comes too early or too late with respect to the respective pixel settling curves. Likewise, the low pulses directly before and after the 6-frame-long onset phase are affected.


Matrix measurements

Flicker-free backlight

According to the settling behavior measurements for several monitor settings, AMA=high at Contrast=40% in "Standard" picture mode is on par with the "FPS1" picture mode (the factory preset) and provides the best settling performance (Figure 4). Contrast=40% is somewhat better than the default, Contrast=50% (matrix plots for AMA=high,C=50% [2nd view]), especially regarding some worst case step sizes, which is probably because of the additional overdrive margin that comes with the Contrast=40% setting. On the other hand, whenever the target luminance is the maximal luminance, settling times are much better with Contrast=50%, but this is just because the panel is maxed out already for maximal pixel values, and it is not overdrive which controls the luminance then but the panel being in saturation.

Surprisingly, the AMA=off case is not really bad either in comparison to the FPS1 and Standard/AMA=high modes, except for the much higher delay deviations (matrix plots for AMA=off [2nd view]). Figure 5 provides a comparison of several test cases.

Figure 4: [2nd view]. Settling measurements with AMA=high and Contrast=40% at 120Hz. For further details on the graphs and the measurement method see Flicker-free settling.


Figure 5: Comparison chart for the some flicker-free backlight test cases (120Hz). Note that smaller bars/values are better. The colored bars refer to the maximum of 90% of all step sizes (max90) and the gray bars refer to the maximum of 100% of all step sizes (max100). For further details on the measurement method see Flicker-free settling.


Strobed backlight

Identifying the optimal setting for the strobed backlight modes boils down to deciding between LightBoost and Motion Blur Reduction (MBR), as both are optimized differently. While LightBoost performs much better at the top of the screen, MBR is much better at the bottom of the screen (Figure 6 and Figure 7). This is somewhat surprising given that LighBoost has two technical advantages over MBR: a faster panel update and pixel line specific overdrive. But obviously, BenQ decided to choose an error balancing (top vs. bottom) which causes the bottom of the screen to fall way behind. Of course, LighBoost is not made for 2D but for 3D, where the timing of the shutter glasses has to be taken into account. This timing, however, is mediated by the external IR emitter which is controlled by the PC and the NVidia driver, so it is likely that BenQ had to stick to some LightBoost timing standard dictated by NVidia.

Note that the MBR modes were not tested with BenQ's standard values for StrobeDuty (20) and StrobePhase (100). StrobeDuty controls the maximal luminance and we have just chosen a value (StrobeDuty=7) that results in approx. 100 cd/m2. More interesting is the StrobePhase value, which BenQ chose to be close to the LightBoost setting as a value of 100 minimizes settling errors at the top of the screen. We, however, have chosen a value (StrobePhase=0) which results in a more balanced error distribution across the screen. Regarding the overdrive setting (AMA), it is clear from the measurements that it is best to use the standard which is setting AMA="high" before activating MBR. Using Contrast=40% results in a rather minor improvement over the standard setting (Contrast 50%). See Figure 8 for the comparison chart.


Figure 6: [2nd view]. Settling error measurements for LightBoost=10% and 120Hz. Note that the Z axes are scaled the same within columns (i.e., across screen locations) but not within rows (i.e. across refresh cycles).
For further details on the measurement method and presentation see LightBoost settling.


Figure 7: [2nd view]. Settling error measurements for Motion Blur Reduction (MBR) (StrobeDuty=7, StrobePhase=0, Contrast=40%, f=120Hz). Note that the Z axes are scaled the same within columns (i.e., across screen locations) but not within rows (i.e. across refresh cycles).
For further details on the measurement method and presentation see LightBoost settling.


Figure 8: Comparison chart for the some 120Hz strobed backlight test cases. Note that smaller bars/values are better. The colored bars refer to the maximum of 90% of all step sizes (max90) and the gray bars refer to the maximum of 100% of all step sizes (max100). The scaling of the bar lengths is the same within columns (i.e., for a particular refresh cycle), but different within the rows.
For the AMA=high and AMA=off conditions the AMA setting has been changed after MBR was activated, whereas for the other conditions AMA was set to "high" before MBR was activated.
For further details on the measurement method see LightBoost settling.


Connector on the power supply board
Wire color Signal
black GND
black GND
n.c.
violet Effectively not connected to
the power supply board
red 5V (supply)
red 5V (supply)
gray Effectively not connected to
the power supply board
green Backlight on/off (3.3V, active high)
brown Fast PWM (18kHz, 3.3V, active high)
white +3.3V (supply)
Connector on the controller board
Wire color Signal
brown Fast PWM (18kHz, 3.3V, active high)
white +3.3V (supply)
black GND
black GND
red 5V (supply)
red 5V (supply)
gray 3.3V
n.c.
violet 0V
green Backlight on/off (3.3V, active high)

Accessing the backlight signals

The controller board of the XL2411Z seems to be identical to that of the XL2411T, but the power supply board with the LED driver is different, at least for the two monitors (Z and T) that were opened up. Especially the LED driver and its wiring has changed for the Z-version, and with it the function of some controller signals.

One signal (brown wire) carries an 18kHz PWM signal that controls the LED average current. At a Brightness setting of 100%, the PWM ratio is either 50:50 for the flicker-free backlight mode or 100:0 for the LightBoost and MBR modes (i.e., always high). Under some circumstances, e.g. for LightBoost at 100Hz, a ratio of about 96.5:3.5 can be observed, but this is close enough to 100:0.

The other signal (green wire) turns the LEDs on or off. So this signal is either constantly high while in the flicker-free mode or carries the strobe pulse signal while in the LightBoost and MBR modes. This signal works like a gating signal for the fast PWM signal. However, the fast PWM signal does not feed through all the way to the LED luminance (although the LED and the photodiode can go that fast), so the fast PWM signal is converted to some analog maximal LED current by the LED driver and the fast PWM frequency cannot be found in the luminance traces anymore – the backlight is truly flicker-free.

In the BenQ XL2411T, the brown wire also carries a PWM signal, which is normally driven at a frequency of around 180Hz. In the T-version, this signal is used for brightness control while in non-LightBoost mode or as strobe signal while in LightBoost mode. The green and gray wires, on the other hand, have constant signal levels depedning on the current mode and just code for 3 different but fixed LED currents (LightBoost off, LightBoost at 100Hz, and LightBoost at 120Hz). Assuming that the T-electronics could be programmed with the Z-firmware and the controller boards were identical, then the brown wire would carry the fast PWM signal and would indeed control the LED brightness – just like it is normally done for PWM controlled brightness except for the much higher PWM frequency (18kHz instead of 180Hz). This might increase the switching loss at the LED driver but is not necessarily a problem. What is a problem though is the strobe signal (green wire) being just interpreted as a signal to switch the LED current between two levels that are higher than even the normal "LightBoost off" level (note that the gray wire would always be at a high level, which would cause some boost of the current all the time). So the strobe signal could never turn off the LED completely and would be rather ineffective for reducing motion blur. Moreover, the average LED current would be too high while in LightBoost or MBR mode. This situation could be remedied though by tweaking the LED driver circuit.

Controller board (left) and power board (right), also hosting the LED driver.
Power and controller board unmounted and flipped over (top view).



Personal tools
Namespaces

Variants
Actions
General
LCDs
Tachistoscope
Toolbox