Figure 1.
Principle of operation of a normally white TN LCD panel.
When no electric field is applied (a), the helical structure of the LC molecules rotates the vertically polarized light so that it can pass the second, horizontal polarizer. When an electric field is applied (b), the molecules tend to align with the electrical field, distort and finally break the helical structure so that the backlight is blocked by the horizontal polarizer and the respective subpixel appears opaque.
Figure 2.
Main components of the LCD luminance transition signal.
(a) shows the recording of a luminance transition from 127 to 255
(maximal luminance of the monitor) for 10 frames and then back to 127
on a Dell 2709 LCD panel. (b) shows the pure transition signal which was generated from (a) by filtering the backlight modulation. (c) shows the backlight modulation signal. Note that (a) is composed of the product of (b) and (c).
Table 1.
Types of LCD monitors.
Figure 3.
Schematic comparison of CRT and LCD luminance signals.
For a single frame presentation of a white object on black background, the CRT signal reaches its maximum rapidly after frame start and decays to nearly zero a few milliseconds later. In the subsequent frame there is still a small phosphor activation at frame start although the frame is supposed to be black. Such a ground activation occurs inevitably when the electron beam traverses the pixel. In contrast, the LCD signal rises considerably slower and holds at maximum until the end of the frame. In the subsequent frame it falls back to its black level.
Figure 4.
Schematic of the different types of dynamic capacitance compensation (DCC).
Figure 5.
Backlight modulations are usually not phase locked to the refresh rate.
The plots combine the recordings of (a) two rising or (b) falling transitions which start at different phases of the backlight signal. Obviously, the resulting transition signals differ substantially.
Figure 6.
Calibration can prolong response times.
The uncalibrated 0% 100% transition shown in (a) has a response time of
ms. Calib
not only increases the amplitude of the backlight ripple a lot but also shifts the target signal from rgb
to rgb
. The resulting transition signal shown in (b) has the considerably longer response time of
ms.
Figure 7.
Examples of the variation of response times (RTs) over different luminance transitions for four differenc monitors.
The small red bars on the top of each bar denote the standard deviation over the five independent measurements. Below the RT bar plots RT mean and standard deviation over the different luminance levels, the coefficient of variation, means over all rising and falling transitions, the transition times from black to white and vice versa, and the manufacturer’s RT specifications are shown.
Figure 8.
Response times (RT) variability over repeated measurements of the same luminance transition.
The plot shows a periodically blinking gray patch between 128 rgb and 191 rgb
for 10 frames per luminance level on a HP LP2480 ZX monitor. Two subsequent falling response times differ substantially.
Figure 9.
Substantial overshoot due to improper DCC.
For illustrative purposes, we specified the time in frames instead of milliseconds. The raw signal was recorded from a gray (25% luminance, calib) patch displayed for 10 frames on a Samsung XL 30 monitor. The response time measured between the 10% level and the 110% level is ten times greater than the 10%/90% response time according to the ISO standard.
Figure 10.
Frame response for static presentations.
(a): part of the PSD of 10 different LCD monitors. (b): Constant signal (50% of the monitor’s luminance maximum) of the monitor with the maximal power at the refresh rate.
Figure 11.
Luminance stepping leads to saturations of the luminance signal before the target level is reached.
The measurement of a transition 63 rgb
127 rgb
(10 frames)
63 rgb
of a BenQ V2400W monitor (uncalibrated) is shown. The target level is not reached in the first two frames of each transition. In addition, the frame response is noticeable at the upper luminance level in absence of any controlled luminance transition.
Figure 12.
Visible motion blur in units of just noticeable differences (JNDs) calculated from the luminance transitions shown in Fig. 7.
See Methods section for details about the motion blur model and respective calculations. The summarizing numbers below each subplot are analogous to those of Fig. 7.
Figure 13.
Impact of the Motion Picture (MP) mode of the NEC 24WMGX monitor on visible motion blur.
With the MP mode disabled (a), the motion blur profile is similar to the typical profiles of other monitors shown in Fig. 12. With the MP mode set to its strongest level (b), the visible motion blur had decreased by about 50% on average.
Figure 14.
Motion Picture (MP) mode of a NEC 24WMGX monitor.
The left hand side plots show luminance signal measurements of a green patch which appears for 10 subsequent frames, followed by 10 black frames, periodically. In the upper row, MP is switched off, in the lower row it is switched to the highest possible level for this monitor. The plots on the right hand side show the respective power spectral densities (PSD) of frequencies between 20 Hz and 200 Hz for the constant level signals (100% green). Obviously, if MP mode is enabled, the dominant backlight frequency of 89 Hz is so weak relative to the strong MP amplitude that it disguised in the PSD.
Figure 15.
Color shifts during a rising 0 rgb
255 rgb
transition of an uncalibrated BenQ V2400W monitor.
Frame boundaries are indicated by the vertical dotted lines. The target luminance was white. Obviously, the luminance distribution of the three color primaries changes over time, as the primaries have different response times. The red primary is fastest whereas the other two primaries are subject to luminance stepping. Therefore, the transition has a red color cast which disappears first during the third frame. The dispersion of the signals is illustrated by the coefficient of variation of the three color primary luminances. The color bar at the bottom sketches the color change of the display over time. Note that the appearance of the colors depends on the calibration of your display and is only a rough approximation to the true color of the transition on the BenQ monitor.
Figure 16.
–tilt voltage. The measurement of the Dell 2408 monitor, green primary, transition 0% 50% (10 frames)
0%, shows that the rising transition starts one frame earlier than expected.