FEATURE. Adaptive Temporal Aperture Control for Improving Motion Image Quality of OLED Display

Transcription

1 Adaptive Temporal Aperture Control for Improving Motion Image Quality of OLED Display Takenobu Usui, Yoshimichi Takano *1 and Toshihiro Yamamoto *2 * 1 Retired May 217, * 2 NHK Engineering System, Inc Degradation of motion image quality by motion blur on hold-type displays, such as LCDs and OLED displays, is a well-known issue To improve motion image quality, a driving method with a shorter temporal aperture has been proposed However, a shorter temporal aperture requires higher instantaneous luminance on displays Higher instantaneous luminance accelerates the lifetime degradation of OLEDs Therefore, we have been developing a driving method with adaptive temporal aperture control for a longer lifetime and better motion image quality However, two image quality degradations were perceived when this driving method was applied One of these degradations was caused at the boundary between the different temporal apertures The other degradation was caused by switching the temporal aperture between frames In this paper, we propose transition area and period insertion methods to suppress these degradations and discuss the mechanism of these degradations We then confirm the effectiveness of our proposed methods by making subjective evaluations and calculating the lifetime of OLED displays when adaptive temporal aperture control is applied 1 Introduction The degradation of video image quality caused by motion blur is a well-known problem for hold-emission *1 display devices such as liquid crystal displays and organic lightemitting diode (OLED) displays 1)-3) Even though OLED displays have a fast response, they are still affected by motion blur that results from active-matrix driving because of the hold-emission operation To reduce motion blur, driving methods of higher frame rates and a shorter emission time within one frame have been proposed 4) 5) From the results of subjective tests on motion blur, Kurita et al concluded that a high frame rate of 36 Hz is required to obtain an image quality that is acceptable to viewers 6)-8) However, achieving such a high frame rate for large-screen high-definition displays requires a major improvement in device performance Therefore, a system that achieves, at a frame rate of 12 Hz, motion blur equivalent to that under operation at a frame rate of 36 Hz by reducing the emission time by one-thirds has also been proposed However, to maintain the same luminance with a shorter emission time, a higher instantaneous luminance is required The luminance of an OLED display is roughly proportional to the driving current and the device lifetime is inversely proportional to an exponential function of the driving current 9) 1), so the increase in instantaneous luminance needed to compensate for the shorter emission time accelerates the degradation of device lifetime To address the above issue, we proposed adaptive temporal aperture *2 control as a method of improving image quality while reducing the degradation of device lifetime and verified the effectiveness of the control by simulation 11) However, dynamically changing the emission time within the window or in the time direction creates a new problem of noticeable artifacts *3, and methods of suppressing such image quality degradation have also been studied 12)-15) We report here the results of subjective evaluation testing of a method for reducing image degradation, conducted using a prototype OLED display equipped with adaptive temporal aperture control The results show that the method is effective in suppressing image quality degradation and * 1 Continuous emission for the duration of one frame * 2 Duty of emission within a frame * 3 False images and shapes that appear as a result of the emission mechanism or image processing 13

2 improving the device lifetime Video player 2 Adaptive Temporal Aperture Control 21 Overview In a hold-emission OLED display, emission continues from the time data is written to each line until data writing for the next frame begins [Fig 1 (a)] When adaptive temporal aperture control is used, motion blur is suppressed by shortening the emission time [to 25% in the case shown in Fig 1 (b)] in the dynamic area of the image where motion occurs In the static area of the image, where there is no motion, emission is held for the entire frame (1% emission time) to avert the degradation of device lifetime by high instantaneous luminance This method can suppress image degradation from motion blur while also averting the degradation of the OLED display lifetime 22 Prototype OLED display testing system In the prototype adaptive temporal aperture testing system (Fig 2), the input from the video player is preprocessed Motion vector detector Scan driver Scan signal Signal converter Data driver 25-inch OLED panel Data signal Figure 2: Structure of prototype adaptive temporal aperture control evaluation system 1 frame Scan line 1 2 1,8 Scanning (data writing) Emission (continuous) (a) Ordinary hold-emission OLED display 1 frame Static areas (long emission time) 1 2 Scan line TAPT:25% TAPT:1% 1,8 TAPT: Temporal aperture Scanning Emission Dynamic area short emission time (b) Adaptive temporal aperture control Figure 1: Adaptive temporal aperture control 14

3 to detect movement vectors and extract the dynamic image area where motion blur is likely to be noticeable Then, the signal converter generates a scanning signal that produces a short emission time within the dynamic area and converts the image signal into a data signal for the compensation of luminance in the parts where the emission time has been shortened Finally, the drive signals are input to a data driver and a scan driver to drive the OLED panel The pixel circuit for the active-matrix *4 OLED panel used in the testing is shown in Fig 3 The temporal aperture of each pixel is controlled by the timing of the voltage applied to each horizontal line of OLEDs (Del), which is, in turn, controlled by a voltage drive signal (Vdrive) and a write signal (Vgate) Referring to a report that reducing blur to four pixels per frame or less is effective in suppressing motion blur 8), we defined the dynamic region of the image as the * 4 A driving method in which each pixel has transistors that control light emission for the pixel V data Tr 1 Tr 2 C 1 D el V gate V drive V drive drive power V gate scan line V data data line D el OLED Figure 3: Pixel circuit for prototype OLED panel region comprising horizontal lines in which there is movement in at least four pixels between frames, and we controlled the temporal aperture in this region to be 25% and the temporal aperture outside this region (the static region) to be 1% The control method illustrated in Fig 2 is considered in more detail The motion detector detects the dynamic area by dividing the image into blocks (16 16 pixels) and calculating a motion vector for each block by matching blocks between frames *5 The motion vectors are used to calculate the proportion of pixels for which there is motion of at least four pixels per frame for each horizontal line Lines for which the proportion exceeds a threshold are determined to be the dynamic area and the other lines are determined to be the static area The threshold value is set with consideration given to motion existing in an area above a certain size and excluding movement vector detection errors Here, we take the dynamic area to be horizontal lines for which the proportion of pixel movement is at least 5% and other areas to be static areas In the next step, the signal converter converts image data in accordance with the temporal aperture so that pixels are not displayed at different luminances for the same input signal level when there is a difference in the temporal aperture between the dynamic areas and static areas of the image Here, the data conversion is performed as illustrated in Fig 4 In ordinary acquired images, there is a nonlinear gamma relationship between the actual luminance and the signal level, so gamma conversion is performed on the signal level to make the relationship between the luminance and signal level linear S = S γ (1) 1 in Gamma conversion Luminance compensation S S In the above equation, S in is the input signal, S 1 is the signal after gamma conversion, and γ is the display gamma value Next, data is converted to correct the difference in luminance due to the difference in emission time (luminance compensation) In this prototype system, the luminance for Display characteristics conversion S out 1 = Disp ( S ) Figure 4: Signal processing for luminance compensation 2 * 5 A method in which the distance of movement between frames is measured by dividing an image into blocks (16 pixels 16 pixels, etc) and comparing each block to neighboring blocks in the previous frame or in the subsequent frame, and taking the distance of movement to be the distance to the block closest to the original block in terms of pixel value 15

4 the 25% temporal aperture, which is the shortest emission duration, is not changed, but the luminance for the areas of the 1% temporal aperture is converted to one-fourth S (2) In the above equation, d is the temporal aperture of the region and S 2 is the signal after data conversion The final step is a nonlinear conversion (display characteristic conversion) of the signal levels so as to match the OLED display voltage luminance characteristic We express the relationship between the display luminance L and the signal level S as Equation (3) L = Disp( S) (3) Then the output signal level from the display characteristic conversion Sout is expressed by Equation (4) S out 1 = Disp ( S ) 2 (4) The conversion described above makes it possible to correctly display the luminance that corresponds to the signal level 23 Temporal aperture control driver circuit In the prototype active-matrix OLED panel, the data signals (Vdata) are written to the pixels line by line in the images displayed at the emission intensity that corresponds to the data signal within the emission period (Fig 5) For arbitrary control of the emission period for each line, we prepared a prototype scan driver that has a scan shift register for the output of the scan signal and an aperture data shift register for the output of the temporal aperture control drive signal (Fig 6) The scan shift register outputs the gate signal (Vgate) to the horizontal lines sequentially in one scan period at a rate of one line per clock cycle [Fig 5 (a)] An example time chart for the aperture data shift register is illustrated in Fig 7 During the scan time for one horizontal line, the aperture data for each line is input to the shift register, with 1 indicating emission for the line and indicating no emission for the line The aperture data thus controls the drive current that is sent to each line in each scan interval By repeating this process for each scan interval, it Scan pulse Emission period Line 1 Line 2 Line 3 Line n Scan time (a) Scan signal Line 1 data signal Line k data signal Scan time (b) Data signal Figure 5: Drive signals of the prototype system 16

5 OLED panel V gate V gate Gate signal CLK1 Scan signal amplifier Scan shift register V drive V drive Latch signal Aperture data CLK2 Drive amplifier Latch Aperture data shift register Figure 6: Circuit configuration of the scan driver for adaptive temporal aperture control Aperture data for each horizontal line Aperture data CLK2 (a) Input signal Line 1 Line 2 Line 3 Line n Scan time (b) Example of output signal (drive current) Figure 7: Example of drive signal for adaptive temporal aperture control (a) OLED display with adaptive temporal aperture control (b) Scan driver circuit board Figure 8: Prototype display and scan driver 17

6 is possible to control the emission of light for each line to an arbitrary time period Photographs of the prototype adaptive temporal aperture control OLED display and scan driver are presented in Fig 8 The panel has a size of 25 inches (diagonal), a resolution of 192 by 18, and a frame rate of 12 Hz This prototype system has been confirmed to suppress motion blur in video images that include movement or scrolling text, for example, by temporal aperture control [Fig 8 (a)] However, verification of the basic operation in actual hardware revealed two major problems: artifacts exist at the control boundaries between dynamic regions and static regions in the image, and there is a noticeable blinking effect when the temporal aperture is switched between frames The causes of these problems are explained and methods of improvement are evaluated in the next section 3 Reducing Image Distortion with Transition Areas and Transition Periods 31 Mechanism behind and suppression of image quality degradation in the spatial direction The mechanism behind the image degradation that occurs when the emission pattern differs in adjacent image areas and a method for suppressing that degradation are illustrated in Fig 9 In Fig 9 (a), the vertical axis is the vertical direction of the image and the horizontal axis is time In the upper part of the figure, the emission time is long and the emission intensity is low; in the lower part of the figure, the emission time is short and the emission intensity is high Because the human eye senses brightness by integration over a certain period of time, brightness is perceived as the same in both areas if the line of sight does not move However, if the line of sight moves in the vertical direction on the screen (when tracking), integration over the movement of the line of sight (indicated by the blue line and arrow in the figure) results in perception of darkness between both areas as shown in Fig 9 (b) Therefore, a transition area in which the emission time is gradually changed, as shown in Fig 9 (c), is established As a result, the change in perceived brightness can be suppressed compared with that in the conventional method, as shown in Fig 9 (d) 32 Mechanism behind and suppression of image quality degradation in the temporal direction The emission intensity for which the pixel emission time changes from short to long between frames and the temporal 1 frame Movement of line of sight Perceived brightness TAPT: 1% Brightness: 25% Perceived dark area TAPT: 25% Brightness: 1% Vertical direction on screen (a) Emission without transition (b) Perceived brightness without transition Perceived brightness TAPT: 1% Brightness: 25% TAPT: 25% Brightness: 1% Transition Suppression of fluctuation TAPT: time aperture Vertical direction on screen (c) Emission with transition (d) Perceived brightness with transition Figure 9: Mechanism behind and suppression of image quality degradation at temporal aperture control boundary 18

7 change in perceived brightness owing to the integration effect in human vision are illustrated in Fig 1, where the tall yellow rectangles indicate the emission intensity adjusted for the temporal aperture and the orange line represents the temporal change in integrated brightness within one frame When there is no transition period for the switch between short and long emission times between frames [Fig 1 (a)], there is a large fluctuation in the integrated brightness within one frame that is perceived as a blinking effect This effect can be reduced by establishing a transition period in which the emission time changes gradually in the time direction, with the center of gravity of the emission at the temporal center of one frame As shown by the orange line in Fig 1 (b), the effect of the transition period is to reduce the fluctuation in the integrated brightness in one frame and to suppress the perceived blinking 33 Subjective evaluation In the previous section, we explained how image quality degradation caused by a change in the temporal aperture can be suppressed by introducing a transition region in the spatial direction and a transition period in the temporal direction In this section, we report the results of subjective experiments conducted to test the effect of this method The conditions for the evaluation experiments are listed in Table 1 The three still images used are a natural scene, Table 1: Conditions of subjective assessment test Viewing environment General office lighting Type: OLED Peak luminance: 2 nits* Display Size: 25-inch diagonal Frame rate: 12 Hz Assessors 1 Viewing distance 3H (3 times the display height) Grading A Blinking or image quality degradation is perceived for method A B Blinking or image quality degradation is perceived for method B X No blinking or image quality degradation is perceived for either method *One nit is the brightness equivalent to 1 cd/m 2 a 7% gray band, and a 7% gray rectangle (respectively images A, B, and C in Fig 11) When video images are used for evaluation, it is difficult to isolate temporal changes and spatial changes, so only still images are used in this evaluation to verify the basic suppression effect in the spatial and temporal dimensions individually The evaluation sequences used to evaluate the effect of introducing the spatial transition region and the temporal transition period are presented in Fig 12 The sequences alternate between method A and method B, in one of which Brightness 1 frame Blink artifact Integration of luminance a Without transition period (without movement of center of gravity) Brightness 1 frame Suppression of change in luminance Integration of luminance 25% temporal aperture 1% temporal aperture Gradual change b With transition period (with movement of center of gravity) Figure 1: Mechanism behind and suppression of image quality degradation when temporal aperture control is switched 19

8 Image A Image B Image C Figure 11: Test images Method A 1 s Method B 1 s Method A 1 s Method B 1 s 15 s (a) Test sequence for spatial transition Method A 4 s Method B 4 s Method A 4 s Method B 4 s (b) Test sequence for temporal transition 15 s Figure 12: Test sequences the temporal aperture is varied by using spatial transition areas or temporal transition periods of various widths, and in the other, a constant temporal aperture is maintained The evaluators respond by indicating in which displayed images they noticed a change such as blinking or image degradation or that they did not notice change in any of the images The results of a subjective evaluation of the change in the spatial transition using images A and B from Fig 11 are presented in Fig 13 Although a gray rectangle image such as image C is often used for evaluation, the tall gray region is expected to enable a more accurate evaluation, considering that the subject of evaluation here is the width of the transition for vertical movement of the line of sight; thus, image B was used The values on the horizontal axis in Fig 13 represent the evaluated widths of the transition region The change in width from line to line was linear For Perception rate (%) Image A Image B Perception rate (%) Image A Image C Transition width (lines) Figure 13: Subjective evaluation results for spatial transition Transition width (frames) Figure 14: Subjective evaluation results for temporal transition 2

9 example, the change in width for each line was 18% for a transition width of 4 lines, which changed the temporal aperture from 25% to 1% (or the reverse) The vertical axis in the figure represents the perception rate as indicated by the responses of the evaluators for each transition width The perception rate is the proportion of responses for which the evaluator could perceive a difference between method A and method B Lower perception rates indicate that changes such as blinking and image degradation were not perceived, so it can be concluded that the hypothesized improvement effect was obtained From the results presented in Fig 13, we can see that the perception rate decreases at a transition width of about 4 lines or more for image A (natural scene) For image B (7% gray), we can see a large decrease in perception rate when the transition rate width is 8 lines or more, but if the number of lines is further increased, there is no proportional decrease in the perception rate We consider the reason for this result to be that the rather poor uniformity of luminance within the OLED display screen used in the evaluations resulted in a change in the uniformity of luminance, causing the perception of a change in the temporal aperture From these results, we conclude that it is possible to suppress image degradation at temporal aperture boundaries by establishing a transition region of 4 to 8 lines The results of a subjective evaluation of change for the temporal transition using images A and C from Fig 11 are presented in Fig 14 The values on the horizontal axis in Fig 14 represent the transition period widths that were evaluated In the same way as for the spatial transition evaluation, the change in width from frame to frame was linear For example, the change in width for one frame was 15%, which changed the temporal aperture from 25% to 1% (or the reverse) over a transition width of 5 frames The evaluation results presented in Fig 14 show that, for both evaluation images, the perception rate decreased for transition widths of 3 frames or more and was less than 4% for a transition width of 5 frames From these results, we conclude that the blinking effect can be suppressed by transition periods of about 5 frames or more (4 s at 12 Hz) 4 Improvement of Device Lifetime We also investigated the effect of adaptive temporal aperture control on OLED lifetime The service life of an OLED device, T, is known to be expressed in terms of initial luminance, L, by 1) 1 T (5) x L where x is an exponent that has a value from 1 to 2 From this relationship, the device lifetime for temporal aperture d, T d, can be expressed as where L represents the instantaneous luminance for the emission interval Accordingly, the relationship between the temporal aperture and the relative lifetime is shown in Fig 15, given that the relative lifetime is 1 when x is 17 and d is 1 We can see that when the temporal aperture is shortened to 25%, for example, the relative lifetime decreases to about 4 Next, we consider dynamic temporal aperture control in which the temporal aperture is set to 25% for the dynamic Relative lifetime Relative lifetime = x L 1 1 x 1 Td d ( < d 1) (6) x d d L Temporal aperture (d) Figure 15: Relationship between temporal aperture and relative lifetime Proportion of dynamic area Figure 16: Relationship between dynamic area proportion and relative lifetime 21

10 image region and to 1% for the static region, and the respective relative lifetime values are taken to be T25 and T1 In this case, the estimated lifetime (T apt ) for when adaptive temporal aperture control is used can be expressed by Equation (7), where α is the proportion of the video image occupied by the dynamic region 1 a 1 = = a T T T apt 25 1 (7) A graph showing the relationship between the relative lifetime and the proportion of the dynamic region with x = 17 is presented in Fig 16 Taking the relative lifetime to be 1 when the dynamic region proportion is (equivalent to d = 1 in Fig 15), we can see that the relative lifetime decreases as the dynamic region proportion increases Detecting motion vectors for the movement of at least four pixels per frame in ordinary TV programs and calculating the proportion of dynamic regions for each program resulted in values of roughly 2% to 4%, so the relative lifetime is expected to range between 6 and 75 when adaptive temporal aperture control is used When the temporal aperture is fixed at 25%, the relative lifetime is 4, as described above, so applying adaptive temporal aperture control can be expected to improve the device lifetime by a factor of about 15 to 18 while maintaining image quality 5 Conclusion We constructed an OLED display panel with adaptive temporal aperture control and used it in evaluation experiments Then, the mechanisms behind the image degradations that occurs at the boundaries of temporal aperture control were revealed In addition to explaining the mechanisms, we presented evaluation results that confirmed its effectiveness in suppressing image degradations Specifically, degradation that occurs when the temporal aperture changes in the spatial direction can be suppressed by establishing a spatial transition region of 8 lines or more, and the blinking phenomenon that occurs when there is a change in the temporal aperture in the time direction can be suppressed by establishing a transition period of 5 frames or more We also considered the effect of adaptive temporal aperture control on OLED lifetime and determined that it is possible to reduce lifetime degradation while maintaining image quality This paper is a revised and corrected version of the following article that appeared in the Journal of the Society for Information Display T Usui, Y Takano and T Yamamoto: A Study on a Driving Method of OLED Displays for a Better Motion Image Quality with Adaptive Temporal Aperture Control, Journal of the SID, Vol 25, No 8, pp (216) References 1) Y Shimodaira, T Hirano and S Fuke: Blur injury caused by motion on the hold type picture display, IECE (Japan, currently IEICE) Trans J68-B, No 12, pp (1985) (in Japanese) 2) H Ishiguro and T Kurita: Consideration on motion picture quality of the hold type display with an octuple-rate CRT, IEICE Tech Rep EID96-4, pp19 26 (1996) (in Japanese) 3) T Kurita, A Saito and I Yuyama: Consideration on Perceived MTF of Hold Type Display for Moving Images, Proc IDW 98, 3D3-4, pp (1998) 4) T Kurita: Moving Picture Quality Improvement for Holdtype AM-LCDs, SID Symposium Digest, 32, pp (21) 5) Y Kuroki and T Nishi: Improvement of Motion Image Quality by High Frame Rate, SID Symposium Digest, 37, pp14-17 (26) 6) T Kurita: A Consideration on Motion Image-quality Improvement of LCD-TVs, SID Symposium Digest, 4, pp (29) 7) T Kurita: A Guideline for Motion Image-quality Improvement of LCDTVs, Proc IMID 9, pp (29) 8) T Kurita: A Consideration on Motion-image-quality Improvement of LCDs and Video Systems, Journal of the SID, Vol18, No12, pp (21) 9) I D Paker, Y Cao and C Y Yang: Lifetime and Degradation Effects in Polymer Light-emitting Diodes, Journal of Applied Physics, Vol85, No4, pp (1999) 1) C Féry, B Racine, D Vaufrey, H Doyeux and S Cinà: Physical Mechanism Responsible for the Stretched Exponential Decay Behavior of Aging Organic Light-emitting Diodes, Applied Physics Letters, Vol87, No21, p21352 (25) 11) T Usui, H Sato, Y Takano, T Yamamoto and K Ishii: A Method of Image Quality Evaluation for Adaptive Temporal Aperture Control with Hold-type Displays, IDW 14, VHF3-2, pp (214) 22

11 12) T Usui, H Sato, Y Takano, T Yamamoto and K Ishii: A Study of Adaptive Temporal Aperture Control for OLED Displays with Motion Vector, SID 215 Digest, pp (215) 13) T Usui, HSato, Y Takano, K Ishii and T Yamamoto: Evaluation System of Adaptive Temporal Aperture Control for OLED Displays, ICCE 216, pp (216) 14) T Usui, HSato, Y Takano, K Ishii and T Yamamoto: A Study of Suppressing Image Quality Degradation Cause bu using Adaptive Temporal Aperture Control for OLED s Displays, ITE Tech Rep IDY216-1, pp73-76(216) (in Japanese) 15) T Usui, Y Takano and T Yamamoto: Development of OLED Display using Adaptive Temporal Aperture Control Driving Method with Transition Area Insertion, IDW 16, DES2-2, pp (216) 23