Chapter 3 Color Optimization for HDR-LCD
3.4 Summary
3.4 Summary
The color model for optimized HDR image was proposed and demonstrated on a 37’’
HDR-LCD. Considering the intensities and light leakages of dimming backlight, the model predicted accurate CIE tristimulus with ∆E00 smaller than 1.4. Using the color optimization method, the color difference of the optimized HDR image achieved a mean ∆E00 < 2.5 (using intensity model, ∆E00 > 11). Moreover, image detail was maintained. Consequently, by using the proposed color model, especially for the color-LED backlight system, the HDR image yielded high color accuracy and maintained clear image detail. The results have been successfully demonstrated on a 37” HDR-LCD for the high quality applications.
(a)
(b) (c) Fig. 3-11 Image details in the: (a) Target, (b) HDR image of intensity model, (c) HDR image of color model
41
Chapter 4
Color Optimization and Power Reduction for FSC-LCD
The color optimization model is applied to achieve high color accuracy. The original RGBWmin and RGBD methods suffer from color distortion and power consumption. By using the conventional FSC-LCD modeling, the optimized RGBWmin and RGBD methods will redistribute liquid crystal signals to maintain image quality. Moreover, the RGBWw and RGBDw method will be proposed using the FSC-LCD modeling of a RGBW LED backlight.
4.1 FSC-LCD Modeling of a RGB Scanning Backlight
The FSC-LCD performs temporal color mixing synthesis to yield a full color image without color filters. By evaluating fundamental differences between HDR-LCD and FSC-LCD, FSC-LCD modeling with RGB LED was simplified using the HDR-LCD model.
The modeling consisted of two stages. The first stage was three non-linear LUTs in RGB channel, as shown in Eq. (3-1). Three transfer matrices (MR’
MG’
MB’
) were described for the RGB LED backlight respectively, as shown in Eq. (4-1).
,max ,min
42
The predicted CIE tristimulus of yielded FSC image were the summation of XYZ values from the RGB scanning backlight, as shown in Eq. (4-2). The color black levels and backlight distribution also increased the color distortion of FSC image.
4.1.1 Optimized RGBWmin Method
In the RGBWmin method, one frame is divided into four fields, R, G, B, and white (W).
The optimized RGBWmin algorithm is illustrated in Fig. 4-1. The LED backlight performs full-on R, G, B, and W backlight sequentially with corresponding LCR, LCG, LCB, and LCW signals. In the W field, RGB LEDs turn on simultaneously. Each pixel’s gray level in the W field depends on the minimum gray level of the corresponding pixel in the original input signals (dr, dg, and db). The input image was transferred into the target tristimulus by using the FSC-LCD model. While assessing the white field luminance and the target tristimulus into the FSC model, the LC signals for R, G, and B field were redistributed. Thus, the optimized RGBWmin method yielded color accuracy on FSC-LCD.
' ' '
Fig.4-1 Optimized RGBWmin method
43
4.1.2 Optimized RGBD Method
The optimized RGBD method displayed R, G, B, and D fields with corresponding global (0D) dimming BL and LC signals. Thus, four sets of BL and LC signals on the RGBD color sequence were determined by image content, as shown in Fig. 4-2. In the D-field, gray levels of three primary color backlights were set as BLR, BLG, and BLB, respectively. The LC signals (LCD) of the D field were defined by using the minimum signals from the input image (dr, dg, and db). In optimized RGBD method, the input signals were evaluated as the target tristimulus by using the FSC-LCD modeling and the mapping function. The optimized RGBD algorithm redistributes the LC signals to the R, G, and B fields to create a colorful image and achieve color accuracy.
44
4.2 FSC-LCD Modeling of a RGBW Scanning Backlight
The FSC-LCD color model with a RGBW scanning backlight is comprised of two stages.
In the first stage, four non-linear LUTs were developed according to individual RGBW LEDs, as illustrated in Eq. (3-1) and (4-3). The LUT for W LED was added into the first stage, as described in Eq. (4-3). The second stage evaluated light leakage and maximum output from RGBW backlight. Therefore, four linear transfer matrices (MR’, MG’, MB’, and MW’) were implemented, as shown in Eq. (4-1) and (4-4), and MW’ function was the transfer function for W LED backlight. With backlight and liquid crystal signals, the color appearance on a only powerful W LEDs provided luminance for reducing power dissipation. As the RGBWw method shows, in Fig. 4-3, the LC signals (LCW) of Ww field were calculated according to
45 field) were evaluated according to the image content. In the RGBDw method, as shown in Fig.
4-4, the luminance was provided by W LEDs and the other RGB LEDs supplied the color. perception according to the target tristimulus to maintain image quality.
Input Image
46
Using the color optimization model, four FSC methods, optimized RGBW, optimized RGBD, RGBWw, and RGBDw, were proposed to maintain image quality. The color accuracy, CBU phenomena, and power reduction between four methods will be explained on a 15.4”
FSC-LCD later.
4.3 Experimental Setup and Results
The experiment results of optimized RGBW/RGBD and proposed RGBWw/RGBDw method are discussed in this section.
4.3.1 Hardware Structure
The optimized RGBWmin, RGBD, RGBWw, and RGBDw methods were demonstrated on a 15.4” FSC-LCD supported by Chunghwa Picture Tubes, LTD, as shown in Fig. 4-5 (a).
The FSC-LCD system was comprised of a 1280x800 TN LC panel and a side-emitting RGB backlight, as shown in Fig. 4-5 (b). The backlight includes 36 RGB LED chips on each side.
The FSC-LCD field rate was 240 Hz and the backlight flashing time was 1.39ms.
Fig.4-4 RGBDw method
47
4.3.2 IEC Picture Library
For evaluating the power consumption of displays, the IEC published International Standards for audio, video, and related equipment [55]. The first edition of the International Standard IEC 62087 was prepared by the IEC technical committee100 to measure power consumption in 2002. The second edition of IEC 62087 specified methods for measurement of the power consumption in television sets, video recording equipment, Set Top Boxes (STBs), audio equipment and multi-function equipment for consumer use in 2008. This second edition canceled and replaced the first edition, published in 2002 and constituted a technical revision.
The concept of IEC 62087 was to develop a new measurement method, APLs (Average Picture Levels). Five videos were collected from USA, UK, Australia, Netherlands, and Japan to produce a standard video that represented TV use, as described in Fig.4-6.
LCD Panel Prism Sheet Diffuser *2
RGB LED LGP
Micro Structure
(a)
(b)
Fig. 4-5 (a) Platform of the 15.4” FSC-LCD, and (b) Backlight structure
48
The objective of survey was to gain a better understanding of the nature of the APLs inherent in standard video. The standard was displayed on a Television under normal use conditions. The methodology adopted was to sample 40 hours of Prime-time viewing to ensure a mix of genre was sampled so as not to bias the results toward any particular content.
This method collected results that do indeed represent average viewing habits. The method was considered by the Working Group to provide data of sufficient accuracy so as to be able to produce the new video clip.
The average of these APL curves has been used to produce a 10 minute natural moving image clip which, in conjunction with the revised TV testing method in IEC 62087, produced a more accurate measurement of the TV sets’ power consumption. This had implicated regulators and energy consumption planners. Highly reliable energy consumption models were developed based on measurements of TV power use using this new method. The comparison showed a strong correlation between the 10 minute clip and the original 40 hrs of collected material.
F re q u en cy
APL Percentage
APL Histograms by County
Fig.4-6 APLs sampled from around the world
49
Furthermore, the 10 minute IEC 62087 natural moving image was captured in stationary images. Several test images fit the average of APL curve. These static images established an IEC picture library which highly correlated to the IEC standard video. The IEC picture library was comprised of 25 static images, as illustrated in Fig. 4-7. The pictures were numbered from darkest (number 1) to brightest (number 25), as illustrated in Fig. 4-8. The IEC picture library was instrumental in examining power consumption between algorithms. Therefore, the following results were estimated based on this picture library.
.
F r e q u e n c y
Dark Bright Luminance
Fig.4-7 IEC picture library
50
4.3.3 Color Accuracy of Static Images
The IEC picture library was prepared to examine the color accuracy of CBU suppression methods, such as optimized RGBWmin, RGBD, RGBWw, and RGBDw. The target tristimuls were determined for each method individually on a 15.4’’ FSC-LCD. The predicted tristimulus of new arranged four fields were calculated according to the color model. The color differences ∆E00 were estimated between the target tristimulus and predicted tristimulus.
1 2 3 4 5
6 7 8 9 10
11 12 13 14 15
16 17 18 19 20
21 22 23 24 25
Fig.4-8 IEC picture number
51
The average ∆E00 in each test image for FSC methods is illustrated in Fig. 4-9. The average color accuracy of FSC methods in the IEC picture library is compared in Fig. 4-10.
The color difference was acceptable when the mean ∆E00 was smaller than 3 [49]. The color accuracy was ∆E00 0.3~0.6 in these FSC methods. Therefore, the optimized RGBWmin, RGBD, RGBWw, and RGBDw method yielded full color images which maintained accurate color reproduction.
0 5 10 15 20 25
0 1 2 3
C o lo r A c c u r a c y ( a v g . E
00)
IEC Picture Library
RGBWmin RGBD RGBWw RGBDw
Acceptable
Fig. 4-9 Color accuracy in IEC static images
52
4.3.4 Relative CBU Suppression
The CBU index was defined using CIEDE2000 color-difference values (∆E00). The ∆E00
was calculated between the original image and the CBU image. The average color difference was obtained to compare CBU suppression in each image, as shown in Fig. 4-11. When the image luminance increased, the CBU increased. Moreover, the relative CBU ratio, as described in Eq. 4-6, was determined to the color difference of the four FSC methods, discussed in this thesis, to the original RGB-driving method [23]. According to Fig. 4-12, the average CBU index in the RGBDw method was reduced to 3.1∆E00 for the IEC picture library.
Furthermore, the RGBDw method suppressed CBU about 77% when related to RGB-driving (the relative CBU ratio~23%).
Fig. 4-10 Color accuracy for FSC methods
RGBWmin RGBD RGBWw RGBDw 0
1 2 3
C o lo r A c c u r a c y ( a v g . E 00 )
FSC Method
Acceptable
0.3
0.6 0.5 0.6
53
0000
% FSC Methods 100%
RGB driving
E Relative CBU Ratio
E
(4-6)
Fig.4-11 CBU index for IEC picture library
Fig.4-12 Average CBU index and relative CBU ratio for FSC methods
54
The Airplane image, as shown in Fig. 4-13 (a), was simulated by RGB-driving and RGBDw method, as shown in Fig. 4-13 (b) and (c). The typical CBU and Dw-field modified images are illustrated in Fig. 4-14. In the marked region, it is obvious that the CBU artifact is reduced when compared to the conventional RGB 3-field sequence.
(a)
R-field G-field B-field (b)
Dw-field R-field G-field B-field (c)
Fig. 4-13 (a) Target image Airplane, (b) Conventional RGB sequence, and (c) RGBDw method
55
4.3.5 Power Consumption
The RGBWw and RGBDw with RGBW 4-in-1 LED backlight resulted in low power consumption. The power dissipation using the difference algorithm is estimated in Fig 4-15.
The IEC picture library represents the normal users’ images. The average power is calculated comparing the FSC methods, as shown in Fig. 4-16. The relative power ratio is defined as the power of the FSC methods to the original RGB-driving method in Eq. 4-7. The RGBWw and RGBDw method reduced power consumption when related to the RGB method.
%
FSC Methods100%
Fig.4-15 Power consumption for IEC picture library
(4-7) (a) (b)
Fig.4-14 CBU images in (a) RGB sequence and (b) RGBDw method
0 5 10 15 20 25
56
4.4 Summary
The color optimization methods, optimized RGBWmin, optimized RGBD, RGBWw, and RGBDw, for perceived FSC-LCD images were proposed and demonstrated on a 15.4” color filter-less LCD. Two models predicted CIE tristimulus of a yielded image using a RGB LED backlight and RGBW LED backlight. Using the RGB LED backlight system, the optimized RGBWmin and RGBD method redistributed the RGB fields to display a colorful FSC image.
In the RGBWw and RGBDw method, the RGBW LED backlight used high powered W LEDs to provide greater luminance. The four color fields for RGBWw and RGBDw methods were arranged to maintain color accuracy.
Moreover, the IEC picture library was established to verify color accuracy, CBU suppression, and power consumption. The optimized RGBWmin, RGBD, RGBWw, and RGBDw method achieved accurate color reproduction (∆E00 0.3~0.6). These methods suppressed CBU phenomena by 59%~77% compared to the RGB-driving sequence. The power dissipation of RGBWw and RGBDw was reduced to 82% and 77% of the conventional RGB sequence respectively.
Fig.4-16 Average power consumption and relative power ratio for FSC methods
57
Consequently, by using the proposed FSC methods, especially RGBDw method, the perceived image on FSC-LCD yielded higher color accuracy, suppressed CBU, and reduced power consumption. The methods were successfully demonstrated on a 15.4” FSC-LCD for the high quality applications.
58
Chapter 5
Conclusion and Future work
5.1 Conclusion
The color model for the optimizing images had been proposed and demonstrated on a 37’’
HDR-LCD. By considering the intensities and light leakages of dimming backlight, the model accurately evaluates the CIE tristimulus of the HDR images. Using the color optimization method, the color difference of the optimized HDR images was achieved to the average ∆E00
< 2.5 (using intensity model, ∆E00 > 11). Moreover the image details were maintained as well.
On the other hand, the FSC methods, optimized RGBWmin and RGBD, were proposed using the color optimization model on a 15.4” FSC-LCD with a RGB-LED scanning backlight system. The observed images performed high color accuracy (∆E00 0.3~0.6) and suppressed CBU by 59~68% comparing to conventional RGB-driving sequence. However, the power consumption increased by 125~136% because of the additional D-field (or W-field).
Thus, the modified color optimization model of a RGBW-LED backlight was proposed.
Novel FSC methods, RGBWw and RGBDw, were improved in RGBW-LED backlight to maintain image quality. These methods yielded high color accuracy (∆E00 0.5~0.6) and suppressed CBU by 59%~77%. The power dissipation of RGBWw and RGBDw was reduced to 82% and 77% of the conventional RGB sequence respectively. Consequently, the RGBDw method outperformed than other FSC methods (optimized RGBWmin, RGBD, RGBWw) in color accuracy, CBU suppression and power reduction.
Using proposed color model for the color-LED backlight system, the perceived image on HDR-LCD and FSC-LCD yielded high image quality and suppressed power dissipation.
59
5.2 Future Work
The RGBDw method suppressed great CBU phenomena and reduced power consumption with global dimming technology on a laptop-size LCD. However, for a TV-size FSC-LCD, human eyes are sensitive to CBU especially for vivid color segments. The Child image, as shown in Fig. 5-1(a), is comprised of yellow, cyan, magenta, and skin color segments. The global BL/LC adaption suppressed insufficient CBU, as shown in Fig. 5-1(b). Thus, the local dimming technology will be implemented into a RGBW LED backlight.
By using the local BL/LC color adaption, the optimized RGBDw method sequentially performs R, G, B, and Dw field, as shown in Fig. 5-1(c). The Dw field provides high color saturation according to original image contents. Comparing to the RGBDw method in global color control, the optimized RGBDw method will suppress CBU and reduce power dissipation effectively in large-sized LCDs.
Fig.5-1 (a) RGB sequence, (b) RGBDw method, and (c) Optimized RGBDw method
60
[2] H. Seetzen, W. Heidrich, W. Stuerzlinger, G. Ward, L. Whitehead M. Trentacoste, A.
Ghosh, and A. Vorozcovs, “High Dynamic Range Display Systems,” SIGGRAPH 2004, ACM Transactions on Graphics, vol. 23(3), pp. 760-768 (2004).
[3] T. Shirai, S. Shimizukawa, T. Shiga, S. Mikoshiba, and K. Käläntär, “RGB-LED Backlights for LCD-TVs with 0D, 1D, and 2D Adaptive Dimming,” SID Symposium Digest Tech Papers 37, pp. 1520-1523 (2006).
[4] R.G. Stewart and W.R. Roach, “Field Sequential Display System Utilizing a Backlit LCD Pixel Array and Method for Forming an Image,” US Patent 5337068 (1994).
[5] T. Uchida, K. Saitoh, T. Miyasita, and M. Suzuki, “Field Sequential Full Color LCD without Color Filter for AM-LCD”, Proc. 17th IDRC, pp. 37 (1997).
[6] N. Koma, T. Miyashita, T. Uchida, and N. Mitani, “Color Field Sequential LCD Using an OCB-TFT-LCD,” SID Symposium Digest Tech Papers 31, pp. 632-635 (2000).
[7] H. Seetzen, Lorne A. Whitehead, and Greg Ward, “A High Dynamic Range Display Using Low and High Resolution Modulators,” SID03 Digest, pp. 1450-1453, 2003.
[8] E. Y. Oh, S. H. Baik, M. H. Sohn, K. D. Kim, H. J. Hong, J. Y. Bang, K. J. Kwon, M. H.
Kim, H. Jang, J. K. Yoon, and I. J. Chung, “IPS-mode dynamic LCD-TV realization with low black luminance and high contrast by adaptive dynamic image control technology,” J Soc Info Display, vol. 13, pp. 215-219, 2005.
[9] Y. W. Wang, Y. K. Cheng, H. P. D. Shieh, T. M. Wang, and H. W. Richard Lin,
“Analyses of Point Spread Function in High Dynamic Range Display System,” Optics and
61
Photonics Taiwan 2005, No. G-SA-X 4-3, 2005
[10] H. Yamakita, M. Sakai, Y. Taniguchi, J. Asatama, and K. Adachi, “Field-Sequential Color LCD Driven by Optimized Method for Color Breakup Reduction,” IDW/AD’05, pp.
83-86, (2005).
[11] T. Ishinable, T. Miyashita, and T. Uchida, “High Performance OCB-mode for Field Sequential Color LCDs,” SID Symposium Digest Tech Papers 38, pp. 987–990 (2007).
[12] K. Käläntär, T. Kishimoto, K. Sekiya, T. Miyashita, and T. Uchida,“Spatio-Temporal Scanning Backlight for Color-Field Sequential Optically Compensated Bend
Liquid-Crystal Display,” SID Symposium Digest Tech Papers 36, pp. 1316–1319 (2005).
[13] Y.-F. Chen, C.-C. Chen, and K.-H. Chen, “Mixed Color Sequential Technique for Reducing Color Breakup and Motion Blur Effects,” IEEE/OSA Journal of Display Technology, Vol.3(4), pp. 377-385 (2007).
[14] C.-C. Chen, Y.-F. Chen, T.-T. Liu, C.-H. Chen, M.-T. Ho, K.-H. Chen, and H.-P. D.
Shieh, “Spatial-temporal Division in Field Sequential Color Technique for Color Filterless LCD,” SID Symposium Digest Tech Papers 38, pp. 1806-1809 (2007).
[15] F. C. Lin, et al, “Dynamic Backlight Gamma on High Dynamic Range LCD TVs”, JOURNAL OF DISPLAY TECHNOLOGY, pp. 139-146 (2008)
[16] L. Kerofsky, et al., “Brightness Preservation for LCD backlight dimming,” J Soc Info Display, pp 1111-1118 (2006)
[17] L. Y. Liao, et al., “Blur-Mask Approach for Real-Time Calculation of Light Spreading Function (LSF) on Spatial Modulated High Dynamic Range LCDs,” (submitting to JDT 2009)
[18] G. Z. Wang, et al., “Delta-Color Adjustment (DCA) for Spatial Modulated Color Backlight Algorithm on High Dynamic Range LCD TVs,’’ (submitting to JDT 2009) [19] G. Z. Wang, Y.P. Huang, and H. P. D. Shieh, ”Segment Color Control (SCC) Method for
62
Color Controlled Backlight of High Dynamic Range LCD-TVs,” SID Symposium Digest Tech (2009)
[20] S. Mateeff, “Saccadic Eye Movements and Localization of Visual Stimuli,” Percept Psychophys 24, pp. 215–224 (1978).
[21] W. Hershberger, “Saccadic Eye Movements and the Perception of Visual Direction,”
Percept Psychophys 41, pp. 35–44 (1987).
[22] H. Honda, “Perceptual Localization of Visual Stimuli Flashed during Saccades,” Percept Psychophys 45, No. 2, pp. 162–174 (1989).
[23] F. C. Lin, et al., “Color Breakup Suppression and Low Power Consumption by Stencil-FSC Method in Field- Sequential LCDs,” J. Soc. Info. Display, pp. 221-228, (2009)
[24] C.-L. Wu, C.-H. Chen, F.-C. Lin, Y.-P. Huang, and H.-P. D. Shieh, “A 5.6-inch Field Sequential Color LCD with Less Color Break-up,” Optics and Photonics Taiwan, GO-004, (2007)
[25] C.-L. Wu, C.-H. Chen, F.-C. Lin, Y.-P. Huang, and H.-P. D. Shieh, “Demonstration of a Mobile-Sized Field Sequential Color LCD for Color Break-up Suppression,”Taiwan Display Conference Proceedings, pp.419-422, (2008).
[26] Yi-Pai Huang, Ke-Horng Chen, Chun-Ho Chen, Fang-Cheng Lin, and Han-Ping D.
Shieh, “Adaptive LC/BL Feedback Control in Field Sequential Color LCD Technique for Color Breakup Minimization,” JOURNAL OF DISPLAY TECHNOLOGY, VOL. 4, NO. 3, SEPTEMBER, pp.290-295 (2008)
[27] Feng Li, “ Deriving LED Driving Signal for Area-Adaptive LED Backlight in High Dynamic Range LCD Displays”, SID07 Digest, pp.1794-1797 (2007).
[28] Mark D. Fairchild, Color Appearance Models, 2nd Edition, Wiley, New York, pp. 310-315 (2005).
[29]M. R. Luo, G. Cui, B. Rigg, “The Development of CIE 2000 Colour-Difference Formula:
CIEDE2000,” Color Research & Application, pp. 340-350 (2000).
[30] Yi-Ling Chen, Yu-Kuo Cheng, Yi-Pai Huang, Han-Ping D. Shieh, and Szu-Che Yeh,
63
“Color Optimization Model for High Dynamic Range LCD with Color-Controlled Backlight System,” Society For Information Display (2009).
[31] Yi-Ling Chen, Yu-Kuo Cheng, Yi-Pai Huang, Han-Ping D. Shieh, and Szu-Che Yeh,
“Color Optimization Model for Color-Controlled Backlight System in High Dynamic Range LCD-TVs,” Internal Displacement Monitoring Centre (2009).
[32] Yi-Ling Chen, Yu-Kuo Cheng, Yi-Pai Huang, Han-Ping D. Shieh, and Szu-Che Yeh,
“Applying Colorimetric Color Reproduction in High Dynamic Range Liquid Crystal Display System,” Optics and Photonics in Taiwan (2008).
[33] R. W. Baloh, A. W. Sills, W. E. Kumely, and V. Honrubia, “Quantitative Measurement of Saccade Amplitude, Duration, and Velocity”, NEUROLOGY, pp.1065-1070 (1975).
[34] D. C. Burr, M. C. Morrone, and J. Ross, “Selective Suppression of the Magnocellular Visual Pathway during Saccadic Eye Movements,” Nature 371, No. 6497, pp. 511–513 (1994).
[35] K. Uchikawa, and M. Sato, “Saccadic Suppression of Achromatic and Chromatic
responses Measured by Increment-threshold Spectral Sensitivity,” J. Opt. Soc. Am. A Opt.
Image Sci. Vis. 12, No. 4, pp. 661–666 (1995).
[36] J. Ross, M. C. Morrone, and D. C. Burr, “Compression of Visual Space before Saccades,” Nature 386, pp. 598–601 (1997).
[37] J. Watanabe, “Perisaccadic Perception of Continuous Flickers,” Vision Res. 45, pp.
413–430 (2005).
[38] T. Jarvenpaa, “Measuring Color Breakup of Stationary Image in Field-Sequential-Color,”
SID Symposium Digest Tech Papers 35, pp. 82-85 (2004)
[39] A. T. Bahill, M. R. Clark, and L. Stark, “ The Main Sequence, a Tool for Studying Human Eye Movements,” Math. Biosci. vol. 24, pp. 191-204 (1975).
[40] T. Jarvenpaa, “Measuring Color Breakup of Stationary Image in Field-Sequential-Color,”
64
J. of Info. Display, Vol. 13, pp. 139-144 (2005).
[41] T. Kurita and T. Kondo, “Evaluation and Improvement of Picture Quality for Moving
[41] T. Kurita and T. Kondo, “Evaluation and Improvement of Picture Quality for Moving