New standardization activity in display picture quality assessment method

Transcription

1 New standardization activity in display picture quality assessment method Jongseo Lee LCD

2 Contents Ⅰ ISO issues Pixel defect Ⅱ Gamma Value Calculation Ⅲ Image Sticking Measurement Ⅳ Power Consumption Measurement Ⅴ Color reproducibility - Gamut VI Contrast related issues Ⅶ Motion Artifact Measurement

3 ISO vs Pixel Fault Classification ISO ISO ISO

4 Gamma Value Calculation Determination of Gamma Both results are not the same, Non-linear s is more accurate But, This is not representative real gamma characteristic There are more missing information

5 Gamma Value Calculation Determination of Gamma Ex) Gamma curves After gamma curve fitting, both gamma values are suppose to be the same Need a mathematical approach to trace the difference Will it be simple to use? Will it represent gamma characteristics correctly?

6 Image Sticking Measurement Luminance measurement Edge detection A B Total Average Luminance in the box Average Luminance in 2 degree range C D Very small luminance difference but, image sticking level is very high

7 Power Consumption Measurement IEC TC100 & TC110 Joint WG

8 Power Consumption Measurement Imaginary Master Histogram of Post-Gamma APLs Frame Counts ( arbitrary units ) Post-Gamma APL ( % )

9 Power Consumption Measurement # JEITA Proposed TV Set Power Measurement Method w b c t Po = x Pw x Pb x Pc x Pt where: Po is the output power value that is used for the final power calculation Pw is the measured power of the 100% white pattern, Pb is the measured power of the full black pattern, Pc is the measured power of the color bar pattern, Pt is the measured power of the white and black bar pattern.

10 Color Gamut Controversy CIE 1931 xy vs. CIE 1976 u v NTCS xx%, EBU, or etc Gamut value is almighty? The larger is the better? What about human color perception?

11 Color Gamut Controversy 1931 vs What is the big deal? - Who cares? CIE 1931 non-uniform color space CIE 1976 uniform color space MacAdam eclipse

12 Color Gamut Controversy 1931 vs CIE 1931 It is very important! LED(105%) 0.7 CCFL(72%) CIE 1976 LED(121%) CCFL(87%) 고색재현 (90%) PDP(83%) 0.6 고색재현 (101%) PDP(83.5%) 0.6 DLP(86.6%) 0.5 DLP(94.6%) 0.5 y 이상적 color(113%) 0.4 v' 이상적 color(131%) x u' Device Color Gamut In 1931 In 1976 Difference LED 105 % 121 % 16 % LCD CCFL 72 % 87 % 15 % Wide Gamut 90 % 101 % 11 % PDP 83 % 83.5 % 0.5 % DLP 86.6 % 94.6 % 8 %

13 Color Gamut Controversy srgb 3 Big Reference Color Gamut - NTSC, Adobe RGB, srgb(bt.709) - Even Big 2 (NTSC, Adobe RGB) do not cover the srgb(bt.709) which is used in real application 0.9 CIE 1931 NTSC(100%) 0.7 CIE 1976 NTSC(100%) 0.8 Adobe RGB(95%) 0.6 Adobe RGB(99.4%) 0.7 srgb(70.8%) 0.5 srgb(87.2%) y v' x u' Is it reasonable? Uncovered Area but larger!

14 Color Gamut Controversy Reference CS NTSC 대비 Color & Gamut - (NTSC) Color & Gamut - 24 인치 LED BL(sRGB 기준한색재현력 ) Color & Gamut - 24 인치 LED BL(Color Locus 전체면적기준 ) C.I.E 1931 C.I.E 1976 C.I.E 1931 C.I.E 1976 C.I.E 1931 C.I.E 1976 x y u' v' x y u' v' x y u' v' White White White Red Red Red Green Green Green Blue Blue Blue Black CCT CCT 6756 CCT 6756 Gamut 100.0% 100.0% Gamut 148.8% 149.6% Gamut?? 49.7% NTSC(100%) NTSC(100%) Color & Gamut - Adobe RGB(1998) Color & Gamut - 24 인치 LED BL(Adobe98 기준한색재현력 ) Color & Gamut - srgb (Color Locus 전체면적기준 ) C.I.E 1931 C.I.E 1976 C.I.E 1931 C.I.E 1976 C.I.E 1931 C.I.E 1976 x y u' v' x y u' v' x y u' v' White White White Red Red Red Green Green Green Blue Blue Blue CCT CCT 6756 CCT Gamut 95.4% 99.4% Gamut 110.3% 131.2% Gamut?? 33.2% Adobe RGB(95%) Adobe RGB(99.4%) Color & Gamut - 24인치 LED BL(NTSC 기준대비 ) Color & Gamut - NTSC (Color Locus 전체면적기준 ) Color & Gamut - (srgb = BT. 709) C.I.E 1931 C.I.E 1976 C.I.E 1931 C.I.E 1976 C.I.E 1931 C.I.E 1976 x y u' v' x y u' v' x y u' v' White White White Red Red Red Green Green Green Blue Blue Blue CCT 6756 CCT CCT Gamut 105.4% 130.5% Gamut?? 38.1% Gamut 70.8% 87.2% 24 LED (105.6%) 24 LED (130.8%) srgb(70.8%) srgb(87.2%) Color & Gamut - (EBU) C.I.E 1931 C.I.E 1976 x y u' v' White Red Green Blue CCT Gamut 71.7% 88.1% srgb(70.8%) srgb(87.2%) Color & Gamut - (SMPTE-C) C.I.E 1931 C.I.E 1976 x y u' v' White Red Green Blue CCT Gamut 65.6% 77.0% srgb(70.8%) srgb(87.2%)

15 Standard Color coordinates

16 Wide color gamut - xvycc

17 Wide color gamut - xvycc

18 Wide color gamut - xvycc

19 Contrast related issues Viewing Angle in LCD Contrast Ratio vs. Viewing Angle Contrast Ratio vs. Viewing Angle Contrast Ratio vs. Viewing Angle Horizontal Vertical 45degree 135degree Horizontal Vertical 45degree 135degree Horizontal Vertical 45degree 135degree Contrast Ra 100 Contrast Ra 100 Contrast Ra Angle 1 Angle 1 Angle What about off-angle picture quality? What else to judge viewing angle?

20 Contrast related issues Ambient (outdoor) CR, Peak CR

21 Motion Artifacts - pursuit camera system Edge Blurring Color Distortion Wire Frame Flickering False Contour

22 Response Time = 0 msec Display and Hold type perception mechanism Impulsive type perception mechanism Retina Retina

23 Abnormal transition M PRC: _V D_LED scanning.csv Human Perception M PRC: _ImpulsiveOCB_q32inch_(Gray 1_1)_9step_1.csv Luminance (Cd m-2, relative) M PRC: _400W2-L01_DCC1_7Q5F04400C_9step_1.csv Luminance (Cd m-2, relative) Measured Data LCD pixel Luminance (Cd m-2, relative) LCD pixel LCD pixel

24 Contrast Sensitivity Function externa l noise photon noise Φ ph neural noise Φ 0 Image Optical MTF M opt (u) Lateral inhibition M lat (u) Integratio n over X,Y,T Decision m signal >km n CSF model outline Mach band

25 Motion Blurring Evaluation - PBET Contrast Sensitivity Function & P-BET (Perceived Blurred Edge Time) Lum. profile CSF P-BET 상대휘도 Distance Fourier Transform Spatial Freq. Spatial Freq. Inverse Fourier Transform time Definition of P-BET

26 More than Motion Blurring IEC : Motion Artifact Measurement - 1 st CD version.

27 We reviewed Ⅰ Pixel defect - ISO Ⅱ Gamma Value Determination Ⅲ Image Sticking - Edge detection Ⅳ Power Consumption Measurement Ⅴ Gamut Controversy VI Ⅶ Viewing Angle, Ambient CR Motion Artifact Measurement - Blurring, CBU, False contour

28 Conclusion Reasonable/Fair Standard Scientific/Engineering background Standard Consider correlation with Human perception We don t want just numbers analyticible Standard More Speed Standard! Our Heart goes out to the late Mr. Nakamura Jumpei!