FPD Standard Activities in Taiwan: Measurements for mura and motion blur

Size: px

Start display at page:

Download "FPD Standard Activities in Taiwan: Measurements for mura and motion blur"

Whitney Brooks
6 years ago
Views:

1 IMID/IDMC 06 The 6 th International Meeting on Information Display & International Display Manufacturing Conference (IMID/IDMC 2006) August 22-25, 2006,Exhibition & Convention Center (EXCO), Daegu, Korea FPD Standard Activities in Taiwan: Measurements for mura and motion blur Taiwan TFT LCD Association Industrial Technology Research Institute Hsinchu, Taiwan Chao-hua Wen August 24, 2006

2 Outline Display Standard Organizations in Taiwan Mura Measurement State of The Art Subjective Test Multi-View Mura Inspection System Motion Blur Measurement Standard Activity in Taiwan Comparisons between TTLA and Ostuka MPRT-1000 Color Shift Analysis Conclusion 2

3 Outlook for LCD Equipment Market Global Market Trend of LCD Equipment Manufactures Array Process Equipment Cell/Module Process Equipment Inspection/Repair Equipment Other 100 Million (USD) Year Data source: 2005 LCD Equipment Data Book, Electronic Journal 3

4 Outlook for Inspection and Repair Equipments 100 Million (USD) Global Market Trend of Inspection and Repair Equipments Repair Equipment Environment Test Equipment Prober Equipment Color Filter Inspection Equipment Array Inspection Equipment Substrate Inspection Year Data source: 2005 LCD Equipment Data Book, Electronic Journal 4

Industry Promotion Group Color Imaging Industry Promotion Office China Taiwan

5 Display Standard Organizations in Taiwan Ministry of Economic Affairs Bureau of Standards, Metrology and Inspection Industrial Development Bureau Digital Video Industry Promotion Group Color Imaging Industry Promotion Office China Taiwan Digital-TV Industry Alliance Sinocon Fundation 8 industry members 台灣平面顯示器材料與元件產業協會 62 industry members 5

TTLA TDMDA Secretary ITRI TDSC Execution Committee Chief Chairman: President of TTLA SEMI Deputy Chairman: President of TDMDA & Director of CIPO Motion Artifacts Committee Mura Measurement Committee

6 TTLA TDMDA Secretary ITRI TDSC Execution Committee Chief Chairman: President of TTLA SEMI Deputy Chairman: President of TDMDA & Director of CIPO Motion Artifacts Committee Mura Measurement Committee Dr. K.H. Yang VP, Hannstar Dr. C.D. Pan Executive Director, Forhouse Charles Yu CEO, Allied Material Tech. Corp. SEMI Taiwan Dr. D.W. Lai President, Optimax Environment Health Safety (EHS) Committee Backlight Unit Committee Color Filter Committee Polarizing Film Committee 6

Modules (1CD, 2006) IEC 61747IEC 61747-6-2: Liquid crystal display devices- Part 6-2: Measuring methods for liquid crystal display modules Reflective type (partly reworked) IEC 61747-6-3: Liquid

7 Related Standard Activity International Electrotechnical Commission Technical Committee No. 110: Flat Panel Display Devices Working Group 2: Liquid Crystal Display Devices IEC : Liquid crystal display devices - Part 5-2: Visual Inspection of Active Matrix Liquid Crystal Display Modules (1CD, 2006) IEC 61747IEC : Liquid crystal display devices- Part 6-2: Measuring methods for liquid crystal display modules Reflective type (partly reworked) IEC : Liquid crystal display devices - Part 6-3: Motion Artifact Measurement of Active Matrix Liquid Crystal Display Modules (1CD, 2006) Video Electronics Standards Association Flat Panel Display Measurements Standard Flat Panel Display Measurements (FPDM2) Update (May, 2005) SEMI International Standards SEMI D : Definition of Measurement Index (SEMU) for Luminance Mura in FPD Image Quality Inspection (Sept. 09, 2002) SEMI D : Measurement Method of SEMI Mura in FPD Image Quality Inspection (Jan. 28, 2005) Standard Panels Working Group SPWG Notebook Panel Specification Version 3.5 (March 29, 2005) 7

8 What is Mura? Rubbing Block V-Band Spacer Mura S-line G-line H-line Tr Gap Backlight Mura Spot Mura Around Mura Color Mura 8

9 Typical Test Conditions Captured under Black Captured under Middle-Gray Captured under White 9

10 Current Works in Taiwan Ergonomic Study To examine SEMI D : Definition of Measurement Index (SEMU) for Luminance Mura in FPD Image Quality Inspection. To examine SEMI D : Measurement Method of SEMI Mura in FPD Image Quality Inspection. To consider more critical factors, such as the different background levels, shapes, etc. Standardized Equipments, Measuring Procedures and Image Processing To develop a prototype of Multiple View Mura Inspection To establish guidelines for visual inspection 10

11 SEMI D Definition of Measurement Index (SEMU) for luminance Mura in FPD Image Quality Inspection C jnd = F( S jnd ) = 1.97/ S 0.33 jnd C S jnd jnd :Contrast of (Uint :% relativeto bracground= 100%) : Area of 2 (Uints:mm ) Semu= C C S X X = = C C X X X : Average (Uint : Surface (Uints mura at JND mura at abovecontrast / C jnd / F( S X ) /(1.97/ S area : mm contrast : %relative 2 ) of 0.33 X of to mura ) mura bracground being being = 100%) measured measured Luminance Mura is related with the size of Mura! 11

12 SEMI D : Measurement Method of SEMI Mura in FPD Image Quality Inspection Light measurement device (LMD) Recommendation of CCD equipments to detect mura Dynamic range: Due to semu should control error within 0.1semu,CCD sensor should be over 12 bits. Luminance accuracy: The accuracy should be ± 0.01 cd/m 2 at 100 cd/m 2. Measuring points: The measuring points should be over 200(H) 200(V) points. The test pattern should be set by 50 cd/m 2 in the white pattern. The permissible error is 10 cd/m 2. The viewing direction should be set at 90º ± 1 º.??? Because of using a ccd camera, the measuring distance is set by the same visual angle in all DUT. This document standardized the 15 display size and the distance is 500mm. Therefore, there will be different distances for each panel size, e.g.,??? The document uses the Cx value to determinate thresholds. In this standard, we can consider mura area when Cx value is over

13 Questions are Luminance is not only unique factor during human visual detection, but also involve Shapes of Mura Orientations of Mura Varied Backgrounds Positions of Mura occurrence Viewing angle, etc. 13

Methodology of Ergonomic Study Experimental Design Two-Alternative Forced-Choice (2AFC) Paradigm Dependent variable Probability of correct Independent variables Background L64 (1.36 cd/m2), L368 (23.

14 Methodology of Ergonomic Study Experimental Design Two-Alternative Forced-Choice (2AFC) Paradigm Dependent variable Probability of correct Independent variables Background L64 (1.36 cd/m2), L368 (23.97 cd/m2), L512 (51.66 cd/m2), L900 ( cd/m2) Area and shape (pseudo Mura) Line: 7 2 pixels (3.96 mm2); pixels (68.56 mm2); 640 2; 1280 pixels ( mm2) Circle (Diameter, mm): 7 pixel (10.76 mm2); 25 pixel ( mm2); 55 pixel ( mm2) Contrast Levels: 14 / 16 levels Apparatus Ambient illumination: 110 lux Viewing distance: 30 cm Display: IBM T221 QUXGA wide TFT-LCD (3840x2400) was reworked to display 7140 gradations from 0.8cd/m 2 to 244.8cd/m 2 ) Contrast Sensitivity Tests - VectorVision CSV-1OOOE 14

15 Results of Subjective Tests on Pseudo Achromatic Mura Contrast Threshold (%) Relation between contrast threshold and area of different Mura types on the different background levels L64-line L368-line L512-line L900-line L64-circle L368-circle L512-circle L900-circle /f(S) Contrast thresholds were dependent on the luminance of backgrounds. L64 was different from L368, L512, or L900 by multiple comparison analysis. The average threshold on L368 was higher than others. In other words, L368 was the best background level to detect Mura for human eyes. To large muras, contrast thresholds were different between linear mura and circle mura. 15

16 Comparisons between SEMU and TTLA Results Contrast (%) f(s)=s 0.33 S: unit (mm 2 ) TTLA vs. SEMU y = x R 2 = /f(s) Contrast threshold on middle gray level nearby 50 cd/m 2 SEMU 1.0 JND SEMU 2.0 JND SEMU 3.0 JND TTLA Cjnd=1.97(1/f(S))+0.72 R 2 =

17 Border Effects on Pseudo Achromatic Mura positive polarity Mura negative polarity Mura Border effects 17

18 Subjective Test for Mura Item 1. Pseudo Achromatic Mura 2. Real Local Mura Factor Background luminance Shape & size (Homogeneously-illuminated pattern) Boundary (Occurrence positions) Typical mura shape (in-homogeneouslyilluminated pattern) Independent Variable L64 (1.36 cd/m 2 ), L368 (23.97 cd/m 2 ), L512 (51.66 cd/m 2 ), L900 ( cd/m 2 ) Line: 7 2 pixels (3.96 mm 2 ); pixels (68.56 mm 2 ); 640 2; 1280 pixels ( mm 2 ) Circle (Diameter, mm): 7 pixel (10.76 mm 2 ); 25 pixel ( mm 2 ); 55 pixel ( mm 2 ) Bottom bezel, Corner bezel 11 types of Mura (bright/dark spot, spacer, weak-line, upper light-leakage, dark corner, rubbing, v-band, non-uniformity) 3. Pseudo Monochromatic Mura Full R/G/B background Size (Homogeneously-illuminated pattern) R: Yxy=(36.23, 0.63, 0.34); G: Yxy=(100.5, 0.28, 0.61); B: Yxy=(26.47, 0.14, 0.11) Circle (Visual angle, degree): 0.33, 1.04, 2.84, 5.13 和

19 Pseudo Achromatic Circle Mura Main Effect: Occurrence Position & Mura Size (Interaction exits between position and size) For Contrast JND, there is no difference between experts and novels. This results are similar to SEMU study. There is also no different effect for bright Mura and dark Mura. 5 Pseudo Achromatic Circle Mura Contrast Threshold (%) P_Circle_Novel N_Circle_Novel P_Edge_Novel P_Coner_Novel P_Circle_Expert N_Circle_Expert P_Edge_Expert P_Coner_Expert Visual Angle (degree) 19

20 Experience Effects Average of Cjnd (%) Subject (Expert vs Novel) Novel Expert Visual Angle (degree) There is no significant difference between experts (inspector) and novels (end-users). 20

21 Pseudo Achromatic Circle Mura 3 Polarity Effect on Circle Mura Average of Cjnd (%) Positive Polarity Negative Polarity Visual Angle (degree) There is no significant difference between Positive Polarity and Negative Polarity. 21

22 Results of Border Effects for Pseudo Achromatic Mura Five Homogeneously-illuminated pattern circles The influence of border: Bottom bezel and corner bezel Eleven subjects: 7 novices and 4 experts 22

23 Contrast threshold(%) Relation between visual angle and contrast threshold under different Mura positions Curve Fitting y = x Center Mura R 2 = Edge Mura Corner Mura y = x R 2 = y = x R 2 = /Visual Angle 23

24 Subjective Test for Real Local Mura 01-kido 02-spacer 03-spacer 04-dark corner 04-v-band 05-rubbing 06-spacer 07-weak line 08-light leak 09-non-uniformity 11-thin_rubbing_line 24

25 Subjective Test for Real Mura Results of statistic test revealed that there were significant effect between Expert and Novel. For most real mura, Expert is easier than Novel to detect mura. This finding differs from the results of experiments of pseudo achromatic mura. Real Mura vs. Contrast Threshold Average of Contrast Threshold (%) Novel Expert 01-kido 02-spacer 03-spacer 04-dark_corner 04-v_band2 05-rubbing 07-spacer Mura Type 07-weak_line2 08-light_leak 08-nonuniformity 11-thin_rubbing_line2 25

26 Summary of Subjective Tests 26

Key Features of Multi-View Mura Inspection Hardware (236cm (W) x 202cm (D) x 184cm (H)) DUT: 10 ~ 19 LCD Module 5 viewing angles: normal, inclination θ H =±45 and θ V = ±25 5 CCD cameras (@ 2kx2k,

27 Key Features of Multi-View Mura Inspection Hardware (236cm (W) x 202cm (D) x 184cm (H)) DUT: 10 ~ 19 LCD Module 5 viewing angles: normal, inclination θ H =±45 and θ V = ±25 5 CCD cameras (@ 2kx2k, mono 12 bits, 66dB dynamic range) Beveled CCD sensors for imaging at 45 & 25 Anti-aliasing (moirê): out of focus Pattern generator: ASTRO VG-845T The Prototype of MVMI Software Processing time 30 each viewing angle Digital keystone correction Image file format: 16 bits Tiff Mura types: Cluster (GAP, Gomi, Spot, Spacer), Rubbing, Stripe, H/V-band, Weak-line, Around, Light-leakage etc. Whole panel quality: GO/NOGO via an vision model Outputs: mura JND map, local JND, and Semu 27

28 Results of Mura Image Processing 28

29 Implementation of Vision Model Image Captured by CCD camera Contrast Computation Contrast map CSF Filtering JND Map The purpose of CSF (Contrast Sensitivity Function) filtering is to simulate the contrast sensitivity of the human vision system. JND (Just Noticeable Difference) value can be used as a totally evaluation of the LCD panel to classify an LCD panel. JND map can help to find out the regions that might contain Mura. Contrast Computation: Minkowski Pooling Minkowski Pooling: JND Value System Flow Chart 29

30 Contrast Sensitivity Function (CSF) Purpose: to simulate the contrast sensitivity of the human vision system. The shape is decided by our experimental results. Human eyes are not sensitive to objects at extremely high and low spatial frequency band. Contrast Sensitivity Function: Simplified Barten s CSF Model CSF curve 30

31 Accuracy of Vision Model Masking effect is taken into account. One model could deal with various shapes. Error between subjects is 2.83 db. Error of vision model is 2.75 db. 31

32 Time-Dependent Mura Will be published in IDW06 Curtain mura appears and varies in a period of time due to chip-on-glass technique. Conditions Two weeks observation, Dark room, Warm-up time for 2 hr, Black test pattern Alternatives Gray value (Intensity), Contrast profile, Local contrast profile, Peak JND, and JND profile Normal Direction No Curtain mura captured Inclination θ V = -25 Curtain mura captured 32

33 Gray value (Intensity) Gray-level profile in X-direction (Colors represent different measuring time slots) 33

34 Contrast profile Global contrast profile in X-direction (Colors represent different measuring time slots) 34

35 Local Contrast Profile Local contrast profile in X-direction (Colors represent different measuring time slots) 35

36 JND Profile Analysis results JND map Inclination θ V = -25 Curtain mura captured JND Position, x

37 Moving Blur-Edge Response Time

38 Response Time Measurement Response Time (on-off / TrTf) In DTV-Taiwan, it is discussing to clarify the different results of measuring form popular instruments. Moving blur-edge response time In this June, we proposed a draft to TTLA/TDSC, titled was Motion blur measurement for displays. Topcon BM series MicroVision RTM-SA Otsuka MPRT-1000 Eldim Optiscope Westar GLRT TTLA MPQ System 38

Tradition Measurement VESA 305-1 Response Time V(t) T 10 T 90 T on L 100 T 90 T 10 T off 時間分辨測光儀 L 90 儲存示波器 L range L 10 L 0 0 0

39 Tradition Measurement VESA Response Time V(t) T 10 T 90 T on L 100 T 90 T 10 T off 時間分辨測光儀 L 90 儲存示波器 L range L 10 L Time, t t = L 100 L 0 T on = T 90 T10 L 0 Lrange+ L T 10 =. 1 L 0 Lrange+ L T 90 = T off f = T 10 T 90 T = T + T on off 39

40 Measurement Setup LCD LCD LCD (a) Tracking Camera System (b) Pursuit Camera System - rotating cameras - rotating mirrors or optics (c) High Speed Camera System 40

41 Measurement Setup (e) Fixed Optical Detectors LCD (f) Software implementation for quantitative visual analysis (d) Pursuit Camera System - rotating mirrors or optics 41

42 Time-base Image Integration Ref: SID 00 DIGEST 30.2: Guiding Principles for High Quality Motion Picture in AMLCDs Applicable to TV Monitors 42

43 Analysis Case 1: No Overshoot or Undershoot L j K ij (s) L i 90% w ij b ij s i s j 10% Relative retinal eye coordinate, s K ij (s): spatial transition b ij : 10 % to 90 % blur-edge width w ij : extended blur-edge width w ij = b ij /0.8 W ij : total blur-edge width W ij = w ij+ w ji (2004) Moving Edge Response Time M 1 = w un i j (2003) where u is scroll speed and N is the number of i j transitions. ij (2005) 43

44 L p Analysis Case 2: Overshoot and/or Undershoot L p L j 90% K ij (s) w ij w ij w ij L i b ij P ij s i s j b ij b ij L v L P ij P v ij Relative retinal eye coordinate, s 10% L j 110% 90% K ij (s) w ij w ij w ij L i s i s j 10% -10% Relative retinal eye coordinate, s 44

45 Implementation of Pursuit Camera System Accuracy and precision Gray-to-Gray 8 gray levels; time for each treatment <10 sec Test speed: 2~20 pixels/frame Pursuit error 1 ms Results were similar to Otsuka MPRT1000 Added Functions Measuring R,G,B,C,M,Y primary and second color motion artifacts. Color Characterization of 3CCD Camera, average of E Motion Picture Quality via Pursuit Camera System Standard Activity Co-work with DTV-Taiwan, TDMDA and TDSC 45

46 Example of Moving Blur Measurement Gray-to-Gray Analysis Color Shift 46

TTLA MPQ vs. Otsuka MPRT-1000 15 pixels per frame Average of Response Time for Gray to Gray (15 ppf) Average= 21.47 Average of Response Time for Gray to Gray (15 ppf) by MPRT-1000 Average= 22.21 35.

47 TTLA MPQ vs. Otsuka MPRT pixels per frame Average of Response Time for Gray to Gray (15 ppf) Average= Average of Response Time for Gray to Gray (15 ppf) by MPRT-1000 Average= Response Time (ms) Response Time (ms) S00 S01 S02 S03 S04 S05 S06 E00 E03 E02 E01 E06 E05 E Start End S00 S01 S02 S03 S04 S05 Start S06 E00 E03 E02 E01 E06 E05 E04 End Standard Deviation of Response Time for Gray to Gray (15 ppf) Standard Deviation of Response Time for Gray to Gray (15 ppf) by MPRT Average= Average= Response Time (ms) Response Time (ms) S00 S01 S02 S03 S04 S05 Start S06 E00 E03 E02 E01 E06 E05 E04 End S00 S01 S02 S03 S04 S05 Start S06 E00 E03 E02 E01 E06 E05 E04 End 47

48 Transition from Red to Green 5 ppf 10 ppf 15 ppf 48

49 Transition from Red to Cyan 5 ppf 10 ppf 15 ppf 49

50 Transition from Red to Magenta 5 ppf 10 ppf 15 ppf 50

51 Tentative Results for 10 pixel per frame Color To Color(10ppf)_MPRT(ms) y/x R1 G1 B1 C1 M1 Y1 R G B C M Y AVG STD Max START R1 Color To Color(10ppf)_MPRT(ms) B1 M1 R M2 Y2 C2 B2 G2 END Color To Color(10ppf)_ColorBiasIndex y/x R1 G1 B1 C1 M1 Y1 R G B C M Y AVG STD Max START Color To Color(10ppf)_ColorBiasIndex R1 B1 M1 R M2 Y2 C2 B2 G2 END 51

52 Conclusions Domestic Industrial International Visual Inspection & Image Quality Continuous Join IEC TC110/WG2, VESA and SEMI activities Harmony of Mura terminology and definition Continuous Efforts to Color blur measurement Mura Contrast Sensitivity Function Semi Mura (SEMU) Japan Modified SEMU Korea, Japan CSF of Mura Taiwan Measuring Systems Standardized test conditions and H/W of mura capture system Standardized S/W of Image processing 52

53 Thank You Very Much!

APPLICATIONS OF HIGH RESOLUTION MEASUREMENT

APPLICATIONS OF HIGH RESOLUTION MEASUREMENT Doug Kreysar, Chief Solutions Officer November 4, 2015 1 AGENDA Welcome to Radiant Vision Systems Trends in Display Technologies Automated Visual Inspection