Optimization of Display-Systems

Transcription

1 Optimization of Display-Systems with respect to Glare, Distinctness of Image and Sparkle Michael E. Becker Display-Messtechnik&Systeme - Rottenburg am Neckar display-messtechnik.de

2 Optimization of Displays with an AG-Layer AG no AG glare control ambient light image blur sparkle scattering AG-layer light carrying visual information 2 display-messtechnik.de

3 Optimization Task + reflections gloss glare image distortions image blur + fingerprints - - visual artefacts sparkle the optimum display task: minimize reflections, image blur, sparkle and costs -... expenses costs + 3 display-messtechnik.de

4 Characterization of Reflectance 4 display-messtechnik.de

5 Characterization of Reflectance Konica Minolta CM display-messtechnik.de

6 Characterization of Reflectance Both devices are made for surfaces covered with paint, pigments, etc. They are NOT suited for characterization of transparent samples. 6 display-messtechnik.de

7 Characterization of Reflectance Helmholtz reciprocity specular included TIS (total integrated scatter) specular only high-resolution reflectance distribution function 7 display-messtechnik.de

8 Characterization of Reflectance? with a gloss-meter LMD LMD Gloss-Meter reading 1 Gloss-Meter reading 2<1 Gloss level of 100% given by specular reflection from a polished black glass surface (n = 589 nm). Increasing gloss level decreasing scattering (at same substrate thickness!) 8 display-messtechnik.de

9 Characterization of Reflectance High-Resolution Scatter Analysis of Anti-Glare Layer Reflection Proc. SID 2016 display-messtechnik.de

10 Scatter Analysis: How Directional Scanning dl ( θ, φ ) = B( θ, φ,θ, φ ;λ,p) de( θ φ ) s s s i i s s i, i The scattering properties of surfaces are generally and completely described by the bidirectional scatter distribution function (BSDF) which is a function of the direction of light incidence, the direction of observation, the wavelength of light and its state of polarization. Assessment and evaluation of the reflective properties of surfaces can be realized by mechanical (motorized) scanning of a range of observation directions with photometric or spectro-radiometric receivers for one direction of light incidence. This can be done with complex, and bulky high-precision mechanisms (gonio-photometer or gonio-spectroradiometer). Scanning of the directions of observation without moving parts: optical scanning (conoscopy), analysis of the spreading of a point- or line-source of illumination (PSF - LSF approach), hemispherical projection and imaging ("imaging hemisphere"). 10 display-messtechnik.de

11 Optical Directional Scanning Conoscopic approach collimated beam illumination Imaging approach point-source illumination θ s θ θ φ Basic components of reflection obvious in PSF-image: specular peak (width of the source) haze ("fuzzy ball" in image, bell shaped profile) Lambertian (constant plateau) 11 display-messtechnik.de

12 Optical Directional Scanning Conoscopic approach Imaging approach directional illumination conoscopic lens E(x,y) point-source illumination E(0,0) directions image ilmd objective lens ilmd directional analysis of light from one area element directional analysis of light from an extended area! properties must be uniform! 12 display-messtechnik.de

13 Optical Directional Scanning Conoscopic imaging system Directions image Fourier transformation??? 13 display-messtechnik.de

14 Optical Directional Scanning Conoscopic imaging system - reflective illumination 1 14 display-messtechnik.de

15 Optical Directional Scanning Conoscopic imaging system - reflective illumination 2 external isotropic point source illumination Conoscopic lens system used as fisheye lens 15 display-messtechnik.de

16 16 display-messtechnik.de PSF - LSF Analyis vs. Directional Scanning BRDF = I(θ r θi θ i θ r S = s s y x : S = r r y x R : R x y ( ) = θ s s i y x x arctan x ( ) = θ r r r y x x arctan x PSF = I(θ*(x) ) θ i S x y = s s y x S: = r r y x R : R θ r Method 1 Method 2 θ* > 0 θ* < 0

17 PSF - LSF Analyis vs. Directional Scanning S θ r y R x s = -50 mm, y s = 100 mm x r = 50 mm, y r = 100 mm θ i θ i (x = 0) = 26,56 x BRDF= R sinθr sinθi ( θ ) exp i w Reflectance vs. Theta-r 0,06 0,05 0,04 0,03 0,02 0, Theta-r 80 [ ] display-messtechnik.de

18 PSF - LSF Analyis vs. Directional Scanning S θ r y R x s = -25 mm, y s = 250 mm x r = 25 mm, y r = 250 mm θ i θ i (x = 0) = 5,7 x BRDF= R sinθr sinθi ( θ ) exp i w Reflectance vs. Theta-r 0,06 0,05 0,04 0,03 0,02 0, Theta-r 80 [ ] display-messtechnik.de

19 PSF - LSF Analyis vs. Directional Scanning Difference between the reflectance from PSF and BRDF vs. θ * for two source orientations: θ i (x=0) = 26 and θ i (x=0) = display-messtechnik.de

20 Analyis of Line-Spread Function camera image 2D geometry with narrow linear light source side view intensity profile, i = f(θ * ) 20 display-messtechnik.de

21 Variation with Angle of Incidence Reflectance + Transmittance t (θ i ) r(θ i,θ r ) AG + rear surface In the case of transparent/translucent layers reflection and transmission are closely coupled. Reflectance distribution of an AG layer without (yellow curve) and with rear surface reflection (blue curve); the red curve shows the light source (system signature). 21 display-messtechnik.de

22 Variation with Angle of Incidence Reflectance + Transmittance r(θ i,θ r ) AG + rear surface AG + reduced rear surface Black absorbing layer reduces reflections at the rear surface to ~10% of the original level. Reflectance distribution of an AG layer without (yellow curve) and with rear surface reflection (blue curve); the red curve shows the light source (system signature). 22 display-messtechnik.de

23 Reflectance Distribution Function 23 display-messtechnik.de

24 Reflectance Distribution - Characteristics R s : reflectance in specular direction sample Haze [%] with user selectable windows (width and location) source 24 display-messtechnik.de

25 High-Resolution Scatter Measurement and Evaluation Directional Scanning by PSF / LSF Analysis simple compact measurement setup without moving parts robust apparatus and procedure centered about the specular direction never miss the specular peak! ilmd images directly show "what's going on" very high directional resolution (improvement of ~10) and wide range of directions (cylinder geometry) colorimetric analysis via ilmd or illumination Requirements & Limitations sample must be uniform over area included in the measurement limited range of inclinations centered about specular or wide angular range at reduced resolution (cylinder). Verification by model calculations and other instruments Proc. SID display-messtechnik.de

26 High-Resolution Scatter Measurement and Evaluation Our model calculations show that the effect of the variation of the Fresnel reflection coefficient and that of the shape asymmetry of the reflectance distribution function with angle of incidence on the results obtained by PSF/LSF analysis is negligible for geometries of practical importance, i.e. when a limited angular range about the specular direction (e.g. up to ±20 ) is evaluated for small angles of light incidence (i.e.< 15 ). We thus conclude that scatter evaluation based on analysis of point-spread and/or line-spread functions (reflective and transmissive case) is well suited for accurate, fast and robust evaluation of the glossy scatter of AG layers. Colorimetric analysis is possible with either monochromatic light sources or with imaging colorimeters in combination with white light sources. PSF and LSF analysis with imaging photometers thus offers a realistic alternative to conventional directional scanning with goniometric or conoscopic instruments due to the reduced instrumental efforts without moving parts or demanding optical systems. This approach offers easy alignment of the setup with adaptation to a wide range of incidence angles and fast measurements in combination with high directional resolution. The required uniformity of the samples across the surface area included in the measurement is required anyway for uniform optical appearance of electronic display devices. 26 display-messtechnik.de

27 Evaluation of Image Blur Image Blur Induced by Scattering Anti-Glare Layers distinctness of image - image clarity - image blur M. E. Becker, T. Fink, U. Krüger: Image Blurring Induced by Scattering Anti-Glare Layers, Proc. SID display-messtechnik.de

28 Image Blur Micro-structured, scattering anti-glare layers reduce mirror images, but they also reduce distinctness of (transmitted) image (DOI), image clarity,.... blurred image increasing distance unblurred image scattering AG-layer distance, d 28 display-messtechnik.de

29 Pixel Crosstalk Display Screen without AG-layer contrast = 500 2D impulse response Display Screen witht AG-layer pixel crosstalk T. Fink - SID2016 x 2D impulse response top view cross section 29 display-messtechnik.de

30 Measurement Setup line of sub-pixels contrast lateral dimension 1D impulse response width contrast 1D impulse response x width top view cross section 30 display-messtechnik.de

31 Image Blur Characterization G R B G G G G pixel crosstalk, PCT line spreading level, S LS L s-1 L s L s+1 averaging L' s-1 L' s L' s+1 PCT = L s 1 + L 2 L s s+ 1 S LS = L' s 1 + L' 2 L' s s display-messtechnik.de

32 Pixel-Crosstalk - Line Spreading Horizontal intensity profiles of activated green subpixels of a non-scattering display screen (0.10 mm pixel pitch, with touch panel) recorded with a microscope objective lens. Red curve: green subpixel column without AG-layer. Yellow curve: green subpixels with AG-layer. Blue curve: a pulse illustrating the ideal subpixel profile. 32 display-messtechnik.de

33 Measurement Setup ilmd W lateral dimension difference angle 2 step responses impulse response M ilmd top view cross section x stationary input MTF 33 display-messtechnik.de

34 Pixel-crosstalk vs. line spreading Pixel-crosstalk, PCT, and line spreading levels, S LS, evaluated for eight AG-samples by the authors in two laboratories (Lab-1, Lab-2) with two different instrumental arrangements plotted vs. reflection gloss from Electronic (specifications Displays of 2010: the manufacturer). Matte - Glossy Displays M. E. Becker, T. Fink, U. Krüger: Image Blurring Induced by Scattering Anti-Glare Layers, Proc. SID display-messtechnik.de

35 Characterization of Image Clarity Slanted edge method - ISO (step analysis) From: 3M AG/AS: A single-film solution to glare and sparkle in HD displays 35 display-messtechnik.de

36 Characterization of Image Clarity Pixel crosstalk 2D impulse response Line spread analysis 1D impulse response distance source - AG-layer introduced by Dr. SID2016 introduced here Transmission distribution function line spread analysis conoscopic imaging Modulation transfer factor or function stationary input (grating) Slanted edge method (step analysis) 36 display-messtechnik.de

37 Line Spreading vs. Distance Transmittance Source 0.5mm 1.5 mm 7mm 30mm AG layer close to pixel matrix: pixel crosstalk increasing distance, d s : transmittance distribution 10 general: line spreading function camera 300 pixel same AG layer, transmittance distribution vs. distance to line source, d s Transmittance line spreading functions Source 0.5mm 1.5 mm 7mm 30mm peak vs. distance pixel crosstalk transmittance distribution ds [mm] line source camera 50 pixel display-messtechnik.de

38 Transmittance Distribution Function camera image Camera 2D geometry with narrow linear light source transmitt. distr. function pixel crosstalk side view intensity profile, i = f(θ * ) 38 display-messtechnik.de

39 Scatter Distribution Functions intensity G-Series 1000 conoscopic imaging G G-Series 100 G130 G110 G source θ * [ ] LSF analysis 100 G70 G110 G source θ * [ ] 5 M. E. Becker, T. Fink, U. Krüger: Image Blurring Induced by Scattering Anti-Glare Layers, Proc. SID2016 luminance 39 display-messtechnik.de

40 Scatter Distribution Functions intensity H-Series conoscopic imaging H H-Series 100 H source θ * [ ] 5 LSF analysis H20 10 luminance H5 source θ * [ ] 5 M. E. Becker, T. Fink, U. Krüger: Image Blurring Induced by Scattering Anti-Glare Layers, Proc. SID display-messtechnik.de

41 Specification of Scatter Specification of scattering by generalized haze formalism: H ( α, δ ) I α 2 I + I α α δ + δ = 0 selection of α important; e.g. 1% level.! α δ! 41 display-messtechnik.de

42 Specification of Scatter Specification of scattering by generalized haze formalism: H ( α, δ ) I α 2 I + I α α δ + δ = 0 α δ 42 display-messtechnik.de

43 Scatter Distribution Functions - Transmittance intensity t: transmittance H-Series in regular direction 1000 H Haze [%] with user selectable windows (width and location) source intensity H G-Series θ * [ ] 5 Transmittance distribution functions 1000 G G130 G110 G90 source θ * [ ] 5 43 display-messtechnik.de

44 Idealized Pixel Raster Fast evaluation of DOI-MT vs. pixel density Support of routine evaluations in the laboratory 44 display-messtechnik.de

45 Distinctness of Image - Modulation Transfer glass M 0 avg(max = avg(max i,0 i,0 ) avg(min ) + avg(min i,0 i,0 ) ) M AG = avg(max avg(max i,ag i,ag ) avg(min ) + avg(min i,ag i,ag ) ) DOI-MT = M AG / M 0 45 display-messtechnik.de

46 Modulation vs. Pixel Frequency Modulation Transfer Function 1 0,95 0,9 0,85 85 ppi Modulation vs. pixel frequency AG1 0,8 0,75 AG2 326 ppi 0, cycles 10 / mm Modulation vs. pixel frequency (= 25.4 / pixel density [ppi]) for two low-scatter AG layers. 46 display-messtechnik.de

47 Transmittance in the Regular Direction The transmittance in the regular direction, t, is a sensitive measure for the "clarity" of scattering samples. 0,80 0,75 0,70 t vs. gloss G-Series H-Series 0,65 0,60 0,55 0, gloss 120 [%] 130 Transmittance in the regular direction, t, of eight AG samples vs. gloss level as specified by the manufacturer 100 % ~92 % θ* [ ] Scattering distribution functions close the the regular direction (±0,3 ) ASTM D Standard Test Method for Transparency of Plastic Sheeting A. C. Webber: "Method for the Measurement of Transparency of Sheet Materials", JOSA 47(1957)9, no scattering 100 % <92 % scattering 47 display-messtechnik.de

48 What is Sparkle? Sparkle: disturbing optical effect on direct view displays that are provided with scattering anti-glare (AG) layers. M. E. Becker, J. Neumeier: Optical Characterization of Scattering AG-Layers, Proc. SID'11 48 display-messtechnik.de

49 Visual Perception of Sparkle We do see: statistic intensity (and chromaticity) modulations vs. location on the display, and vs. direction of observation. We do not see: individual pixels of the display screen. AG1 AG2 Sparkle is not restricted to one plane in space. 49 display-messtechnik.de

50 Generation of Sparkle display green Water-droplets sprayed onto an LCD-monitor screen: formation of convex micro-lenses. Refraction explains what we see. display white only green subpixels ON all subpixels ON 50 display-messtechnik.de

51 Sparkle Sparkle: statistic intensity and chromaticity modulations vs. location on the display and vs. direction of observation. surface topography + Generation of sparkle: Superposition of two structured layers, Modulation of transmitted light by refraction, diffraction and scattering. 51 display-messtechnik.de

52 Sparkle Sparkle: statistic intensity modulations vs. location on the display and vs. direction of observation. + Sparkle is distinctly visible with green illumination: Superposition of two structured layers, Modulation of transmitted light by refraction, diffraction and scattering. 52 display-messtechnik.de

53 Sparkle 85 ppi 104 ppi 134 ppi Image of display pixel matrix with sparkle = superposition of regular (pixel matrix) and statistical (sparkle) intensity modulations (with green illumination). 190 ppi 326 ppi 53 display-messtechnik.de

54 Visual Perceptionof Sparkle p d one cycle β [ ] viewing distance, d v Observation condition Adjustment of viewing distance C r = 1,01 to make pixel pattern disappear while sparkle remains visible. f [cycles/ ] = 1/ arctan vis,0 ( p / d ) d v Kathy Mullen: J. Physiol. (1985), 359, pp display-messtechnik.de

55 Periodicities and Frequencies periodicities λ p frequencies Display pixel matrix introduces periodic intensity variations while visual sparkle is caused by statistic intensity modulations (vs. location on the display and vs. direction of observation). Display pixel dimensions range from 0,3 mm (PC desktop monitors for office work) to about 0,07 mm (high resolution display screens for handheld devices). Subpixel dimensions thus are in the range from 0,1 mm to 0,02 mm. Sampling: >5 camera pixels (3.75 µm pixel pitch) per display pixel, up to 1:1 imaging. Surface structures of AG-layer with average wavelengths in the range from 5 µm to 50 µm. 55 display-messtechnik.de

56 Periodicities and Frequencies periodicities λ p frequencies Separation of periodic intensity variations from statistic intensity modulations in order to extract sparkle requires a low-pass filter for the spatial frequencies. 56 display-messtechnik.de

57 Sparkle Evaluation Spatial filtering (convolution) with rational kernel bare pixel matrix without noise filtering Sparkle level = σ / µ σ = 0 pixel matrix with noise filtering Sparkle level = σ / µ σ > 0 57 display-messtechnik.de

58 Sparkle Evaluation Single Image Method (SIM) Difference Image Method (DIM) Record one image of pixel matrix with AG-layer, Record first image of pixel matrix with AG-layer, evaluate pixel ratio (>5). evaluate pixel ratio (>5). Shift AG-layer (some mm). Record second image of pixel matrix with AG-layer. Calculate difference image Apply spatial filtering to remove periodic modulations caused by the pixel matrix of the display.y Evaluate Evaluate Apply spatial filtering to remove periodic modulations caused by the pixel matrix of the display. Sparkle level = σ / m Sparkle level = σ / m 58 display-messtechnik.de

59 Sparkle in the Spatial Domain Pixel matrix of tablet computer, green subpixels only. Pixel pitch = 0.1 mm Pixel matrix of tablet computer, green subpixels only with low-sparkle AG-layer. 59 display-messtechnik.de

60 Sparkle in the Frequency Domain f 0 Result of 2D Fourier transformation of an image of the pixel matrix without AG-layer, Fourier amplitudes (indicated by grey levels with linear scaling) vs. spatial frequencies in x and y-direction with same scaling of frequency axes as below. f 0 amplitudes - linear scaling Result of 2D Fourier transformation of an image of a display with a low sparkle AG-layer, Fourier amplitudes (indicated by grey levels with logarithmic scaling) vs. spatial frequencies in x and y-direction. The length of the arrow represents the spatial frequency at the corresponding azimuth angle, φ. f 0 is the fundamental frequency of the display pixel matrix. amplitudes - log. scaling 60 display-messtechnik.de

61 Sparkle in the Frequency Domain Image 2DFT Filtered Image 2DFT s<1% sparkle offset / bias s = 4.95% s = 6.60% 61 display-messtechnik.de

62 Effect of Filtering in the Frequency Domain 62 display-messtechnik.de

63 Sparkle Evaluation R. Adler, et al.: US Patent , 1990 sparkle = "random moiré" D. R. Cairns, P. Evans: "Laser Speckle of Textured Surfaces: Towards High Performance Anti-Glare Surfaces", Proc. SID2007 laser speckle D. K. P. Huckaby, D. R. Cairns: "Quantifying "Sparkle" of Anti-Glare Surfaces", Proc. SID2009 laser speckle M. E. Becker, J. Neumeier: "Optical Characterization of Scattering Anti-Glare Layers", Proc SID2011 J. Gollier, et al.: "Display Sparkle Measurement and Human Response", Proc. SID2013 T.-W. Hsu, et al.: "Novel Evaluation Method of Sparkle for LCDs with Different Anti-Glare Films", Proc IDW2014 M. E. Becker: JSID 2015, Proc. SID display-messtechnik.de

64 Sparkle Evaluation M. E. Becker, J. Neumeier: "Optical Characterization of Scattering Anti-Glare Layers", Proc SID2011 Spatial filtering with rational kernel matched to display pixel matrix. J. Gollier, et al.: "Display Sparkle Measurement and Human Response", Proc. SID2013 phase Pixel power deviation (PPD): Boundaries between adjacent pixels are identified. Total intensity within each pixel is integrated and normalized by dividing by the pixel powers from the bare pixel matrix image. The standard deviation of the distribution of pixel powers is then calculated to give the PPDr value. T.-W. Hsu, et al.: "Novel Evaluation Method of Sparkle for LCDs with Different Anti-Glare Films", Proc IDW2014 Summation of Fourier amplitudes. Spatial filtering with rational kernel (matched to display pixel matrix) offers highest sparkle sensitivity. 64 display-messtechnik.de

65 Effect of Lens Aperture α: angle subtended by the pupil at the target, aperture angle of the pupil. α viewing distance (VD): 250 mm mm α : min arctan (3 mm / 800 mm) = 0,22 (~13'), α : max arctan (9 mm / 250mm) = 2,06. pupil diameter (PD): 3 mm to 9 mm depending on the state of adaptation. Larger aperture of objective lens averages out directional variations reduction of sparkle level. Pronounced increase of sparkle level with F# (= f / D) can be measured. x 3.5 Only measurements with the same aperture angle should be compared. 65 display-messtechnik.de

66 SMS-1000 Features Measurement and Evaluation of Sparkle (2 methods) Distinctness of Image (MTF) Transmittance Distribution Reflectance Distribution optional microscope head, features continuously expanded. 66 display-messtechnik.de

67 Accessories Microscope head The optional microscope head is provided for detailed photometric analysis of small features, i. e. picture elements (pixels) and sub-pixels of display devices, for example, for evaluation of "pixel crosstalk". It comprises a translational stage with fine adjustment (FA), a photometric camera and a variety of objective lenses with an imaging ratio of 1:2 or 1:4. Pixel pattern matrix replacing actual displays FA Slit mask for transmissive illumination (line source) 67 display-messtechnik.de

68 Sparkle Measurement System SMS-1000 Measurement & Evaluation of Sparkle, DOI (MTF) in transmission, Reflectance distribution function, Transmittance distribution function. 68 display-messtechnik.de