Display Measurement Standard from SID s ICDM CORM Council for Optical Radiation Measurements May 9-11, 2010 Las Vegas, Nevada Edward F. Kelley, Ph.D., LLC. ed@keltekresearch.comcom 303-651-0787 www.keltekresearch.com (or 4keltek.com) Portions of this presentation are copyrighted. E. F. Kelley of, 2010.
OUTLINE Display Measurement Standard from SID's ICDM ABSTRACT: We review the content of the new standard from the International Committee for Display Metrology (ICDM) of the Society for Information Display (SID). This measurement standard is an extension of a previous standard completed under a different umbrella. Attention will be focused on the structure, new parts, and philosophy of the document. ORIGINS FEATURES SETUP OF DISPLAY TYPES OF DISPLAY METRICS (NEW IN BLUE) (These will be discussed in more detail this afternoon.) Fundamental Gray & Color Scales Spatial Uniformity Viewing-Angle Temporal CONFOUNDING PROBLEMS Reflection Motion Physical, Mechanical, Electrical Front Projector Front-Projection Screen 3D & Stereo
ICDM ORIGINS 1992 NIST begins Flat Panel Display Lab. NIST joins VESA (Video Electronics Standard Association) for rapid development of FPD standard. FPDM version 1 released May 1998. VESA Display Metrology Committee chair (EFK) humiliated by VESA BOD chair, reduced to editor. FPDM version 2 released June 2001 (begrudgingly). Motion Artifacts Update released May 19, 2005. FPDM version 3 started, but. Chair (Joe Miseli) and editor (EFK) were fired by grumpy BOD (we ignored too many rules from micromanagers). Took our toys and went to SID (Society for Information Display) most (70 % to 89 %) of FPDM2 is in public domain by virtue of NIST s contribution. Parent SID Committee: Definitions and Standards Committee Created new: International Committee for Display Metrology (Joe Miseli, chair; EFK editor) VESA is now friends with SID and will be a document distributor of the new SID-ICDM standard.
FEATURES Metrology document NOT performance criteria! No, we don t tell you what you should get. That is the job of other standards bodies. This document tells you how to measure it correctly. But the eye can t see it anyway, so why measure it so accurately?! No, we DON T care if somebody thinks the eye can t see something; we want to measure it correctly regardless! Besides, in many cases, we suspect that what the eye can and can t see has been measured incorrectly anyway! Idea was to supplement other standards bodies that don t want to spend a lot of time writing methods in gory detail a niche.
FPDM2 Worldwide Hit! Specification of good metrology for displays Self-contained measurement procedures Nope, it s not a spaghetti document!!! We have distinct meatballs! Buffet ffetofmeas measurements use se what you need Easy to use and read Avoids technospeak and uses simple sentence structure as much as possible. Appreciated by non-native English speakers Extensible more will be added as needed Adaptable affords a variety of equipment Accommodating special needs permitted Metrology Section Tutorial Section Includes diagnostics, cautions and hints Can be found on a messy desk (lab)! Uses different fonts for headings! (Wow!!!) Now, we need much more! And has cartoons! 5
Migration of Effort Expanding FPDM an ambitious undertaking VIDEO ELECTRONICS STANDARDS ASSOCIATION DISPLAY METROLOGY COMMITTEE FLAT PANEL DISPLAY MEASUREMENTS STANDARD Version 2.0 Abstract: This is a standard to provide measurement procedures to quantify flat panel display characteristics. Performance criteria or performance minima are not specified; rather, a series of measurements are clearly detailed to enable unambiguous and reproducible measurements of displays using the simplest instrumentation that will provide adequate results. All mea surements need not be performed. The measurements that are most applicable to the display purposes can be selected as desired. Diagnostics and metrological difficulties are addressed, and technical discussions are presented to assist those unfamiliar with light measurements. June 1, 2001 This is a sample cover. Changes will be made in the final edition; the name may be changed as well. PDF version freely distributed Printed version in color (good $) FPDM 1 was 322 pages. This puppy is about 500 pages! Target date: By the end of 2010, hopefully sooner. October 6
OLD FPDM2 STYLE CHANGE NEW ICDM 7
SPECSMANSHIP SETUP OF DISPLAY How do manufacturers obtain specifications? Some honestly, some not so honestly How important to consumers are those specifications? Quite important unfortunately. It would be nice if the specifications were always as meaningful as people p think. Are those specifications realistic? Not always. MANUFACTURING SPECIFICATIONS? Some manufacturers have deliberately specified a setup procedure that provides inappropriate display settings. The ICDM standard d will address these matters and not permit such procedures even if the manufacturer specifies them.
SPECSMANSHIP Unrealistic adjustment to enhance specifications Some manufacturer setup specifications dictate adjusting for brightest white to get L W then adjusting for darkest black to get L K gives much greater contrast ratio that way. Never mind that the display couldn t be used with those adjustments Tweak for L W Tweak for L K Nobody would want to use the display with those adjustments! 9
ICDM requires display setup to be reasonable! Adjust display settings to obtain the best possible image under darkroom conditions or other carefully specified ambient conditions. Then L W measure full-screen L W and L K without changing the settings. These are the luminance values to be reported and used to calculate contrast (full-screen darkroom contrast = sequential contrast). L K 10
TYPES OF DISPLAY MEASUREMENTS Fundamental Gray & Color Scales Spatial Uniformity Viewing-Angle Temporal Reflection Motion Physical Mechanical Electrical Front Projector Front-Projection Screen 3D&Stereo Touchscreens??? (Maybe) New in blue. Extensive detail in introductory remarks for Motion, Projector, and 3D sections. Reflection section greatly expanded. 183 measurements & growing.
GRAY & COLOR SCALES Gray Scale INPUT OUTPUT Gray Level Gray Shade (Command Level) L = L K, L 2, L 3,, L W V i = 0, 1, 2, 255 (8-bit) L 1 L K, L 256 L W V 1 V K = 0, V 256 V W = 255 γ Typical form: L a= L W-L K i = av i + L K, 255 γ See the Gray-Scale & Color-Scale Metrics Section for the prescription of selecting a subset of levels from the full number of available levels. 12
SIMPLE GAMMA MODEL This functional relationship between the luminance and gray level has a variety of names depending di upon the group you re with: tone curve, electro-optical l transfer function (EOTF), gamma (just the word), gamma curve, gray-scale curve, tone scale, etc. γ L i = av i +L K a=(l γ W - L K )/(V W V K ) log(l i L K ) = γ log(v i V K ) + log(a) Usually V K = 0 /m 2 ) Gray Shade Luminance, L (cd/ 400 350 300 250 200 150 100 50 0 0 50 100 150 200 250 300 Gray Level, V g(l - LK ) lo 3.0 2.5 2.0 15 1.5 1.0 05 0.5 Linear Regression (y = mx + b): γ = 2.25 ±0.012 b = log(a) = -2..84 ±0.026 r = 0.9999 b = log(a) is where it crosses the vertical axis at log(v) = 0. 1.4 1.6 1.8 2.0 2.2 2.4 2.6 13 log(v)
CONFOUNDING PROBLEMS WHITE PRIMARIES (WHITE SUBPIXELS) MULTIPRIMARY DISPLAYS: RGBY, RGBCMY, GLOBAL DIMMING LOCAL DIMMING WHITE WHAT S THAT??? BLACK & ZERO BLACK LUMINANCE INFINITE CONTRAST
WHITE & MULTIPLE PRIMARIES What are the effects of a white primary (subpixel)??? RGB RGBW What are the effects of multiple primaries??? RGBY RGBCY RGBCMY RGBCVY RGBWCMY Commercially Available! Existing Hypothetical 15
MULTIPLE PRIMARIES RGBY RGBCY RGBCMY RGBCVY RGBWCMY N > 3 color primaries means each color inside the gamut can be achieved in more than one way. Possible metamerisms! Beyond the present scope of ICDM work. Hypothetical (White primary can be used for highlights.) 16
GLOBAL & LOCAL DIMMING GLOBAL DIMMING DISPLAYS: Displays that can change the luminance of the entire backlight (LCD) or the luminance of the full screen to suit the imagery. May be able to achieve absolute black as with full-screen black. To avoid this, use a constant average-picture-level (APL) pattern that cycles the different gray shades through the center region (c, d & f) or adjusts the center and background (e) for the same APL. LOCAL DIMMING DISPLAYS: Even more problematic displays (LCD) that have a low-resolution LED backlight to modify contrast in different regions. Large APL patterns (c) may not work, still might get too black. Pattern (d) may work better. BEST IDEA for both local and global dimming may be to turn the feature off, if possible, for gray-scale and color-scale measurements. Preview of ICDM DMS L S L X (a) (b) (c) (d) (e) INTRO SCPL01 SCPL02 SCPL03 SCPL04 SCPL05 SCPL06 (f) 17
WHITE WHAT S THAT??? FULL SCREEN WHITE: Often reported for displays that t can sustain a bright full screen without a loss of luminance. May include a white primary. PEAK WHITE (HIGHLIGHT WHITE): For screens that cannot maintain their brightest luminance at full screen because of power loading. May include a white primary. Often this is not distinguished i d from full-screen white, unfortunately. t LOADING: Method to determine power loading. Measure increasing white box. L = 100%( L max L W )/L W L W = Full-Screen White L = 63% 18
Fundamental Metrics, Cont. Full Screen & Box IMAGERY WHITE (WITH NO WHITE PRIMARY): If the display employs a white primary, and if the white primary can be turned off, then you can measure a full-screen white. If the white primary cannot be turned off, then use the nonatile trisequence patterns measuring at the center screen. This is an attempt to eliminate or reduce the use of a white primary as much as possible. 3x3 matrices of maximum RGB L IW = L R + L G + L B The name imagery white is not yet fixed as a standard term. It is proposed in the ICDM DMS, but has not been voted in yet. 19
BLACK & ZERO LUMINANCE BLACK FULL SCREEN BLACK: Often reported for displays that exhibit a visible black luminance L K > 0 when the entire screen is at the black level V K = V 1 = 0. IMAGERY BLACK: For displays that can either turn off the pixel entirely (e.g. OLED displays) or turn off the backlight as in global dimming or local dimming displays. We use a pattern that may allow light from a surrounding bright area to leak or bleed into the black area. That pattern has not been agreed upon. Candidates: Star Patterns SCPL01 CAN WE MEASURE IT? If we have low or zero luminance levels, can our instrumentation measure it? It is only zero if we can t see it in a good dark room with adapted eyes. 20
INFINITE CONTRASTS C = L L W K True for L K = 0 no matter what L W is! Already had these displays with some CRTs. Use an intermediate t level l rather than black and call it mid-contrast? (Very gamma dependent.) Are they practical? Really? YES! Some applications: Theme park tunnel ride wants to scare people when they pass by. They won t see the display until it lights up. A gamer wants to feel like he s in outer space. We don t know what to do about such displays yet. 21
THANKS FOR LISTENING! Edward F. Kelley, Ph.D.;, LLC.; ed@keltekresearch.com E. F. Kelley of, 2010, portions of this presentation are copyrighted.