The nCRI Colour Rendering Index János Schanda University of Pannonia Virtual Environments and Imaging Technologies Laboratory Overview The CIE system of colour rendring evaluation Colour fidelity and colour preference Why updating the CIE CRI calculation method? The colour fidelity index proposal Summary Brief History of CIE colour rendering 1948: 8 band Spectral Band Method (SBM), deviation from full radiator 1955: established WC 1.3.2 to address terminology and compare SBM with test sample method 1961: agreed on test color method, with 8 test samples 1964: Publication 13 (1st edition), published test sample method Brief History of CIE colour rendering 1974: Publication 13 (2nd edition) Defined reference illuminants Test samples: 8 + 6 Von Kries chromatic adaptation transform Use the CIE 1964 UCS Scaling: Warm white halophosphate lamp = Ra=51 This is the CRI as we know it today CIE 1974 Test method Select a reference illuminant Planckian if TCC<5000 K Phase of Daylight if TCC>5000K Determine Test Sample Tristimulus Values Transform into U*V*W* space Apply von Kries chromatic adaptation transform Calculate colour differences Calculate special and general colour rendering indices T es t S o u rce R e f. CC T Ilul m . T es t C h rom . A d a p t. S am p el s XY , Z , XY , Z , U *V , *W , * U *V , *W , * E R i R i a Brief History of CIE colour rendering 1980s: new TC worked on subject, but closed without recommendations due to member disagreements 1995: Publication 13 (3rd edition), no basic changes from 2nd edition, just fixed some errors 1999: Another TC (TC 1-33) closed without reaching consensus. Industry opposed proposed new procedures due to lack of visual experiments Recent CIE Activity 2007: TC 1-62 published CIE 177 Studied color rendering of white LED sources Recommended the development of a new color rendering metric (or set of metrics) Recommended concurrent use of CRI and new metric, at least at first The new metric should be applicable to all types of light sources Current Standardization Work TC 1-69: Colour Rendition by White Light Sources Started in late 2006 Chair: Wendy Davis, US Terms of Reference: To investigate new methods for assessing the colour rendition properties of white-light sources used for illumination, including solid-state light sources, with the goal of recommending new assessment procedures. Closing the TC? CIE Div.1 will have to decide in 2012 on future; most probably TC will close with description of some possible colour fidelity and colour preference metrics. Proposed metrics • Rank-order based color rendering index (RCRI) • Feel of contrast index (FCI) • CRI-CAM02UCS nCRI • Color quality scale (CQS) • Harmony rendering index (HRI) • Categorical color rendering index (CCRI) • Memory CRI (MCRI) Questions raised with current CIE Test method Reference illuminant: Use single reference? CIE D65? What about visual scene in the evening? Questions raised with current CIE Test method Colorimetry U*,V*,W* space outdated Von Kries chromatic adaptation transform outdated Use CIECAM02, with built in chromatic adaptation transform and its extendsion: UCS colour space Questions raised with current CIE test method Test samples CIE 8 unsaturated plus 4 basic colours + complexion & foliage Problem: Samples from few pigments Not representative real world samples CIE Test Samples Ph-LED YAG 4-LED with yellow CAM02-Ri 4,93 3,54 1,07 93,5 68,0 91,0 77,9 77,9 95,1 80,7 61,5 90,5 92,1 89,8 75,5 79,4 89,0 78,7 66,5 97,6 88,5 97,5 97,9 89,7 82,0 91,4 70,9 88,6 92,8 88,3 94,7 98,7 87,9 96,3 95,7 83,2 76,5 88,5 83,3 5,57 2,20 9,32 4,13 5,77 6,38 8,87 89,3 70,9 78,5 71,8 88,2 86,8 74,2 86,4 82,5 91,3 71,2 78,3 72,1 85,2 89,1 85,2 98,0 74,3 77,8 88,7 92,8 67,8 91,0 75,6 88,8 71,1 78,0 86,3 84,6 51,9 83,9 48,4 93,7 90,9 85,9 94,8 97,4 92,5 82,8 89,5 90,3 96,1 84,9 88,6 87,7 88,2 82,5 91,1 89,5 92,4 70,6 80,5 90,5 84,4 89,4 78,7 88,7 90,7 91,6 84,2 84,2 88,3 76,7 84,1 97,3 91,7 91,3 94,8 85,6 93,5 98,1 90,7 CIE2 2 4 14 18 20 24 25 4-LED no yellow 5 19 31 4-LED-2 max R(9-12) CIE1 4-LED-1 max Ra 3-LED model gamut expanded 3-LED-3 3-LED-2 3-LED-1 E*ab Colour fidelity metric Colour appearance should be the same under the test source and under the reference illuminant. Test samples should provide unique results, and/or number of samples should be high enough to avoid scaling differences using different sample sets. The method should replace the current colour rendering test method Visual observations in double booth with Macbeth Color Checker Chart samples Fidelity experiment Correlation between visual ranking and measured colour differences Light source CIE 13.2 CIELAB CIECAM02 CoolWhite CFL(940) -0,39 -0,42 -0,29 CoolWhite CFL(840) 0,67 0,57 0,64 CoolWhite Fl., tradit. 0,69 0,50 0,75 CoolWhite Tri-band 0,49 0,47 0,60 P-LED -0,62 -0,37 0,04 RGB-LED 1 0,75 0,74 0,79 RGB-LED 2 0,82 0,82 0,89 Recommended new colour fidelity metric: nCRI Flowchart of the new metric: Determine. Ref. Ill. SPD Use new test samples Calc. Tristim. Values Transform to CIECAMUCS Calc. Col. Diff. Average Calculate CRI2012 Reference illuminant in nCRI Reference illuminant: Stay with CIE 13.3 ref. illuminants: Below 5000 K: Planck distribution Above 5000 K: Phases of daylight Determination of equal CCT: u,v – diagram, 2° observer One could consider to move the crossing temperature further down, as at sunset the daylight CCT might be below 5000 K. Test samples in nCRI New test samples to avoid lamp spectrum optimisation to maximize CRI 3 sets of samples: HL17, artificial set of 17 colours sampling in the colour solid 90 samples of high colour constancy 90 samples of low colour constancy Artist colours & skin tones Tristimulus and CIECAM02 calculations USE CIE 10° observer CIECAM 02 parameters: Background Yb=20 Average surround condition with F, Nc, c parameters of 1, 1, 0.69, respectively. Luminance of test adapting field: LA=100 cd/m2 Adaptation factor: D =1. Colour difference and averaging calculation Colour difference in CIECAM 02-UCS space N Root-mean-square averaging: Avoid negative Ri values: Δ E rm s ΔE 2 i i 1 N Sigmoid type functions: 2 2 General colour rendering index: R a ,2012 100 k Δ E 1 e Similar equations for the special colour rendering indices Constant k sets the average of the Ra,2012 of the CIE F 1 to F12 lamps equal to the Ra index, preliminary value is k=1/55. 1 ,5 rm s Practical examples nCRI output The Excel program provides spectrum of the lamp the CCT and (uv) of the test lamp the following general colour rendering indices: Ra, 2012 CIE Ra CQS Qa nCRI output Pictorial information: Gamut of the 17 samples under test lamp and reference illuminant Visualisation of the 17 samples’ colours Information of the worst colour from the 210 sample set Auxiliary information If requested the program can output also such informations as vectors between corresponding test lamp and reference illuminant illuminated 210 samples, detailed information on the sample lowest Ri index, for the different groups of test samples Correlation between Ra,2012 and CIE Ra, CQS Qa 100.0 R a, Qa 80.0 60.0 CIE Ra 40.0 CQS Qa 20.0 0.0 0.0 20.0 40.0 60.0 80.0 100.0 Ra,2012 Above Ra,2012=80 correlation high, For Warm White CFLs Ra,2012 is lower than other metrics provide The program is available for trial. Thanks for your kind attention! This publication/research has been supported by the TÁMOP-4.2.2/B-10/1-2010-0025 project.