EPA Energy Star Lighting Webinar Series Evaluating Color Quality – March 31, 2016 Color Quality of Lighting and Metrics – Where are we going to? Yoshi Ohno (Ph.D., CIE President, IES Fellow) NIST Fellow, Sensor Science Division National Institute of Standards and Technology Gaithersburg, Maryland USA Energy Star Webinar March 2016 1
Color Quality of Lighting Color Quality Chromaticity Color Rendition (white light) • Color fidelity (CRI) • Color preference 400 500 600 700 400 450 500 550 600 650 700 • CCT, Duv • Chromaticity coordinates ( x,y ), ( u ’ ,v ’ ) Energy Star Webinar March 2016 2
OUTLINE 1. Research on White Light Chromaticity 2. Research on Color Saturation Preference 3. Understanding TM-30 4. Where are we going to? Energy Star Webinar March 2016 3
OUTLINE 1. Research on White Light Chromaticity 2. Research on Color Saturation Preference 3. Understanding TM-30 4. Where are we going to? Energy Star Webinar March 2016 4
Color Quality = CCT and CRI ? CCT: Correlated Color Temperature (CIE S017 ILV) CRI: Color Rendering Index (CIE 13.3) Some example This product is not acceptable. Why? CCT = 3050 K CRI (R a )= 91 Duv - Too yellowish! Normal CCT does not tell the whole story of chromaticity. Energy Star Webinar March 2016 5
White Light Chromaticity ANSI C78.377 for SSL • Widely used, in Energy Star, DLC, ANSI C78.376-2001 for IEA SSL Annex ,etc. • First published in 2008 linear fluorescent lamps • Revisions in 2011, 2015 Energy Star Webinar March 2016 6
ANSI C78.377 Future Considerations Annex B. 4-step Annex C. 4-step u’v’ circles quadrangle tolerances (CIE TN001) Energy Star Webinar March 2016 7
Preference to below blackbody Known for many years but not covered in any standards. An example: Neodymium lamp Energy Star Webinar March 2016 8
Recent Studies Experiment on NIST vision experiment Experiment on preferred perceived white point in 2013 lighting by NRC, Canada by LRC (2013) (2013) Ohno, Y., Fein, M., Vision Experiment on Dikel et al, ”Preferred Chromaticity of Color - M. S. Rea,* J. P. Freyssinier , “White Acceptable and Preferred White Light Tunable LED Lighting”, LEUKOS, 10:2, 101- Lighting”, CR&A , 30-2 , 82-92, 2013. Chromaticity for Lighting, CIE x039:2014, 115, DOI: 10.1080/15502724.2013.855614 pp. 192-199 (2014). (2013). Energy Star Webinar March 2016 9
NIST Spectrally Tunable Lighting Facility Energy Star Webinar March 2016 10
NIST Spectrally Tunable Lighting Facility Energy Star Webinar March 2016 11
2013 Vision Experiment at NIST on Preferred and Acceptable level of Duv Adaptation points Pair • NIST STLF at ~ 300 lx. 6 Duv points at each CCT, • 18 subjects (20 to 70 yrs old) 4 CCTs, • Viewed fruits/vegetables on the table, at total 23 points. his/her skin tone and the whole room. Total 50 spectra used. • Adapted to each Duv point before judgement: which light is “more natural”. Energy Star Webinar March 2016 12
Which light looks more natural? A Energy Star Webinar March 2016 13
Which light looks more natural? B Which light looks more natural? A Energy Star Webinar March 2016 14
2013 Vision Experiment at NIST on Preferred and Acceptable level of Duv Average results of all subjects • NIST STLF at ~ 300 lx. • 18 subjects (20 to 70 yrs old) • Viewed fruits/vegetables on the table, his/her skin tone and the whole room. Duv=~ -0.015 at all CCTs appeared most natural. • Adapted to each Duv point before judgement: which light is “more natural”. Energy Star Webinar March 2016 15
2013 Vision Experiment at NIST on Preferred and Acceptable level of Duv Duv defined in ANSI C78.377. • NIST STLF at ~ 300 lx. • 18 subjects (20 to 70 yrs old) • Viewed fruits/vegetables on the table, his/her skin tone and the whole room. Further reference Y , Ohno, “Practical Use and Calculation of • Adapted to each Duv point before CCT and Duv” LEUKOS 10:1, 47-55, DOI: 10.1080/15502724.2014.839020 (2013). judgement: which light is “more natural”. Energy Star Webinar March 2016 16
Discussion by Minchen Wei & Kevin Houser LEUKOS, DOI: 10.1080/15502724.2015.1029131, April 2015 Vol. 12, Issue 1-2, 2016 “We infer that the preference expressed by participants in the studies by Dikel and others [2014] and Ohno and Fein [2014] may not be solely related to chromaticity .” “ aspects of color rendition (that is, color fidelity and relative gamut) may also influence preference.” Energy Star Webinar March 2016 17
2015 Experiment 2013 Experiment Gamut area Q g Gamut area Q g 3500 K 3500 K Chroma saturation (red, green) Chroma saturation (red, green) Chroma saturation 3500 K 3500 K Energy Star Webinar March 2016 18
2015 Experiment 2013 Experiment Gamut area Q g Gamut area Q g 3500 K 3500 K Chroma saturation (red, green) Chroma saturation (red, green) Chroma saturation 3500 K 3500 K Energy Star Webinar March 2016 19
Average Results of 2015 vs 2013 Experiment 2013 results 2015 results Std.dev No significant differences found. References Ohno, Y., Fein, M., Vision Experiment on Acceptable Ohno, Y., Oh, S., Vision Experiment II on White Light and Preferred White Light Chromaticity for Lighting, Chromaticity for Lighting, CIE x042:2016, pp. 175-184 CIE x039:2014, pp. 192-199 (2014) (2016) Energy Star Webinar March 2016 20
Proposal for addition to ANSI C78.377 Proposal Current specification Flexible Duv (Center Duv -0.002 to -0.012) “Preference - based” (Center Duv -0.012 ) Energy Star Webinar March 2016 21
OUTLINE 1. Research on White Light Chromaticity 2. Research on Color Saturation Preference 3. Understanding TM-30 and CRI 4. Where are we going to? Energy Star Webinar March 2016 22
Why perception differ from the CRI scores? CRI R a =94 CRI R a =78 Looks better (for most people) Plots of 15 CQS samples Red and Test light on CIELAB green are ( R a =78) (a*,b*) critical The area Reference within such (Planckian) plots is called “gamut area” • CRI is a color fidelity metric based on Reference Illuminant. • Perception is different from color fidelity Energy Star Webinar March 2016 23
Color Gamut and Perception De-saturated Neutral Over-saturated (to some extent) Looks poor Good Looks better Perception and CRI do not Perception and CRI agree agree (CRI works okay) R a =82 R a =82 R a =85 R a =63 R a =100 Energy Star Webinar March 2016 24
2014 Vision Experiment at NIST on preferred chroma saturation level • 20 subjects • 3 CCTs (2700, 3500, 5000 K), D uv = 0 • 3500 K, D uv = -0.015 100 CIELAB 80 D C * ab = 16 60 Most saturated 40 20 a* 0 -80 -60 -40 -20 0 20 40 60 80 -20 -40 D C * ab = -16 -60 b* -80 Most de-saturated Energy Star Webinar March 2016 25
Which light looks better? A Energy Star Webinar March 2016 26
Which light looks better? A Which light looks better? B Energy Star Webinar March 2016 27
Results (1) Entire room Average Preference for D C 100% 100 CRI Ra Percent of Participants who 90% 80% 90 preferred the Point 70% (2) Skin tone of subject 80 60% 50% Average 40% 70 Room 30% Skin 20% 60 Green (3) Red fruits/vegetables 10% Red 0% -16 -12 -8 -4 0 4 8 12 16 D C ab * More saturated De-saturated (4) Green fruits/vegetables Neutral Reference: Y. Ohno, M. Fein, C. Miller, Vision Experiment on Chroma Saturation for Color Quality Preference, CIE 216 :2015, pp. 60 – 69 (2015) Energy Star Webinar March 2016 28
OUTLINE 1. Research on White Light Chromaticity 2. Research on Color Saturation Preference 3. Understanding TM-30 4. Where are we going to? Energy Star Webinar March 2016 29
IES TM-30 IES Method for Evaluating Light Source Color Rendition Color fidelity: Two-metric system Improvement • Fidelity index R f of CRI R a • Gamut index R g Information for 99 test samples preference Latest color space Works accurately for all colors of and chromatic object and all adaptation formula spectra Calculation tool Provide details beyond R f and R g . Energy Star Webinar March 2016 30
Concept of Two-Metric System • R g provides preference-related information. More • TM-30 does not provide saturated design guidance for R g Preferred? • Limitation of R g - ave. of all hues R g equal but different shapes Gamut R g R g =100 R g =100 R f =75 R f =75 Less saturated Image from Michael Royer Fidelity R f Gamut area is not sufficient. Energy Star Webinar March 2016 31
IES Position statement of CRI – 2015 Issued Sep. 2015 IES recognize that CIE CRI has shortcomings, especially with new lighting technologies. TM-30 was developed to solve this problem. “ TM-30-15 is not a required standard, and it does not provide design guidance or criteria for best practices.” “the issuance of TM-30-15 will enable the international lighting community to carefully evaluate it, providing a path leading to improved standards and design guidance.” Link: https://www.ies.org/PDF/PositionStatements/PS-8-15.pdf Energy Star Webinar March 2016 32
Recommend
More recommend
