Are the light sources oflight cabinets temporarily stable?
Almuñécar, September 24th 2008
R. Roa, L. Gómez-Robledo, R. Huertas, M. Melgosa
Workshop on “Colorimetry and Color Imaging”
Departamento de Óptica, Facultad de Ciencias, Universidad de Granada.18071 Granada (SPAIN). email: [email protected]
Introduction
Method
Results and discussion
Conclusions
References
2
Contents Are the light sources of light cabinets…?
Light source
Object
ObserverIn a first approach, color specification depends on :
Light source Object Observer
X E R x
Y E R y
Z E R z
To standardize color coordinates:
CIEInternational Commission
on Illumination
Standards for the observer: Standard Observer 1931 and Standard Observer 1964
Standards for the light source: illuminants (relative spectral distribution)
3
CIE 15:2004, 3rd Edition, Colorimetry (Technical report), CIE Central Bureau, Vienna (2004).
Introduction Are the light sources of light cabinets…?
CIE proposes the following illuminants:
A (tungsten), C (average daylight), D50, D55, D65 y D75 (daylight simulators)
Typical fluorescents lamps (FL1 to FL12)
New fluorescents lamps (FL3.1 a FL3.15) and high pressure discharge lamps (HP1 to HP5)
On visual experiments:
Cabinets with source simulators which represents properly the considered illuminant.
Daylight Tungsten Fluorescent
4
CIE 15:2004, 3rd Edition, Colorimetry (Technical report), CIE Central Bureau, Vienna (2004).
Introduction Are the light sources of light cabinets…?
From previous works:
There is an important color variability in commercial light cabinets
There are relevant differences between light sources and illuminants
R. Huertas, M. J. Rivas, M. Melgosa, M. Sánchez-Marañón, S. Bhosle, J. J. Damelincourt. “Uniformirty of lighting in color assessment cabinets”, 19th Congress of the International Commission for Optics (ICO XIX), SPIE Proceedings, Volume 4829, 879-880, Firenze, Italy (2002).
R. Roa, R. Huertas, M. A. López-Álvarez, L. Gómez-Robledo, M. Melgosa. “Comparación entre iluminantes y fuentes simuladoras”, Opt. Pura Apl. 41 (3) 291-300 (2008).
5
Objective:Study the temporal evolution of some light sources of commercial light cabinets
Introduction Are the light sources of light cabinets…?
It is important to know the state of the light we are using (temporarily dependent)R. Roa, L. Gómez-Robledo, R. Huertas, M. Melgosa. “Temporal evolution of light sources available in commercial
color cabinets”, 4th Balkan Conference on Lighting - Balkan Light 2008, Ljubljana, Slovenia (2008) (in press)
EXPERIMENTAL SETUP
2 light cabinetsGretagMacbeth Spectralight III
VeriVide CAC 60
Photo Research SpectraScan PR-704 spectroradiometer
PTFE reference white
3 sources studied in each light cabinet
D65 (illuminant D65, daylight, 6500K)
A or F (illuminant A, tungsten, 2856K)
TL84 (illuminant F11, fluorescent, 4000K)
We measured the spectral radiance of the reference white at the following hours of use of the light sources:
0, 0.5, 1, 2, 3, 4, 5, 10, 20, 30, 40, 50, 75 and 100 hours
Total: 14 hours x 3 light sources x 2 light booths = 84 measurements
6
Method Are the light sources of light cabinets…?
WAYS OF COMPARISON
Normalized spectral distributions of the light sources with their respective illuminants.
Metrics for comparison of spectral values:
RMSE (Root Mean Square Error) 2
1
( ) '( )n
i ii
RMSE S S
WRMSE (Weighted Root Mean Square Error) 2
1
( ) ( ) '( )n
i ii
WRMSE R i S S
SCI (Spectral Comparison Index)1
( ) ( )n
i ii
SCI w
CSCM (Colorimetric and Spectral Combined Metric)
*1 1000 1 (%)abCSCM Ln GFC E IRE
7
F. H. Imai, M. R. Rosen, R. S. Berns, “Comparative study of metrics for spectral match quality”, Proceedings of the First European Conference on Colour in Graphics, Imaging and Vision, 492-496 (2002).
J. A. S. Viggiano, “A perception-referenced method for comparison of radiance ratio spectra and its application as an index of metamerism”, Proceedings of AIC Color 01, 701-704 (2004).
M. A. López-Álvarez, J. Hernández-Andrés, J. Romero, R. L. Lee Jr., “Designing a practical system for spectral imaging of skylight”, Appl. Opt. 44, 5688-5695 (2005).
Method Are the light sources of light cabinets…?
Correlated color temperature (CCT)
CIE Color Rendering Index (CRI)
8
CIEDE2000 color differences
CIEDE2000 color differences between GretagMacbeth ColorChecker chips illuminated by the source simulator or its respective illuminant
1/222 2' ' ' ' '
00 TL L C C H H C C H H
L C H C HE R
k S k S k S k S k S
J. Hernández-Andrés, R. L. Lee Jr., J. Romero, “Calculating correlated color temperatures across the entire gamut of daylight and skylight chromaticities”, Appl. Opt. 38, 5703-5709 (1999).
CIE Publication 13.3-1995, Method of Measuring and Specifying Colour Rendering Properties of Light Sources (Technical report), CIE Central Bureau, Vienna (1995).
CIE Publication 142-2001, Improvement to Industrial Colour-Difference Evaluation (Technical report), CIE Central Bureau, Vienna (2001).
Method Are the light sources of light cabinets…?
SPECTRAL DISTRIBUTIONS
There are some differences between light sources and their respective illuminants.
Sources are stabilized after half an hour of use, except sources simulating the A illuminant (approximately stable from the first measurement).
9
R. Roa, R. Huertas, M. A. López-Álvarez, L. Gómez-Robledo, M. Melgosa. “Comparación entre iluminantes y fuentes simuladoras”, Opt. Pura Apl. 41 (3) 291-300 (2008).
Spectralight III D65 Spectralight III A
Spectralight III TL84
CAC 60 D65CAC 60 F
CAC 60 TL84
Results and discussion Are the light sources of light cabinets…?
METRICS
RMSE WRMSE
Cabinet Source Average Std. Dev. (CV) Max Min Average Std. Dev. (CV) Max Min
Spectralight III
D65 0.1315 0.0029 (2.2%) 0.1364 (10h) 0.1210 (100h) 0.0121 0.0005 (4.1%) 0.0135 (0h) 0.0110 (100h)
A 0.0419 0.0012 (2.9%) 0.0440 (75h) 0.0397 (2h) 0.0040 0.0001 (2.8%) 0.0042 (75h) 0.0037 (2h)
TL84 0.1712 0.0103 (6.0%) 0.2042 (0h) 0.1642 (0.5h) 0.0306 0.0005 (1.8%) 0.0314 (100h) 0.0296 (0h)
CAC 60
D65 0.3796 0.0087 (2.3%) 0.4078 (0h) 0.3738 (1h) 0.0450 0.0030 (6.6%) 0.0549 (0h) 0.0434 (1h)
F 0.0841 0.0037 (4.4%) 0.0909 (100h) 0.0768 (0h) 0.0113 0.0005 (4.8%) 0.0122 (100h) 0.0102 (0h)
TL84 0.1843 0.0071 (3.8%) 0.1905 (100h) 0.1615 (0h) 0.0325 0.0016 (4.8%) 0.0338 (100h) 0.0275 (0h)
Values in green are those with the lowest coefficient of variation. The red ones have the largest coefficient of variation.
10
Results and discussion Are the light sources of light cabinets…?
Values slightly rise over the time except for the source D65 from Spectralight III cabinet.
The results obtained with the 4 metrics are not the same.
11
SCI CSCM
Cabinet Source Average Std. Dev. (CV) Max Min Average Std. Dev. (CV) Max Min
Spectralight III
D65 0.0473 0.0016 (3.3%) 0.0519 (0h) 0.0439 (100h) 8.81 0.51 (5.8%) 9.78 (0h) 7.36 (100h)
A 0.0095 0.0003 (3.1%) 0.0101 (75h) 0.0090 (2h) 9.34 0.40 (4.3%) 9.88 (2h) 8.51 (0h)
TL84 0.1297 0.0039 (3.0%) 0.1349 (100h) 0.1220 (1h) 24.84 1.90 (7.7%) 26.16 (5h) 18.84 (0h)
CAC 60
D65 0.1795 0.0114 (6.4%) 0.2169 (0h) 0.1728 (1h) 21.16 0.82 (3.9%) 22.42 (100h) 19.40 (0h)
F 0.0262 0.0012 (4.5%) 0.0283 (100h) 0.0239 (0h) 30.82 1.01 (3.3%) 32.65 (100h) 28.88 (0h)
TL84 0.1373 0.0051 (3.7%) 0.1434 (100h) 0.1210 (0h) 24.17 2.79 (11.6%) 26.02 (100h) 15.18 (0h)
Results and discussion Are the light sources of light cabinets…?
CORRELATED COLOR TEMPERATURE
CCT (K)
Cabinet Source Average Std. Dev. (CV) Max Min
Spectralight III
D65 6129 (6500) 36 (0.6%) 6263 (100h) 6084 (2h)
A 2823 (2856) 6 (0.2%) 2833 (0h) 2814 (2h)
TL84 3698 (4000) 30 (0.8%) 3761 (0.5h) 3663 (100h)
CAC 60
D65 6117 (6500) 133 (2.2%) 6539 (0h) 5992 (100h)
F 2549 (2856) 13 (0.5%) 2574 (0h) 2527 (100h)
TL84 3747 (4000) 47 (1.2%) 3883 (0h) 3705 (100h) 12
Source A from Spectralight III
Lowest variation over time
Source D65 from Spectralight III
Shows an increase over time
Source D65 from CAC 60
Largest decrease over time
Results and discussion Are the light sources of light cabinets…?
COLOR RENDERING INDEX
CRI
Cabinet Source Average Std. Dev. (CV) Max Min
Spectralight III
D65 95 0.44 (0.5%) 95 (20h) 94 (0h)
A 98 0.05 (0.1%) 98 (2h) 97 (0h)
TL84 97 0.27 (0.3%) 98 (0h) 96 (1h)
CAC 60
D65 95 0.44 (0.5%) 96 (0h) 94 (100h)
F 91 0.44 (0.5%) 92 (0h) 91 (100h)
TL84 97 0.09 (0.1%) 97 (0h) 97 (75h)
Sources A from Spectralight III and TL84 from CAC 60
Variation on CRI
Source F from CAC 60
13
Lowest variation over time
Suffers (insignificant) oscillations while evolves
Almost insignificant in all the light sources
Results and discussion Are the light sources of light cabinets…?
CIEDE2000 COLOR DIFFERENCES
14
Results and discussion Are the light sources of light cabinets…?
Chips and standard deviations in CIEDE2000 units
Cabinet Source Average Max Min
Spectralight III
D65 0.055 0.158 (13) 0.003 (24)
A 0.012 0.030 (15) 0.000 (23)
TL84 0.056 0.192 (13) 0.001 (24)
CAC 60
D65 0.091 0.409 (13) 0.002 (20)
F 0.039 0.118 (15) 0.002 (24)
TL84 0.120 0.437 (13) 0.000 (24)
Bluish chips 8 and 13
Source A from Spectralight III
Achromatic samples (chips 19 to 24)
15
largest standard deviations in color differences
smallest variations when the sources age
Source TL84 from CAC 60
smallest color-difference variations over time
greatest standards deviations
1 2 3 4 5 6
7 8 9 10 11 12
13 14 15 16 17 18
19 20 21 22 23 24
Results and discussion Are the light sources of light cabinets…?
Source A from Spectralight III cabinet is the source which suffers the lowest variations over time, attending to the most of indices.
When high precision color measurements (e.g. some threshold visual experiments) are needed, it should be necessary to know the state of the light source we are using, because it is temporally dependent, and it cannot be assumed that it is identical to a CIE illuminant.
To complete this analysis, it would be interesting to study the evolution of the light sources at more hours of use, increase the number of light sources and use other indices for the comparison.
Future work:
Source D65 from Spectralight III cabinet gets closer to its illuminant with age. On the other side, source D65 from CAC 60 cabinet gets older faster than others.
16
Conclusions Are the light sources of light cabinets…?
This work is related to current TC1-44 “Practical Sources for Daylight Colorimetry”.
Final conclusion:
Thanks for your attention
Acknowledgements:Research Project FIS2007-64266, Ministerio de Educación y Ciencia (Spain), with FEDER support.
Vicerrectorado de Política Científica e Investigación de la Universidad de Granada, for the research grant to Rafael Roa.
Contact: [email protected]