Professional Documents
Culture Documents
Museum National dHistoire Naturelle, Centre de Recherche sur la Conservation des Collections,
36 rue Geoffroy Saint-Hilaire, 75005 Paris, France; bSociete Nationale des Chemins de fer
Francais, Direction de lInnovation et de la Recherche, Unite Confort et Services,
45 rue de Londres, 75379 Paris Cedex 8, France
1. Background
In 1941, Kruithof [1] conducted a study on the colour
of lamps for general illumination, when fluorescent
lighting had just emerged. A fluorescent lamp is a
low pressure gas-discharge emitter, the glass of which is
lined with a phosphor that absorbs ultraviolet radiations and emits light in the visible range. By modifying
the composition of the fluorescent material, one can
adjust the spectrum of the emitted light. By mixing
different luminescent substances in any desired proportion, Kruithof was able to control the colour of the
emitted light. Indeed, he used three phosphors respectively emitting in the blue, green and red domain of
the visible spectrum. Precisely, he could obtain white
lights with chromaticity close to the Planckian locus,
at correlated colour temperature above 2500 K.
Kruithof introduced a variable number of lamps in a
room, so that the observer could not guess which type
of illumination was on. All lights could be graded in
intensity while controlling the correlated colour temperature (CCT). Beneath 2850 K, he would use incandescent lamps with variable current to produce
light sources. For higher CCT, the experiment was
conducted with daylight originating from a window
or daylight luminescence lamps. There is little
1434
F. Vienot et al.
Figure 1. Reproduction of Kruithofs diagram published in 1941 [1]. Correlated colour temperature (CCT) is scaled in inverse of
megaKelvins along the abscissa. Illuminance is reported on a log scale along the ordinate axis. Two curves show the limits
between which an illumination is considered pleasing after Kruithof. The squares show the nine illumination conditions of our
experiment. Reprinted with permission from Kruithof, Philips Technical Review, 1941, 6, 6596. 1941, Koninklijke Philips
Electronics N.V.
1435
2. Materials
We built a light booth which could house several LED
clusters and provide controlled white illumination at
various correlated colour temperatures and various
illuminance levels. Then we selected several visual tests
in order to investigate performance, colour appearance, cognitive responses and subjective feelings of
observers.
1436
F. Vienot et al.
Figure 2. Relative spectral power distribution of every LED type used in the experiments. Ordinate in arbitrary units. (The
colour version of this figure is included in the online version of the journal.)
1437
Figure 3. Spectral power distribution of the light for the nine illumination conditions. Measurements were taken with a Minolta
CS-1000 spectroradiometer aimed at a white spectralon tile placed on the floor of the light booth. Precise correlated colour
temperature and luminance values are given in Table 1. (The colour version of this figure is included in the online version of the
journal.)
Table 1. Photometric parameters of the nine illuminations used in the experiment. Measurements are given in terms of
luminance expressed in cd m2. They correspond to about 150, 300 and 600 lx.
2700 K
2700 K
2700 K
4000 K
4000 K
4000 K
6500 K
6500 K
6500 K
150 lx
300 lx
600 lx
150 lx
300 lx
600 lx
150 lx
300 lx
600 lx
Correlated
colour
temperature (K)
Luminance
(cd m2)
Colour
rendering index
Ra
Average total
colour rendering
index Ra14
2711.2
2703.5
2699.3
4019.5
4014.0
3997.4
6489.7
6503.7
6491.7
50.4
99.6
198.0
50.7
98.9
196.5
48.7
105.2
192.2
91.7
92.0
93.2
94.3
93.6
96.1
90.7
92.0
91.0
88.2
90.9
91.5
92.5
92.2
93.3
86.6
89.1
90.6
1438
F. Vienot et al.
with
1439
combination
was
3. Results
For every illumination we could average the results of
20 observers.
The aim was to assess whether the visual response
varies with illuminance and/or with correlated colour
temperature. Nevertheless, whichever family of tests
performance, or subjective judgements there might be
systematic variations between observers. Therefore, we
performed a multivariate ANOVA on the results with
three factors: illuminance, CCT and observers.
We used the anovan command from MatLab which
can accommodate missing elements. A 30 3 3
matrix was prepared with individual results, in which
NaN elements were entered when the ith observer had
not been screened under the jth CCT and kth
illuminance lighting combination.
Detailed information about means and standard
deviations of the results and about the statistical
analysis are given in Tables 2 to 5.
3.1. Examples of results
As an example, we present in Figure 4 the results for
low contrast acuity and high contrast acuity. Low
contrast acuity significantly depends upon the illuminance ( p50.0001) but not upon the CCT ( p40.05)
(Table 3). High contrast acuity significantly depends
upon the illuminance ( p50.001) and upon the
CCT ( p50.05). Further, we could investigate which
CCT had a significant effect on high contrast acuity.
Indeed, we checked that high contrast acuity improves
between 2700 and 6500 K ( p50.05), but not between
2700 and 4000 K ( p40.05) or between 4000 and
6500 K ( p40.05) (Table 6).
Table 2. Results of the performance tests. Mean and standard deviation of 20 observers.
Low contrast
acuity
High contrast
acuity
Visual performance
with Landolt rings
Reading speed
Test
condition
2700 K
150 lx
4000 K
150 lx
6500 K
150 lx
2700 K
300 lx
4000 K
300 lx
6500 K
300 lx
2700 K
600 lx
4000 K
600 lx
6500 K
600 lx
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
0.1660
0.0724
0.1805
0.0764
0.1475
0.0767
0.1547
0.0717
0.1305
0.0727
0.1325
0.0647
0.1105
0.0620
0.0930
0.0440
0.1095
0.0707
0.3189
0.0926
41.20
5.51
0.3380
0.0699
41.70
3.69
0.3184
0.0754
42.11
3.65
0.3098
0.0664
42.65
5.41
0.3163
0.0695
42.05
6.82
0.3302
0.0881
41.00
6.76
0.3124
0.0897
41.85
4.22
0.3104
0.0968
41.10
5.57
1440
F. Vienot et al.
Table 3. Results of the analysis of variance for performance tests. F-values and significance. Degree of
freedom for correlated colour temperature, illuminance and participant is, respectively, 2, 2 and 29.
F
p
F
p
F
p
F
p
CCT
Illuminance
Participant
1.76
0.1761
3.28
0.0404
0.81
0.4466
0.38
0.6857
33.4
0.0000
8.82
0.0002
1.05
0.3522
1.84
0.1629
11.78
0.0000
11.71
0.0000
7.78
0.0000
31.61
0.0000
Bright
Clear
Glaring
Pleasant
Comfortable
Relaxing
Warm
Cheerful
Colour
rendering
Test
condition
2700 K
150 lx
4000 K
150 lx
6500 K
150 lx
2700 K
300 lx
4000 K
300 lx
6500 K
300 lx
2700 K
600 lx
4000 K
600 lx
6500 K
600 lx
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
Mean
Standard
deviation
3.85
1.27
3.50
1.10
5.25
1.52
4.35
1.04
5.20
1.28
5.60
1.10
5.05
1.05
5.95
0.83
6.16
0.69
3.55
1.32
3.35
1.31
5.10
1.55
4.00
1.34
4.50
1.40
5.70
1.42
4.25
1.37
6.05
1.00
6.21
0.63
5.70
1.30
5.05
1.43
4.00
1.81
5.00
1.52
4.60
1.39
4.00
1.38
5.15
1.66
4.35
1.69
3.16
1.74
4.55
1.36
3.60
1.60
3.30
1.42
4.75
1.52
4.85
1.42
4.35
1.73
4.80
1.28
4.45
1.39
3.79
1.51
4.95
1.23
3.45
1.73
3.45
1.43
4.45
1.39
4.85
1.46
4.35
1.63
5.15
1.09
4.15
1.50
3.68
1.57
4.75
1.41
3.60
1.60
3.20
1.44
4.40
1.31
4.60
1.50
3.85
1.39
4.75
1.48
3.60
1.57
3.11
1.29
5.15
1.42
3.85
1.53
2.75
1.29
5.55
1.15
4.25
1.48
3.25
1.45
5.05
1.43
3.75
1.62
3.42
1.26
3.70
1.42
2.85
1.18
3.55
1.54
4.20
1.32
4.00
1.17
4.05
1.47
4.05
1.10
4.65
1.14
4.84
1.01
4.40
1.14
3.70
1.13
4.25
1.65
4.10
1.41
4.40
1.31
4.25
1.48
3.95
110
500
1.56
4.26
1.28
1441
Bright
Clear
Glaring
Pleasant
Comfortable
Relaxing
Warm
Cheerful
Colour rendering
F
p
F
p
F
p
F
p
F
p
F
p
F
p
F
p
F
p
CCT
Illuminance
Participant
20.41
0.0000
28.5300
0.0000
18.2200
0.0000
5.6900
0.0042
7.6700
0.0007
10.7000
0.0000
37.7600
0.0000
1.3700
0.2506
0.380
0.686
29.39
0.0000
21.2000
0.0000
3.5800
0.0302
5.2800
0.0061
2.6500
0.0743
1.8500
0.1610
1.6200
0.2019
15.9100
0.0000
0.790
0.455
1.65
0.0294
1.4400
0.0839
2.1200
0.0020
1.4100
0.0974
1.3300
0.1397
0.8900
0.6324
1.5900
0.0401
2.5100
0.0002
0.960
0.525
F
p
High contrast
F
p
F
p
1442
F. Vienot et al.
Between 4000
and 6500K
F
15.77
p
0.0001
Luminance
Participant
19.24
0.0000
1.53
0.0692
"
"
"
#
#
#
#
0.0000
0.0000
0.0000
0.0042
0.0007
0.0000
0.0000
Effect of
illuminance
"
"
"
50"100
0.0000
0.0000
0.03
0.002
"
0.0000
Effect of
participants
Yes
Yes
Yes
Yes
Effect of
Effect of
illuminance participants
Yes
"
0.0002
Yes
Yes
Yes
1443
macular pigmentation because the ipRGCs are supposed to be evenly distributed over the total retina. The
lens spectral absorbance proposed by Stockman and
Sharpe could apply to our young observers. Finally,
the stimulus that elicits the response of melanopsin is
proportional to the convolution of the spectral power
distribution of the LED illumination and the action
spectrum of ipRGCs. Our calculation implies additivity of the menalopsin phototransduction, a property
that has been assessed in the circadian response in
rodents [39].
To explain our results about brightness, a fair
scrutiny requires us to discuss possible mechanisms to
explain the brightness response other than the melanopsin stimulation such as the S-cones participation
or the rods participation. We calculated the S-cone
stimulus using the fundamental response curve proposed by the CIE [40] and we calculated the rod
response using the V0 () scotopic luminous efficiency
function. Results of the calculation of the S-cones, the
melanopsin and the rods stimuli are displayed in
Figure 6 where every stimulus is normalised to its
average value. The figure clearly shows the increase of
the stimulus with CCT when luminance is constant
whichever photoreceptor is involved. We note the same
tendency as for the subjective scaling of brightness
(Figure 5). Therefore, we propose to consider as a
possible explanation of the relative increase of brightness with CCT an increase of intensity of the stimulus
for the melanopsin, which occurs when luminance
increases as well as when CCT increases at constant
luminance.
1444
F. Vienot et al.
1445
Acknowledgements
We would like to thank Corinne Talotte from the Direction
de la recherche de la Societe Nationale des Chemins de fer
Francais for initiating this study, Alain Bricoune and Frantz
Dennery from LedToLite for providing the LED illumination system, Eric Castet and Francois Vital-Durand for
counselling us on how to choose reading texts, Danie`le
Dubois and Gaelle Delepeau from Laboratoire dAcoustique
Musicale for guidance in preparing the questionnaire, Jean
Le Rohellec for photographs and comments and Kenneth
Knoblauch for comments on the manuscript.
Notes
1. The English version is a translation from French of the
subjective scales proposed to evaluate light quality: Non
eblouissante Eblouissante; Sombre Lumineuse;
Froide Chaude; Rendu des couleurs artificiel
Rendu des couleurs naturel; Crepusculaire Claire;
Triste Gaie; Fatigante Relaxante; Inconfortable
Confortable; Desagreable Agreable.
2. Introducing a reference CCT Tcp,o would yield the model
n
References
[1] Kruithof, A.A. Philips Techn. Rev. 1941, 6, 6596.
[2] Commission Internationale de lEclairage. Method of
Measuring and Specifying Colour Rendering Properties
of Light Sources, (CIE Pub. 13.2-1974), CIE Pub.13.
3-1995.
[3] Tiller, D.K. Qualite de leclairage. http://irc.nrccnrc.gc.ca/pubs/bsi/92-5-print_f.html (accessed 2009).
[4] Fotios, S.A. Lighting Res. Technol. 2001, 3, 163178.
[5] Bodmann, H.W.; Sollner, G.; Voit, E. Bewertung von
Beleuchtungniveaus bei verschiedenen lichtarten.
Proceedings of the CIE 15th Session, Vienna, 1963.
[6] Bodmann, H.W. Trans. Illum. Eng. Soc. 1967, 32, 2240.
[7] Smith, S.W.; Rea, M.S. Relationships between Office
Task Performance and Ratings of Feelings and Task
Evaluations under Different Light Sources and Levels.
Proceedings of the CIE 19th Session, Kyoto, 1979.
[8] Boyce, P.R. Lighting Res. Technol. 1977, 9, 1124.
[9] Flynn, J.E.; Spender, T.J. J. Illum. Eng. Soc. 1979, 6,
167179.
[10] Harrington, R.E. J. Opt. Soc. Am. 1954, 44, 113116.
[11] Berman, S.M. Lighting Res. Technol. 2008, 40, 373376.
[12] Aston, S.M.; Bellchambers, H.E. Lighting Res. Technol.
1969, 1, 259261.
[13] Bellchambers, H.E.; Godby, A.C. Lighting Res. Technol.
1972, 4, 104106.
[14] Kanaya, S.; Hashimoto, K.; Kichize, E. Subjective
Balance Between General Color Rendering Index Color
1446
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
F. Vienot et al.
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]