Vous êtes sur la page 1sur 16

Experimental Modeling of Colour Harmony

Ferenc Szabó,* Peter Bodrogi, János Schanda


Virtual Environment and Imaging Technologies Laboratory, Department of Electrical Engineering and Information Systems,
Faculty of Information Technology, University of Pannonia, Veszprém, Hungary

Received 3 March 2008; revised 15 September 2008; accepted 30 November 2008

Abstract: This study investigates colour harmony in visual Ou9) defined colour harmony as an interrelationship of col-
experiments in order to develop a new quantitative colour ours. The main principles of these studies are ‘‘complemen-
harmony model. On the basis of new experimental results, tary’’ and ‘‘analogous’’ but these concepts are not consistent
colour harmony formulae were developed to predict colour among the studies. Also, the colour wheel10 was often
harmony from the CIECAM02 hue, chroma, and lightness adopted as a tool to define these basic relationships. Other
correlates of the members of two- or three-colour combina- authors (Judd and Wyszecki10) define colour harmony as a
tions. In the experiments, observers were presented two- more universal concept: ‘‘when two or more colours seen in
and three-colour combinations displayed on a well-charac- neighboring areas produce a pleasing effect, they are said
terized CRT monitor in a dark room. Colour harmony was to produce colour harmony.’’ Also, there is no consistency
estimated visually on an 11 category scale from 25 among the principles and the keywords of colour harmony.
(meaning completely disharmonious) to þ5 (meaning com- It is completeness according to Goethe,6 order according to
pletely harmonious), including 0 as the neutral colour har- Nemcsics5 and Chevreul,7 and balance according to Mun-
mony impression. From these results, mathematical models sell.2
of colour harmony were developed. The visual results were A quantitative model for two-colour combinations based
also compared with classical colour harmony theories. Two on the CIELAB colour space was developed by Ou et al.9
supplementary experiments were also carried out: one of We carried out computations11 to investigate and compare
them tested the main principles of colour harmony with real this model with classical colour harmony theories. Results
Munsell colour chips, and another one compared the visual showed that there was a significant difference between the
rating of the new models with existing colour harmony predictions of this model and classical harmony theories.
theories. Ó 2009 Wiley Periodicals, Inc. Col Res Appl, 35, 34 – 49, We also conducted a visual experiment11 to investigate the
2010; Published online 17 November 2009 in Wiley InterScience (www. correlation between Ou’s model and our visual dataset, but
interscience.wiley.com). DOI 10.1002/col.20558 only a weak correlation of r2 ¼ 0.30 was found. One possi-
ble reason of this phenomenon can be the different ethnic
Key words: colour harmony; colour harmony formula; origins of the observers, hence our observers were univer-
colour harmony rendering; colour rendering sity students from Hungary, while in Ou’s model, they were
Chinese observers.
INTRODUCTION
Nowadays, choosing of colours that are harmonious in
Colour theory has long had the goal of predicting or speci- a defined lighting situation is an important task not only
fying those colour combinations that would ‘‘work well to- for artists but also for engineers and for the overwhelming
gether’’ or appear harmonious. Scientists and artists of the majority of end users of colour products. It is important
last centuries (Itten,1 Munsell,2 Ostwald3) and nowadays in fashion design just as in developing programs for reha-
(Nemcsics4,5) developed colour order systems, where they bilitation12 and teaching13. If we want to describe the col-
defined rules to establish harmonic sets of colours. These our quality of light sources with a new harmony rendering
colour harmony studies were based on the orderly arrange- metric, correlating well with visual results, a new model
ment of colours in the colour order system. The second of colour harmony, based on a set of new visual experi-
group of authors (Goethe,6 Chevreul,7 Moon and Spencer,8 ments needs to be developed.
In the following sections, first the experiments per-
formed on a CRT monitor will be described. Based on the
*Correspondence to: Ferenc Szabó (e-mail: szafe@szafeonline.hu). results of this experiment, colour harmony models for the
different colour sample combinations will be developed.
V
C 2009 Wiley Periodicals, Inc. Then the model will be tested on an independent data set,

34 COLOR research and application


using material samples. Finally the new models will be
compared with literature data and existing theories of colour
harmony.

FIRST EXPERIMENT
Stimuli and Procedure
A flat-screen colour CRT monitor with 6500K white
point and 116 cd/m2 peak white luminance was used to
display all stimuli. It had good colour channel independ-
ence hence the additive phosphor matrix characterization
model14 could be used together with a polynomial CRT
tone characteristics model. The monitor was characterized
and calibrated using a well-calibrated Photo Research PR-
705 spectroradiometer. Verification measurements showed
that the accuracy was within 2% for the tri-stimulus values
using this monitor characterization model.
All observations took place in a dark room where the
colour monitor was the only light source. Three series of
observations were carried out within 4 months with nine
observers. They were five female and four male university
students familiar with computers and computer monitors.
Colour vision was tested by the Munsell-Farnsworth 100
Hue Test. The viewing distance was 60 cm, the viewing
area of the monitor was 32.5 cm 3 24.5 cm.
Test colour samples were displayed on a middle gray
background (J ¼ 50). The background covered a rectan-
gular area of 29 cm 3 20 cm. There was a white frame
around the screen boundary outside the background. The FIG. 1. (a, b) Experimental images on the CRT monitor.
size of each colour patch was 4.5 3 4.5 cm2. The colour Examples of two- and three-colour combinations.
patches were spaced next to each other, see Figs. 1(a) and
1(b). The task of the observer was to rate his/her colour
harmony impression on a scale from 25 to þ5, by click- were selected from eight hues (with CIECAM02 hue
ing on the corresponding numbered button in the experi- angles of 0, 45, 90, 135, 180, 225, 270, and 315 degrees)
mental computer program. Five corresponded to best har- and each hue consisted of 8 tones: vivid, pale, dull, dark,
mony and 25 corresponded to worst harmony and 0 cor- light grayish, grayish, dark grayish, and the ‘‘spectral’’
responded to neutral colour harmony impression. After (i.e., most saturated) colours. The five achromatic colours
each scaling, a middle-gray background with a white included the grayscale colours of CIECAM at J ¼ 15, 30,
frame was shown for 2 s between the slides to eliminate 60, 80, and 95. To examine the repeatability of the ob-
the after-image effect. The colour stimuli were selected server responses, the whole experiment was repeated.
from all parts of the CIECAM02 colour appearance
model. Three-colour Harmony Experiments
Three-colour combinations consisted of three square col-
Two-colour Harmony Experiment
our patches presented in a triangle position, having a com-
In the two-colour harmony experiment, there were mon vertex in the middle, see Fig. 1(b). Test patches of
2346 colour stimuli. Each stimulus consisted of two three-colour combinations were arranged similarly to the
squared colour patches presented side by side, in the mid- patches used in the two-colour harmony experiments with
dle of the gray background. These stimuli were generated the three squares having common edges. Two series of
from 69 colours, combining all colours from this set of experiments were carried out in case of three-colour com-
69 [(69 3 68) 7 2 ¼ 2346]. This set also contained 273 binations. One of them investigated monochrome three-
two-colour combinations of monochromatic harmonies colour combinations. The other experiment investigated
(containing colours which share the same hue). The 69 trichromatic three-colour combinations, which consists of
colours were selected systematically from the CIECAM02 three samples of different hues. The monochrome experi-
colour space containing five achromatic and 64 chromatic ment consisted of 3322 monochromatic three-colour com-
test samples. Positions of the colour stimuli in the CIE- binations. Test samples were selected from 12 hues (with
CAM ac–bc diagram can be seen in Fig. 2(a), CIECAM02 CIECAM02 hue angles of 0, 30, 60, 90, 120, 150, 180,
correlates can be seen in Table Ia. The chromatic colours 210, 240, 270, 300, and 330 degrees) and each hue con-

Volume 35, Number 1, February 2010 35


FIG. 2. CIECAM02 ac, bc plot of test samples selected to investigate (a) two colour combinations; (b) monochromatic
three-colour combinations; (c) trichromatic three-colour combinations (having equal chroma and different lightness levels);
(d) tri-chromatic three-colour combinations (having equal lightness and different chroma levels).

sisted of eight tones: vivid, pale, dull, dark, light grayish, reduce the number of stimuli, but have still enough stim-
grayish, dark grayish, and the ‘‘spectral’’ (i.e., most satu- uli for statistical calculations and regression. The number
rated) colours. The five achromatic colours included the of stimuli was reduced to 14,280.
grayscale colours of CIECAM at J ¼ 15, 30, 60, 80, and About 7140 stimuli were generated by using 54 colours
95. Monochromatic three-colour combinations were com- having moderate chroma (C ¼ 20) and three different
posed from these set of samples, selecting three of them lightness levels [J ¼ 30, 50, 80; see Fig. 2(c)] by combin-
which had the same hue. The colour stimuli of the mono- ing these colours within the set. Another 7140 stimuli
chromatic three-colour harmony experiment are shown in were generated by using 54 colours having medium light-
Fig. 2(b), CIECAM02 correlates can be seen in Table Ib. ness (J ¼ 50) and three different values of chroma [C ¼
To examine the repeatability of the observer responses, the 10, 20, 30; see Fig. 2(d)], by combining these colours
whole experiment with monochromatic three-colour com- within the set. Latter 54 colours contained five achromatic
binations was repeated. and 49 chromatic colour samples. The chromatic colours
Before starting three-colour harmony experiments, we were selected from 12 hues (CIECAM02 hue angle: 0,
had the results of the two-colour harmony experiments 30, 60, 90, 120, 150, 180, 210, 240, 270, 300, and 330
and our hypothesis was that colour harmony impression is degrees). The five achromatic colours included the gray-
influenced by differences and sums of chroma, lightness, scale colours of CIECAM J ¼ 15, 30, 60, 80, and 95.
and hue. Using the same sampling of the CIECAM02 Chromatic colours can be divided into ‘‘equal property sec-
appearance model as in case of the two-colour harmony tions’’ by their lightness and chroma values, see Figs. 2(c)
experiments (five levels of lightness, three levels of and (d), CIECAM02 correlates can be seen in Table Ic, d.
chroma, and 12 hues) produces more than 50,000 three- To examine the repeatability of observer responses, the
colour combinations, which is impossible to use in a vis- presentation of 2 3 2000 randomly chosen stimuli from
ual experiment. Some simplification had to be chosen to both sets was repeated once.

36 COLOR research and application


TABLE I. Specification of colour samples used in TABLE I. (Continued)
harmony experiments.
# J C h ac bc
# Tone J C h ac bc
(b) Monochrome three-colour harmony experiment
(a) Two-colour harmony experiment 1 15 0.0 0 0.00 0.00
0 Achromatic 15 0.0 0 0.00 0.00 2 30 0.0 0 0.00 0.00
1 30 0.0 0 0.00 0.00 3 60 0.0 0 0.00 0.00
2 60 0.0 0 0.00 0.00 4 80 0.0 0 0.00 0.00
3 80 0.0 0 0.00 0.00 5 95 0.0 0 0.00 0.00
4 95 0.0 0 0.00 0.00 6 30 10.0 0 10.00 0.00
5 Dark Greyish 27 8.8 0 8.80 0.00 7 30 10.0 30 8.66 5.00
6 29 9.4 45 6.65 6.65 8 30 10.0 60 5.00 8.66
7 31 10.1 90 0.00 10.10 9 30 10.0 90 0.00 10.00
8 33 11.3 135 27.99 7.99 10 30 10.0 120 25.00 8.66
9 27 11.9 180 211.90 0.00 11 30 10.0 150 28.66 5.00
10 29 12.4 225 28.73 28.73 12 30 10.0 180 210.00 0.00
11 31 12.8 270 0.00 212.75 13 30 10.0 210 28.66 25.00
12 33 13.1 315 9.26 29.26 14 30 10.0 240 25.00 28.66
13 Greyish 47 8.8 0 8.80 0.00 15 30 10.0 270 0.00 210.00
14 49 9.4 45 6.65 6.65 16 30 10.0 300 5.00 28.66
15 51 10.1 90 0.00 10.10 17 30 10.0 330 8.66 25.00
16 53 11.3 135 27.99 7.99 18 50 10.0 0 10.00 0.00
17 47 11.9 180 211.90 0.00 19 50 10.0 30 8.66 5.00
18 49 12.4 225 28.73 28.73 20 50 10.0 60 5.00 8.66
19 51 12.8 270 0.00 212.75 21 50 10.0 90 0.00 10.00
20 53 13.1 315 9.26 29.26 22 50 10.0 120 25.00 8.66
21 Light Greyish 77 8.8 0 8.80 0.00 23 50 10.0 150 28.66 5.00
22 79 9.4 45 6.65 6.65 24 50 10.0 180 210.00 0.00
23 81 10.1 90 0.00 10.10 25 50 10.0 210 28.66 25.00
24 83 11.3 135 27.99 7.99 26 50 10.0 240 25.00 28.66
25 77 11.9 180 211.90 0.00 27 50 10.0 270 0.00 210.00
26 79 12.4 225 28.73 28.73 28 50 10.0 300 5.00 28.66
27 81 12.8 270 0.00 212.75 29 50 10.0 330 8.66 25.00
28 83 13.1 315 9.26 29.26 30 80 10.0 0 10.00 0.00
29 Dark 27 17.7 0 17.70 0.00 31 80 10.0 30 8.66 5.00
30 29 18.2 45 12.90 12.90 32 80 10.0 60 5.00 8.66
31 31 18.9 90 0.00 18.90 33 80 10.0 90 0.00 10.00
32 33 19.3 135 213.61 13.61 34 80 10.0 120 25.00 8.66
33 27 20.2 180 220.15 0.00 35 80 10.0 150 28.66 5.00
34 29 21.2 225 214.99 214.99 36 80 10.0 180 210.00 0.00
35 31 22.0 270 0.00 221.95 37 80 10.0 210 28.66 25.00
36 33 22.5 315 15.87 215.87 38 80 10.0 240 25.00 28.66
37 Dull 47 17.7 0 17.70 0.00 39 80 10.0 270 0.00 210.00
38 49 18.2 45 12.90 12.90 40 80 10.0 300 5.00 28.66
39 51 18.9 90 0.00 18.90 41 80 10.0 330 8.66 25.00
40 53 19.3 135 213.61 13.61 42 30 20.0 0 20.00 0.00
41 47 20.2 180 220.15 0.00 43 30 20.0 30 17.32 10.00
42 49 21.2 225 214.99 214.99 44 30 20.0 60 10.00 17.32
43 51 22.0 270 0.00 221.95 45 30 20.0 90 0.00 20.00
44 53 22.5 315 15.87 215.87 46 30 20.0 120 210.00 17.32
45 Pale 77 17.7 0 17.70 0.00 47 30 20.0 150 217.32 10.00
46 79 18.2 45 12.90 12.90 48 30 20.0 180 220.00 0.00
47 81 18.9 90 0.00 18.90 49 30 20.0 210 217.32 210.00
48 83 19.3 135 213.61 13.61 50 30 20.0 240 210.00 217.32
49 77 20.2 180 220.15 0.00 51 30 20.0 270 0.00 220.00
50 79 21.2 225 214.99 214.99 52 30 20.0 300 10.00 217.32
51 81 22.0 270 0.00 221.95 53 30 20.0 330 17.32 210.00
52 83 22.5 315 15.87 215.87 54 50 20.0 0 20.00 0.00
53 Vivid 47 60.8 0 60.78 0.00 55 50 20.0 30 17.32 10.00
54 49 51.0 45 36.06 36.06 56 50 20.0 60 10.00 17.32
55 51 38.0 90 0.00 38.00 57 50 20.0 90 0.00 20.00
56 53 54.3 135 238.38 38.38 58 50 20.0 120 210.00 17.32
57 47 35.2 180 235.24 0.00 59 50 20.0 150 217.32 10.00
58 49 37.5 225 226.49 226.49 60 50 20.0 180 220.00 0.00
59 51 47.9 270 0.00 247.87 61 50 20.0 210 217.32 210.00
60 53 51.1 315 36.15 236.15 62 50 20.0 240 210.00 217.32
61 Most saturated 67 27.1 0 27.12 0.00 63 50 20.0 270 0.00 220.00
62 69 29.3 45 20.75 20.75 64 50 20.0 300 10.00 217.32
63 71 31.2 90 0.00 31.18 65 50 20.0 330 17.32 210.00
64 73 33.3 135 223.56 23.56 66 80 20.0 0 20.00 0.00
65 67 27.1 180 227.12 0.00 67 80 20.0 30 17.32 10.00
66 69 29.3 225 220.75 220.75 68 80 20.0 60 10.00 17.32
67 71 31.2 270 0.00 231.18 69 80 20.0 90 0.00 20.00
68 73 33.3 315 23.56 223.56 70 80 20.0 120 210.00 17.32
71 80 20.0 150 217.32 10.00

Volume 35, Number 1, February 2010 37


TABLE I. (Continued) TABLE I. (Continued)
# J C h ac bc # J C h ac bc

72 80 20.0 180 220.00 0.00 3 51 10.1 60 5.05 8.75


73 80 20.0 210 217.32 210.00 4 53 11.3 90 0.00 11.30
74 80 20.0 240 210.00 217.32 5 47 11.9 120 25.95 10.31
75 80 20.0 270 0.00 220.00 6 49 12.4 150 210.70 6.18
76 80 20.0 300 10.00 217.32 7 51 12.8 180 212.75 0.00
77 80 20.0 330 17.32 210.00 8 53 13.1 210 211.34 26.55
78 70 27.1 0 27.12 0.00 9 47 13.1 240 26.55 211.34
79 70 27.1 30 23.49 13.56 10 49 13.1 270 0.00 213.10
80 70 27.1 60 13.56 23.49 11 51 13.1 300 6.55 211.34
81 70 27.1 90 0.00 27.12 12 53 13.1 330 11.34 26.55
82 70 27.1 120 213.56 23.49 13 47 17.7 0 17.70 0.00
83 70 27.1 150 223.49 13.56 14 49 18.2 30 15.80 9.12
84 70 27.1 180 227.12 0.00 15 51 18.9 60 9.45 16.37
85 70 27.1 210 223.49 213.56 16 53 19.3 90 0.00 19.25
86 70 27.1 240 213.56 223.49 17 47 20.2 120 210.08 17.45
87 70 27.1 270 0.00 227.12 18 49 21.2 150 218.36 10.60
88 70 27.1 300 13.56 223.49 19 51 22.0 180 221.95 0.00
89 70 27.1 330 23.49 213.56 20 53 22.5 210 219.44 211.23
90 50 60.8 0 60.78 0.00 21 53 22.5 240 211.23 219.44
91 50 60.8 30 52.64 30.39 22 53 22.5 270 0.00 222.45
92 50 47.0 60 23.50 40.70 23 53 22.5 300 11.23 219.44
93 50 39.0 90 0.00 39.00 24 53 22.5 330 19.44 211.23
94 50 47.0 120 223.50 40.70 25 47 60.8 0 60.78 0.00
95 50 53.0 150 245.90 26.50 26 49 51.0 30 44.17 25.50
96 50 40.0 180 240.00 0.00 27 51 38.0 60 19.00 32.91
97 50 41.0 210 235.51 220.50 28 53 39.0 90 0.00 39.00
98 50 53.0 240 226.50 245.90 29 47 35.2 120 217.62 30.52
99 50 69.0 270 0.00 269.00 30 49 37.5 150 232.44 18.73
100 50 75.0 300 37.50 264.95 31 51 38.0 180 238.00 0.00
101 50 72.0 330 62.35 236.00 32 53 40.0 210 234.64 220.00
(c) Trichromatic three-colour harmony experiment—Equal chroma 33 53 38.0 240 219.00 232.91
section 34 53 51.1 270 0.00 251.12
1 27 17.7 0 17.70 0.00 35 53 51.1 300 25.56 244.27
2 29 18.2 30 15.80 9.12 36 53 51.1 330 44.27 225.56
3 31 18.8 60 9.39 16.26
4 33 19.3 90 0.00 19.32
5 27 19.9 120 29.93 17.20
6 29 20.4 150 217.67 10.20
RESULTS
7 31 20.9 180 220.94 0.00
8 33 21.5 210 218.60 210.74
9 27 22.0 240 211.01 219.07
The consistency of observer answers was examined by
10 29 22.6 270 0.00 222.56 testing the interobserver variability and the intraobserver
11 31 23.1 300 11.55 220.01 variability of the harmony scores. Interobserver variability
12 33 23.6 330 20.47 211.82
13 47 17.7 0 17.70 0.00
(root mean squared) refers to the correspondence of the
14 49 18.2 30 15.80 9.12 nine observers, and intraobserver variability refers to the
15 51 18.8 60 9.39 16.26 repeatability of each observer’s judgments. The mean
16 53 19.3 90 0.00 19.32
17 47 19.9 120 29.93 17.20
value of interobserver variability for the whole experi-
18 49 20.4 150 217.67 10.20 mental dataset (measured on 2346 two-colour combina-
19 51 20.9 180 220.94 0.00 tions, 3322 monochromatic three-colour combinations,
20 53 21.5 210 218.60 210.74
21 47 22.0 240 211.01 219.07
and 2 3 2000 randomly chosen trichromatic three-colour
22 49 22.6 270 0.00 222.56 combinations) was 2.12, and the mean value of intraob-
23 51 23.1 300 11.55 220.01 server variability was 2.44.
24 53 23.6 330 20.47 211.82
25 77 17.7 0 17.70 0.00
The visual colour harmony results showed normal dis-
26 79 18.2 30 15.80 9.12 tributions. An ANOVA test was carried out to determine
27 81 18.8 60 9.39 16.26 those factors that possibly influence the visual colour har-
28 83 19.3 90 0.00 19.32
29 77 19.9 120 29.93 17.20
mony impression. For two-colour combinations lightness
30 79 20.4 150 217.67 10.20 difference, chroma difference, and hue difference were
31 81 20.9 180 220.94 0.00 identified as the most important factors. Hue itself also
32 83 21.5 210 218.60 210.74
33 77 22.0 240 211.01 219.07
influenced harmony: observers differentiated monochro-
34 79 22.6 270 0.00 222.56 matic colour combinations according to the hue of the
35 81 23.1 300 11.55 220.01 constituent colours.
36 83 23.6 330 20.47 211.82
If the constituent colours have the same hue (i.e.,
(d) Trichromatic three-colour harmony experiment—Equal lightness
section
monochrome colour combinations) then they are more
1 47 8.8 0 8.80 0.00 harmonious than the colour combinations of different
2 49 9.4 30 8.14 4.70 hues. To describe these harmonies more accurately, we

38 COLOR research and application


FIG. 3. Predicting perceived colour harmony for two-colour combinations as a function of (a) CIECAM02 hue [see Eq.
(1)], (b) absolute lightness difference [see Eq. (2)], (c) absolute chroma difference [see Eq. (3)], and (d) lightness sum [see
Eq. (4)]. Each dot represents the average of nine observers 3 two observations.

need separate mathematical models for monochrome har- visual colour harmony scores tend to increase if we con-
monies, see Sections ‘‘Colour harmony model for mono- sider not only one perceptual attribute but also other per-
chrome two-colour combinations’’ and ‘‘Colour harmony ceptual attributes or factors (e.g., the equal hue, nonequal
model for monochrome three-colour combinations.’’ lightness group has higher visual harmony score than the
Results indicated that small lightness differences had nonequal lightness group).
low visual colour harmony scores, see Figs. 3(b) and 6(b). The following models hypothesize that the sums and
As lightness difference increases, the harmony score also differences of the perceptual colour attributes (J, C, h) of
increases. Colour pairs having higher average lightness the constituting colours of the compositions influence the
were proved to be more harmonious than pairs of lower av- impression of colour harmony additively. Our descriptor
erage lightness, see Figs. 3(d) and 6(d). For chroma differ- formulae of harmony are assembled from the lightness
ences, a linear decrease of colour harmony can be observed difference, chroma difference, hue difference, hue, and
if chroma difference increases, see Figs. 3(c) and 6(c). the average lightness of the composition.
From these results, some general rules can be formu- Our aim is the accurate modeling of colour harmony in
lated for colour harmony. By depicting these results in a case of two- and three-colour combinations. To satisfy
colour space, general patterns of colour harmony can be this criteria, visual data were collected in case of the four
observed. Attributes of colour combinations producing models separately, and—instead of averaging—four dif-
harmony or disharmony can be defined. From the ferent (but similar) curves were fitted to visual results to
perceptual attributes of the constituent colours, the extent model the effect of the hues of the constituent colours in
of perceived harmony can be predicted, see the next Sec- case of the monochrome and dichrome two-colour combi-
tion. nations (see Fig. 3) and monochrome and trichrome three-
colour combinations (see Fig. 4). The mathematical dif-
ference between these hue preference curves and the
NEW MODELS OF COLOUR HARMONY
modeled hue effect in9 can be seen immediately. At the
Classical colour harmony theories most often formulate moment, we can present only colour harmony models that
general directives in terms of hue, chroma, and lightness are valid in case of Hungarian observers at the same age
and their differences to make harmonious compositions. (the similar facts can be concluded in,9 which model is
Appendix has a comparative description of the present valid only for young Chinese observers). One kind of hue
visual results and classical colour harmony theories. As preference effect in case of different ethnic groups of
can be seen from Table AI (Appendix), the present mean observers can be studied only during a collaborative work,

Volume 35, Number 1, February 2010 39


lightness differences are in turn less harmonious. The best
fitting formula is defined by Eq. (2).

CHF2M;Jdiff ¼ 2:33  105 jDJj3  0:004jDJj2


þ 0:211jDJj þ 0:246 ð2Þ

In Fig. 3(c), colour harmony scores can be seen as a


function of |DC| (the absolute value of chroma differ-
ence). Maximum colour harmony can be realized if the
chroma difference is at its minimum. As the chroma dif-
ference increases the colour harmony score decreases line-
arly. The best fitting formula is defined by Eq. (3).
CHF2M;Cdiff ¼ 3:87  0:066jDCj (3)

In Fig. 3(d), colour harmony scores can be seen as a


function of the lightness sum (Jsum). Minimal colour har-
mony score can be observed in case of low lightness sum,
maximum is at high lightness sum. The best fitting for-
mula is defined by Eq. (4).
CHF2M;Jsum ¼ 0:0268 Jsum  0:656 (4)

Equations (1)–(4) are further transformed to obtain


CHF2M, the correlate of colour harmony, see Eq. (5).
FIG. 4. Predicting perceived colour harmony for nonmo- CHF2M ¼ 0:283  3:275 CHF2M;Jdiff  0:643 CHF2M;Jsum
nochromatic (‘‘dichromatic’’) two-colour combinations as a 
 2:749 CHF2M;Cdiff  4:773 CHF2M;HP  5:305 ð5Þ
function of (a) ‘‘hue preference’’ [see Eq. (12)], (b) sum of
chroma [see Eq. (10)].
In Eqs. (1)–(5), |DJ| is the absolute value of lightness
difference between the two colours, Jsum is the lightness
and conclusions for hue preference can be made only af- sum of the samples, |DC| is the absolute value of chroma
ter having the experimental results. The comparison of difference, the subscript HP indicates hue preference, and
the experimental results are planned to be published in a h is the hue of the composition. (Note that all correlates
future article, which could include an age/culture depend- of colour appearance are defined in CIECAM02.)
ent or age/culture independent model of colour harmony.
Colour Harmony Model for Dichromatic
Two-Colour Combinations
Colour Harmony Model for Monochromatic
Two-Colour Combinations In case of dichromatic two-colour combinations, hue
difference was also found to be a relevant factor of colour
In case of monochromatic two-colour combinations, the harmony impression. This can be seen in Fig. 4(a) show-
hues of the two constituent colours are equal or almost ing mean colour harmony score as a function of absolute
equal (|h1 2 h2| \ 58), and that is why hue difference is hue difference (|h12h2|).
not a relevant factor affecting colour harmony. To investi- The curve fitted to lightness difference [Fig. 3(b)] shows
gate the dependence of visual colour harmony from hue similar tendencies to monochromatic harmonies. This curve
angle, monochromatic two-colour combinations were is defined by Eq. (7). Perceived harmony as a function of
grouped into same property groups, according to hue lightness difference has a global maximum at the value
angle. In Fig. 3(a), colour harmony rankings are depicted of |DJ| ¼ 60 and dichromatic two-colour combinations of
as a function of the hue angles of the constituent colours extreme lightness difference were found to be less harmoni-
(of equal hue angle: h). As can be seen on Fig. 3(a), the ous. A further lightness-related effect is that those dichro-
hues of some colours produce higher perceived harmony matic two-colour combinations in which both constituent
than others. This phenomenon is defined as hue preference colours have high lightness are perceived to be more harmo-
factor, and the best fitting formula is defined by Eq. (1). nious than those of low lightness. The mean perceived colour
CHF2M;HP ¼ 0:361 sinð1; 511hÞ  2:512 (1) harmony score shows a linear tendency with the Jsum value
of the two constituent colours, see Fig. 3(d) and Eq. (8).
As can be seen from Fig. 3(b), combinations of equal In Fig. 3(c), the mean colour harmony score can be
lightness have the lowest colour harmony scores. Light- seen as a function of |DC|, the absolute value of the
ness difference has a global maximum at about |DJ| ¼ 30 chroma difference between the constituent colours of the
absolute lightness difference and combinations of extreme dichromatic two-colour combination. Perceived colour

40 COLOR research and application


harmony is at its maximum if the chroma difference is
minimal. With increasing chroma difference, the colour
harmony score decreases linearly, see Eq. (9). Figure 4(b)
shows the dependence of colour harmony impression from
chroma sum.
On the basis of Figs. 3(a)–3(d) and 4(a) and (b) the fol-
lowing formulae were constructed to predict the visual
colour harmony for ‘‘dichromatic’’ (or ‘‘nonmonochro-
matic’’) two-colour combinations.

CHF2D ¼ 0:47  ð0:515 CHF2D;Jdiff þ 0:391 CHF2D;Jsum


þ 0:205 CHF2D;Cdiff þ 1:736 CHF2D;Csum
þ 2:187 CHF2D;hdiff þ 5:104 CHF2D;HP Þ  2:283
ð6Þ

where,

CHF2D;Jdiff ¼ 2:5  105 jDJj3 þ 3:103 jDJj2


FIG. 5. Result of the method of resolving three colour
 2:2  102 jDJj þ 0:158 ð7Þ combinations to two-colour combinations. Correlation
between visual colour harmony impression and predictions
CHF2D;Jsum ¼ 0:027 Jsum  0:656 (8) with the help of the optimized two colour harmony formula
[see Eqs. (13), (14)].
CHF2D;Cdiff ¼ 0:053jDCj þ 1:172 (9)

CHF2D;Csum ¼ 0:051 Csum þ 2:36 (10) CHF2Dð12Þ  CHF2Dð23Þ  CHF2Dð31Þ


CHF3T ¼ (13)
3
CHF2D;hdiff ¼ 8  105 ðjh1 h2 jÞ  0:0279jh1  h2 j þ 2:3428
2

CHF2DðnmÞ ¼ 4:766  ð0:089 CHF2D;Jdiff


(11)
1 5
CHF2D;HP ¼ 2
4  10 ðh1 Þ  0:0127 h1 þ 1:4035 þ 0:196 CHF2D;Jsum þ 0:333 CHF2D;Cdiff
2 
þ 4  105 ðh2 Þ2  0:0127 h2 þ 1:4035 ð12Þ þ 1:217 CHF2D;Csum þ 5 CHF2D;hdiff
þ 0:595 CHF2D;HP Þ  2:522 ð14Þ

Colour Harmony Models for Three-Colour Combinations The meaning of the subscript (n-m) in Eq. (14) is num-
If the aim is to model colour harmony not only for bering the constituent two-colour combinations i.e. (1-2),
two-colour combinations but also for three-colour combi- (2-3), and (3-1).
nations or even for four or five constituent colours then As can be seen, only a small correlation coefficient of r2 ¼
the following very exciting question arises. Is it possible 0.49 was achieved by this method (by optimizing the weight-
to use the above mathematical models of two-colour har- ing factors of Eq. (6) to achieve maximum correlation). Fig-
monies to describe three, four, or five-colour combina- ure 5 shows that a new model must be formulated for three-
tions? A straightforward solution is resolving a three-col- colour combinations to predict perceived colour harmony.
our combination into three two-colour combinations. Thus, authors constructed colour harmony models for three-
Before developing new mathematical models for three- colour combinations. The principle was the same as in case of
colour combinations, authors tried to model the present two-colour combinations: descriptor models for monochro-
three-colour combination experimental dataset with two- matic and tri-chromatic three-colour combinations were
colour harmony formulae. Latter two-colour combinations developed.
can be described by the formulae introduced earlier, see
Eqs. (1)–(12). Each three-colour combination used in the
Colour Harmony Model for Monochromatic
visual harmony experiments (see Section ‘‘Three-colour
Three-Colour Combinations
harmony experiment’’ earlier) was divided into three two-
colour combinations and colour harmony prediction was In case of monochromatic three-colour combinations,
made by averaging the values of the two colour harmony the hues of the three constituent colours are equal or
formulae. This method lead to a correlation coefficient of within the following limit: (|h1 2 h2| \ 58, |h1 2 h3| \ 58,
r2 ¼ 0.32. Reweighting the weights in Eq. (6) improved |h2 2 h3| \ 58). The lightness difference can be defined as
the method (r2 ¼ 0.49) and lead to Eqs. (13), (14) and to the sum of absolute lightness differences among any two of
the results depicted in Fig. 5. the constituents:

Volume 35, Number 1, February 2010 41


DJ ¼ j j1  J2 j þ j j1  J3 j þ j j2  J3 j (15) colours, they have to be arranged to Jmin, Jmid, Jmax val-
ues, where Jmin denotes the smallest lightness, Jmid
Lightness sum can be defined as the sum of lightness denotes the medium lightness and Jmax denotes the maxi-
correlates: mum lightness among J1, J2, J3. After that, absolute light-
ness difference can be defined as the sum of lightness dif-
Jsum ¼ J1  J2  J3 (16) ferences in the following way:

The chroma difference can be defined in the same way: DJ ¼ jJmid  Jmin j þ jJmid  Jmax j (25)

The chroma difference can be defined in the same way,


DC ¼ jC1  C2 j þ jC1  C3 j þ jC2  C3 j (17)
similar to the lightness difference:
The above formulae of Eqs. (15)–(17) are used in Eqs. DC ¼ jCmid  Cmin j þ jCmid  Cmax j (26)
(18)–(22) below intended to predict the colour harmony
impression of monochromatic three-colour combinations. The Based on the experimental results, the optimized DJ
‘‘Hue preference effect’’ can be defined in the same way as in equation is shown in Eq. (28) and the correlation with the
the model of monochromatic two-colour combinations; hence, visual data in Fig. 6(b). Similarly, Eq. (30) and Fig. 6(c)
the hue of the three constituent colours is equal or similar. show the interrelationship between absolute chroma dif-
ference and visual harmony score data, and Eq. (29) and
CHF3M ¼ 2:523  ð0:0325 CHF3M;Jdiff Fig. 6(d) for lightness sum.
þ 2:2659 CHF3M;Jsum þ 3:4194 CHF3M;Cdiff The ‘‘Hue preference effect’’ and the ‘‘lightness sum’’
 1:355 CHF3M;Csum þ 0:721 CHF3M;HP Þ þ 1:318 can be defined for each constituent colour in the same
way, as in the model of monochromatic three-colour com-
ð18Þ binations. After that, the ‘‘hue preference effect’’ for the
three colour combinations can be evaluated as a sum of
where:
these three hue preference values [see Eq. (31)].
CHF3M;Jdiff ¼ 5:419  107 DJ 4 þ 1:173  104 DJ 3
CHF3T ¼ 2:187  ð0:967 CHF3T;Jdiff þ 1:9397 CHF3T;Jsum
 8:008  103 DJ 2 þ 1:813  101 DJ þ 1:1 ð19Þ þ 0:4399 CHF3T;Cdiff  0:1153 CHF3T;Chrdiff
CHF3M;Jsum ¼ 0:0121 Jsum þ 0:1222 (20) þ 0:11 CHF3t;HP Þ þ 0:134 ð27Þ

CHF3M;Cdiff ¼ 0:0342 DC þ 3:8324 (21) where:


7 3 5 2
CHF3M;HP ¼ 10 h  9  10 h þ 0:0172 h þ 0:4291
CHF3T;Jdiff ¼ 5  105 DJ 3 þ 0:0042 DJ 2
(22)
 0:0694 DJ  0:4898 ð28Þ
Colour Harmony Model for Trichromatic
CHF3T;Jsum ¼ 0:0121 Jsum  0:1222 (29)
Three-Colour Combinations
In case of trichromatic three-colour combinations, the CHF3T;Cdiff ¼ 6  104 DC2  0:0647 DC þ 0:7711 (30)
hues of the three constituent colours are not equal. To cal-
culate absolute hue difference for the h1, h2, h3 hue corre- CHF3T;HP ¼ ððsinðh1  4:836Þ  0:151  7:378Þ  3:796
lates of the three constituent colours, they have to be þ 3:97Þ þ ððsinðh2  4:836Þ  0:151  7:378Þ
arranged to hmax, hmid, hmax values, where hmin refers to
the smallest hue angle among h1, h2, h3 starting from the  3:796 þ 3:97Þ þ ððsinðh3  4:836Þ  0:151  7:378Þ
ac axis and heading counter-clockwise, hmid denotes the  3:796 þ 3:97Þ ð31Þ
medium hue angle and hmax means the maximum hue CHF3T;hdiff ¼ 1:923 Dh21 Dh22 þ 3:781 Dh21
angle among h1, h2, h3. After that, two hue differences þ 3:781 Dh22 þ 5:793 Dh1 Dh2  7:605 Dh1
can be defined [Eqs. (23), (24)]:
 7:605 Dh2 þ 3:6766 ð32Þ
Dh1 ¼ hmid  hmin (23)
Coefficients of the hue difference matrix [see Fig. 7(a)]
Dh2 ¼ hmax  hmid (24) were determined by three-dimensional regression.
The effect of these two hue difference values on colour
Performance of the New Colour Harmony Models
harmony can be seen with the help of a three dimensional
surface graph [see Fig. 7(a) and Eq. (32)], having the Dh1 The performance of the new colour harmony models
and Dh2 values as independent variables. can be measured by computing the correlation between
To calculate the absolute lightness difference for the the experimental dataset and the predicted colour har-
three J1, J2, J3 lightness correlates of the three constituent mony scores, see Figs. 8(a)–8(d).

42 COLOR research and application


FIG. 6. Predicting perceived colour harmony for three-colour combinations as a function of (a) CIECAM02 hue [see Eq.
(22)], (b) absolute lightness difference [see Eq. (19)], (c) absolute chroma difference [see Eq. (21)], and d) lightness sum
[see Eq. (20)]. Each dot represents the average of nine observers 3 two observations.

As can be seen from Figs. 8(a)–8(d), the present visual


dataset can be modeled with the present colour harmony
formulae. High correlation coefficients (r2 ¼ 0.759 and r2
¼ 0.768) were obtained in case of colour harmony models
for two-colour combinations. In case of three-colour com-
binations, slightly lower but still significant correlations
were found (r2 ¼ 0.671 and r2 ¼ 0.627).

Testing the Colour Harmony Models


Comparing CRT Simulation Results with Real Illumi-
nated Munsell Samples. In the above discussion, only the
correlation of the models was discussed, how well they
describe the visual dataset from which they were derived.
CRT simulation is an easy and comfortable way of doing
colour harmony experiments but it is important to verify
FIG. 7. Predicting perceived colour harmony for tri-chro-
these CRT simulation results with real reflecting test sam- matic three-colour combinations as a function of CIE-
ples drawn from an independent experiment. To do this CAM02 hue difference [see Eq. (32)]. Each dot represents
verification for hue difference, chroma difference, and the average of 9 observers 3 2 observations.
lightness difference, reflective test sample sets were
selected from the Munsell atlas. Four test samples were
selected as primary samples with the following Munsell standard observer. These XYZ values were then trans-
notations: 5PB 7/6, 5YR 7/6, 5RP 7/6, and 5G 7/6. Obser- formed into the CIECAM02 colour appearance space,
vations took place in a lighting booth under a B-category using the white point of the ‘‘B’’ category D65 simulator
D65 simulator, L ¼ 170 cd/m2, with the same observers as white point descriptor and other parameters of viewing
as in the CRT simulation experiments. Reflectance spectra situation was considered with the help of the correspond-
of the selected Munsell chips were also measured and ing CIECAM02 parameters (LA, Yb, F, c, Nc). After calcu-
tristimulus values were calculated using CIE 1931 28 lating CIECAM02 correlates, colour harmony predictions

Volume 35, Number 1, February 2010 43


FIG. 8. (a) Prediction of monochromatic two-colour combinations. (b) Prediction of dichromatic two-colour combinations.
(c) Prediction of monochromatic three-colour combinations. (d) Prediction of tri-chromatic three-colour combinations.

of these two-colour combinations (which were composed was eliminated (at the time of the ranking secondary sam-
from two Munsell chips) were made with the help of our ples for the primary 5PB 7/6 sample, the 5PB 5/6 second-
two-colour harmony formula. Viewing situation was con- ary sample was eliminated). So, observers were not
sidered as defining CIECAM02 parameters. allowed to form monochromatic harmonies.
For hue difference investigation, nine other Munsell The investigation of Munsell Value differences was
chips were selected with the same chroma and medium done using two procedures: The four primary samples
lightness difference as secondary samples, from 10 hues: were selected for investigation, and for each hue the Mun-
5PB 5/6, 5P 5/6, 5RP 5/6, 5R 5/6, 5YR 5/6, 5Y 5/6, 5GY sell Value levels 2/, 4/, 6/, 8/, 10/ were tested. In the first
5/6, 5G 5/6, 5BG 5/6, and 5B 5/6. Observers had two case the hue and chroma of the secondary samples (with
tasks. First, they had to choose for each of the four pri- Munsell value levels 2/, 4/, 6/, 8/, 10/) were the same as
mary samples the other member of the combination from the primary samples, so observers composed monochro-
the secondary set of nine chips to produce the most har- matic two-colour combinations. In the second case the
monious and the most disharmonious two-colour composi- chroma of the test samples was the same, but the hue dif-
tions. After that, they had to rank the other seven sam- ference was chosen to be different but optimal (neigh-
ples, and to order the nine two-colour combinations. boring hues with 368 of hue difference), so observers
Notice that all of the 10 main Munsell hues were between were asked to compose dichromatic two-colour combina-
the choices and therefore, observers could compose tions. Observers had the same task as in the case of the
monochromatic harmonies, too. hue difference tests.
By doing the investigation of hue difference, a second The investigation of chroma difference was carried out
part of observation was also carried out: in case of every by the aid of Munsell chips having equal hue and light-
primary sample, the secondary sample of the same hue ness. Four hues were selected for investigation, and for

44 COLOR research and application


FIG. 10. (a) Investigating value difference. Ranking of col-
our combinations having equal hue and chroma, influence
factor: lightness difference. Scale values show the place of
the colour combination in a scale from 9 (best harmony) to
1 (worst harmony) after averaging the ratings of observers.
Results are shown with 95% confidence intervals.

FIG. 9. Investigating hue difference. Ranking of colour


combinations having equal chroma, different lightness, From the figures mentioned earlier, the same tendencies
influence factor is: hue difference. Scale values show the can be realized in case of two-colour combinations for
place of the colour combination in a scale from 9 (best Munsell chroma, Munsell value and Munsell hue as the
harmony) to 1 (worst harmony) after averaging the ratings
of observers. fitted curves show in Figs. 3(a)–3(c).

each hue the Munsell Chroma levels /2, /4, /6, /8, /10 Comparison of Classical and Modern Theories
were tested. Observers had the same task, as in the case of Colour Harmony
of hue difference and lightness difference tests.
In our opinion, a verbal scale with elements: ‘‘best har- This section investigates and compares colour harmony
mony, second best harmony, third best harmony. . .’’ is theories and models, in order to reveal the theory which
much easier to use for the observer during this task as the produces the most harmonious colour composition for the
original scale from 25 to þ5. The expression ‘‘best har- observer. Beyond classical colour harmony theories,2,4,15
mony’’ was then transformed to maximum harmony score the predictive harmony model of Ou9 and our new
(which was the number of secondary samples used in the colour harmony formula for two-colour combinations are
experiment) and so on the ‘‘worst harmony’’ was trans- compared.
formed to minimum harmony score (þ1). To investigate colour harmony theories, a comparison
of the above mentioned colour harmony models and theo-

RESULTS
As can be seen in Fig. 9, by investigating hue differences,
the ‘‘similar hue’’ property was accentuated. Monochro-
matic harmonies or two-colour combinations with neighbor-
ing hues were found more harmonious than combinations
with large hue angle difference i.e., opposite hues.
Results showed that observers found the two-colour
combinations having different lightness more harmonious.
As can be seen from Fig. 10 if the hue angles are defined
to be the same and the two chroma levels are equal then
observers choose a sample with a medium lightness dif-
ference as the most harmonious. Samples with equal
lightness difference are the least harmonious.
The opposite result can be experienced in case of
chroma differences. If the hue is chosen to be the same, FIG. 11. Investigating chroma difference. Ranking of col-
and lightness difference is chosen to be optimal (three our combinations having equal hue and different lightness,
Munsell Value Difference) then the most preferred two- influence factor is: chroma difference. Scale values show
the place of the colour combination in a scale from 9 (best
colour combination can be arranged by minimal chroma harmony) to 1 (worst harmony) after averaging the ratings
differences [see section Models of Colour Harmony, Figs. of observers. Results are shown with 95% confidence
3(c) and 6(c)]. intervals.

Volume 35, Number 1, February 2010 45


1. Lightness difference must have an optimal value to
achieve high colour harmony scores [see Figs. 3(b) and
6(b)]. Small lightness differences produce lower colour
harmony scores. If the lightness difference increases then
the colour harmony score also increases.
2. In case of chroma differences, highest colour har-
mony scores are produced by colours having small
chroma differences [see Figs. 3(c) and 6(c)]. Colour com-
binations of a high chroma difference produce lower col-
our harmony scores.
3. Similar hues produce high colour harmony scores
(see Figs. 4(a) and 7(a)]. In case of tri-chromatic three-
colour combinations, a two-variable function has to be
used to predict colour harmony as a function of the hue
FIG. 12. Comparison of colour harmony theories for two- differences among the three test samples.
and three-colour harmony. Experimental images on the 4. The sum of lightness also influences colour harmony
CRT monitor. impression—a higher lightness sum resulting in a higher
harmony score.
ries were carried out in a visual experiment on a cali- Considering the earlier results, a mathematical model
brated CRT monitor. Forty colours were selected as of colour harmony was developed. The correlation with
anchor stimuli, and harmonious dichromatic colour com- the present visual results turned out to be satisfactory.
binations were arranged along with them from the Mun-
sell,2 Colouroid4 and RAL Design15 colour systems. The
most harmonious colour for each anchor stimulus was
also selected with the help of the ‘‘CH formula’’9 and our
own colour harmony model. Seven observers were pre-
sented these five types of two-colour combinations. All of
them contained the same colour as anchor colour and the
second colour has been selected based on the different
recommendations of the different theories and models
(see Fig. 12). The order of the presented models was ran-
domly generated. Observers had to assign a rank order
number, starting from 5 (for the most harmonious compo-
sition) to 1 (the least harmonious one). The observation
of the samples was repeated five times.
Results showed that two-colour harmony models based
on visual experiments gave more accurate predictions to
define colour harmony. As can be seen from Fig. 13(a),
there is a significant difference between the mean result of
the new predictive models and classical colour harmony
theories. The difference between the two experimental pre-
dictive models9,16 is not significant. In Fig. 13(b), the num-
ber of occasions can be seen when the models were
selected to be the less harmonious. As can be seen, colour
combinations according to the recently developed harmony
models were selected least harmonious the least.

DISCUSSION AND CONCLUSIONS

In this article, visual colour harmony experiments are


described to develop a new quantitative model. Four new
colour harmony formulae were developed to predict col-
our harmony based on CIECAM02 correlates of the mem-
ber colours of two and three colour compositions. Visual
colour harmony was measured on a 11 category scale FIG. 13. (a) Mean colour harmony scores for different col-
our harmony models. Mean rankings of seven observer’s
from 25 (meaning completely disharmonious) to þ5 five repetitions. Results are visualized with 95% confi-
(meaning completely harmonious). The following funda- dence intervals. (b) Number of occasions when the models
mental colour harmony rules were identified. selected to be the less harmonious.

46 COLOR research and application


Comparing these findings with the previous colour har- The present colour harmony formulae (CHF) proved to
mony formula derived by Ou and Luo similar tendencies be good descriptors of colour harmony in a validation
can be realized except the so-called hue preference func- experiment (see section Comparing CRT simulation
tion, which may differ as a reason of cultural differences results with real illuminated Munsell samples). It was also
between the two groups of observers. We recently agreed proved that the different conditions of colour harmony
to collaborate with constructors of the previous colour (monochromatic and dichromatic two-colour combina-
harmony formula, to compare the differences and similar- tions, as well as monochromatic and tri-chromatic three-
ities in details. As a result of this collaboration, a culture colour combinations) cannot be predicted by one single
dependent (or independent) model of colour harmony will formula. By analyzing the other affecting factors i.e., the
be constructed and published in a future article. number of test samples, age, gender, and cultural origin
The present visual results were also compared with of observers, and by investigating the visual results of dif-
classical colour harmony theories, see Appendix. Several ferent groups of observers, different parameter sets can be
classical colour harmony principles and conditions were defined to derive different sets of colour harmony formu-
investigated to compare classical theories and the present lae (CHF) for the different groups of observers, but this is
visual results. The ‘‘equal hue’’ or the ‘‘equal chroma’’ out of scope of the present article.
properties of test colour samples produced high colour Five theories and models of colour harmony were com-
harmony scores. If both properties are true at the same pared (see section Comparison of classical and modern
time then the colour harmony score will be even higher theories of colour harmony), including three classical col-
than in case of the ‘‘equal chroma’’ or ‘‘equal hue’’ prop- our harmony theories2,4,15 and two recently developed
erty alone. The ‘‘equal lightness’’ property produced mathematical models. This investigation was the first
always lower colour harmony scores, and combinations comparison of the two recently developed colour harmony
with nonequal lightness produced higher colour harmony models. A significant difference can be discovered
scores. The ‘‘complementary hue’’ property produced between the colour harmony impression in case of classi-
lower harmony scores than the ‘‘equal’’ or ‘‘neighbouring cal theories and newly developed models. Newly devel-
hues’’ property. oped models gave a more accurate prediction for colour
A second experiment with real Munsell colour samples combinations. From accurate mathematical modeling of
was also carried out with young and elderly observers to colour harmony, a new method can be developed, to
investigate the main principles of colour harmony, and to describe light source colour quality, with the extent of
discover the difference between the colour harmony distortion of colour harmony impression. This new Har-
impressions of young and elderly observers. By compar- mony Rendering Index16 can supplement the current Col-
ing the present visual results and the formulae of a former our Rendering Index17 of the CIE.
mathematical model of colour harmony,9 similar tenden-
cies were observed for lightness, chroma and hue differ-
ences in case of young observers. The present ‘‘hue pref- ACKNOWLEDGMENTS
erence’’ curves (the definition of ‘‘Hue Preference’’ can
The authors thank to Li-Chen Ou and M. Ronnier Luo for
be seen in Section Colour harmony model for monochro-
the preceding work, upon which ‘‘Modeling colour har-
matic two-colour combinations) were different from a for-
mony in case of two-colour combinations’’ section of this
mer study.9 Authors think that these deviations of the
article is based.
‘‘hue preference’’ curves are responsible for the low cor-
relations between the predictions of this former model,
and the present model. We assume that the ‘‘hue prefer-
ence’’ of the observers of different ethnic origin can be APPENDIX: COMPARING THE PRESENT
described by ‘‘hue preference’’ curves optimized for each VISUAL RESULTS WITH CLASSICAL
ethnic group separately. Then, these curves can be COLOUR HARMONY PRINCIPLES
adapted to different colour harmony models.
By comparing the colour harmony impression of young To find out the main differences between classical colour
and elderly observers, significant differences could be harmony principles and the present visual experiments,
observed. For elderly observers colour combinations with our experimental dataset was compared with the principles
small lightness difference resulted in higher perceived of classical harmony theories. In our visual experiment, all
harmony than for young observers. A significant differ- possible colour combinations were generated to each
ence was also found for the chroma difference functions. anchor sample (See Section Method). This dataset contains
Elderly observers chose a less harmonious sample as the colour combinations which are predicted to be harmonious
anchor sample to be the most harmonious, young observ- according to the different classical colour harmony princi-
ers found that the combination with equal chroma had ples. From our visual dataset, different combinations can
been the most harmonious. The results of young observers be selected and categorized with the help of data filtering.
for chroma and lightness differences are consistent with Visual colour harmony rankings were averaged for each
our experiments carried out on a CRT monitor (see Sec- category separately. These visual averages of two and
tion First Experiment). three colour combinations can be seen in Table AI.

Volume 35, Number 1, February 2010 47


TABLE AI. Mean visual colour harmony scores of categorized visual rankings calculated for the classical
harmony principles.
Mean visual
Description with colour harmony
Nr. Name of principle CIECAM02 attributes score STD N

Two-colour combinations
1/a Equal hue h1 ¼ h2 2.517 1.51 273
1/b Equal hue, equal lightness h1 ¼ h2 20.02 1.276 40
J1 ¼ J2
1/c Equal hue, nonequal lightness h1 ¼ h2 2.952 1.052 233
J1 = J2
1/d Equal hue, equal chroma h1 ¼ h2 3.495 0.575 58
C1 ¼ C2
1/e Equal hue, nonequal chroma h1 ¼ h2 2.253 1.577 215
C1 = C 2
2/a Neighboring hues |h12h2| ¼ 458 1.174 1.782 512
2/b Neighboring hues, equal lightness |h12h2| ¼ 458 0.34 1.628 160
J1 ¼ J2
2/c Neighboring hues, nonequal lightness |h12h2| ¼ 458 1.949 1.41 352
J1 = J2
3 Complementary hue |h12h2| ¼ 1808 20.22 1.51 256
3/a Complementary hue, equal lightness |h12h2| ¼ 1808 20.886 1.55 72
J1 ¼ J2
4 Equal lightness J1 ¼ J2 20.476 1.54 528
5 Nonequal lightness J1 = J2 1.91 1.48 1264
6 Equal lightness distance from medium lightness |J1 2 50| ¼ |J2 250| 2.19 1.31 693
7 Equal chroma C1 ¼ C2 3.26 1.06 568
8 Nonequal chroma C1 = C2 0.595 1.31 1224
Monochromatic three-colour combinations
1 Equal lightness J1 ¼ J2 ¼ J3 1.055 1.05 36
2 Nonequal lightness J1 = J2 = J3 2.23 1.11 2206
3 Equal chroma C1 ¼ C2 ¼ C3 3.305 1.18 34
4 Nonequal chroma C1 = C2 = C3 1.465 1.27 1608
Trichromatic three-colour combinations
1/a Neighboring hues |h12h2| ¼ 308 1.36 1.16 852
|h22h3| ¼ 308
1/b Neighboring hues, equal lightness |h12h2| ¼ 308 1.17 0.99 660
|h22h3| ¼ 308
J1 ¼ J2 ¼ J3
1/c Neighboring hues, nonequal lightness |h12h2| ¼ 308 1.42 1.37 504
|h22h3| ¼ 308
J1 = J2 = J3
1/d Neighboring hues, equal chroma |h12h2| ¼ 308 1.45 0.67 664
|h22h3| ¼ 308
C1 ¼ C2 ¼ C3
1/e Neighboring hues, nonequal chroma |h12h2| ¼ 308 1.37 1.39 514
|h22h3| ¼ 308
C1 = C2 = C3
2 Equal hue differences |h12h2| ¼ 1208 21 1.03 216
|h12h3| ¼ 1208
|h22h3| ¼ 1208
3 Equal lightness J1 ¼ J2 ¼ J3 0.02 1.23 7800
4 Nonequal lightness J1 = J2 = J3 0.08 1.49 1728
5 Equal chroma C1 ¼ C2 ¼ C3 2.14 2.04 7876
6 Nonequal chroma C1 = C2 = C3 20.18 1.32 2164

Hue difference is one of the most important and most produce a strong contrast or tension, because they com-
referred factors in classical colour harmony theories. plement each other when mixed, and they complete the
In case of two-colour combinations, equal hue and colour sense; others thought that the juxtapositions of
complementary hue are mentioned to be the most impor- complementary colours produce harmonious colour inter-
tant principles of creating harmony by many authors.2–8 actions.18 According to the present dataset, there is a sig-
The present visual results showed that observers found nificant contradiction between this principle and the pres-
the combination of two-colours harmonious if the member ent visual results, see Table AI and section New Models
colours of the combination shared the same hue. It can be of Colour Harmony, Figs. 4(a) and 7(a).
also seen that lightness and chroma differences [see Figs. The harmony of neighbouring hues is also frequently
3(b) and 3(c)] also have a relevant effect even if they are used by artist or colour designers.2–4 They tend to pro-
monochromatic harmonies. Some theorists and artists duce a single-hued or a dominant colour experience.
believed that juxtapositions of complementary colours These combinations produced also a pleasing effect in our

48 COLOR research and application


experiment except if there was a small lightness differ- 5. Nemcsics A. Experimental determination of laws of color harmony.
ence or a large chroma difference between the constituent Part 1: Harmony content of different scales with similar hues. Colour
Res Appl 2007;32:477–488.
samples. 6. Goethe JW. Theory of Colours. Translation by C. L. Eastlake (1840)
Lightness difference is also a factor often cited in litera- from the German‘‘Farbenlehre’’ of 1810. Reprinted in 1970, Massa-
ture.1–6 The equal lightness property of colours tends to chusetts: The MIT Press.
lower visual colour harmony scores. According to Mun- 7. Chevreul ME. The Principles of Harmony and Contrast of Colors.
Translation by C. Martel (1854) from the French edition of 1839.
sell’s theory,2 if the lightness of the two colour samples
New York: Van Nostrand Reinhold; 1981.
crosses the middle lightness (Munsell Value ¼ 5) harmony 8. Moon P, Spencer DE. Geometric formulation of classical colour har-
is produced. This type of harmony can also be imagined in mony. J Opt Soc Am 1944;34:46–59.
the CIECAM02 colour appearance model if the two con- 9. Li-Chen OU, Ming Ronnier LUO. A study of colour harmony for
stituent colours are in equal distance from the J ¼ 50 light- two-colour combinations. Color Res Appl 2006;31:191–204.
10. Judd DB, Wyszecki G. Colour in Business. Science and Industry,
ness level. According to the present visual results, in this
3rd edition. New York: Wiley; 1975.
case, good colour harmony impression was observed; see 11. Szabó F, Bodrogi P, Schanda J. Comparison of colour harmony
Table AI and Section Models of Colour harmony. models: Visual experiment with reflecting samples simulated on a
Chroma difference is also an important factor2–8 - equal colour CRT monitor. In: Proceedings of the 3rd European Confer-
chroma or similar chroma always generates high harmony ence on Colour in Graphics, Imaging, and vision (CGIV), 2006.
12. Sik Lányi C. Multimedia Medical Informatics System in Healthcare.
scores [see Figs. 3(c) and 6(c)]. By increasing chroma dif- In: Ichalkaranje N, Ichalkaranje A, Jain LC, editors. Intelligent Para-
ference, the present visual dataset also showed a decrease digms for Assistive and Preventive Healthcare. Berlin, Heidelberg:
of colour harmony scores. Springer-Verlag; 2006. p 37–68. ISBN 3-540-31762-7.
In case of three-colour combinations, the main harmony 13. Sik Lányi C, Mátrai R, Molnár G, Lányi Zs. User interface design
question of developing multimedia games and education programs
principle was formulated to be the concept of triadic com-
for visual impairment children, Special Issue of Elektrotechnik &
binations.4 A triadic colour scheme adopts any three col- Informationstechnik (e&i) by Springer Wien/New York, ‘‘Technol-
ours approximately equidistant around the hue circle. The ogy Enhanced Learning (TEL)’’, 2005 (12) pp.488–494. ISSN 0932-
present experiments seem to contradict to this princi- 383X.
ple4—these triadic colour schemes are not as harmonious 14. Berns RS, Katoh N. Methods of characterizing displays. In: Green
P, MacDonald LW, editors. Colour Engineering: Achieving Device
as expected in the light of this classic theory.4
Independent Colour. England: Wiley; 2002. p 127–164.
15. Web page of RAL Design System. Available at: http://www.ral.de/en/
ral_farben/farbkarten/ral_design.php. Accessed on September 20, 2009.
16. Szabó F, Bodrogi P, Schanda J. Visual experiments on colour har-
1. Itten J. The Art of Colour. Translation by Ernst van Haagen from
mony: A formula and a rendering index. In: Proceedings of the 26th
German ‘‘Kunst der Farbe’’. New York: Van Nostrand Reinhold;
Session of the CIE, 2007.
1961.
17. CIE Publ 177: 2007 Colour rendering of white LED light sources.
2. Munsell AH. A Grammar of Color, From the original version of
Vienna: CIE Central Bureau; 2007.
1921. New York: Van Nostrand Reinhold; 1969.
18. Sully J. Harmony of colours. Mind 1879;14:172–191.
3. Ostwald W. The Colour Primer. Edition by F. Birren from the German
19. Szabó F, Schanda J, Bodrogi P. Experimental investigation of the
‘‘Die Farbenfibel’’ of 1916. New York: Van Nostrand Reinhold; 1969.
distortion of colour harmony. In: Proceedings of the AIC Interna-
4. Nemcsics A. Farbenlehre und Farbendynamik. Budapest: Akadémiai
tional Conference on Colour Harmony, 2007.
Kiadó; 1993.

Volume 35, Number 1, February 2010 49

Vous aimerez peut-être aussi