Basic HTML Version
NORDIC LIGHT & COLOUR
115
Significant differences between the luminance pictures of the
striped models can be observed. Obviously, the Yellow/Blue
room is the brightest one according to the luminance pattern
of the false-colour picture as well as the luminance values
(44.7/46.9 cd/m
2
on the window wall and 61.7/65.4 cd/m
2
on the
side walls). There is also a noticeable striped pattern on the
walls here, see Figure 23. In the Red/Green room the pattern
is less visible and luminance values are lower according to the
graph (37.1/38.6 cd/m
2
on the window wall and 53.8/57.6 cd/m
2
on the side walls). The darkest room is Grey/Blue, where the
pattern is almost invisible and the luminance values are even
lower (34/35.4 cd/m
2
on the window wall and 49.3/50.8 cd/m
2
on
the side walls). This is consistent with the the test results from
the experiment respondents.
Results for the one-coloured models are not so evident, Figure
25. According to the luminance patterns on the false-colour
pictures, two rooms have quite dark frontal window walls – Yel-
low (37.2/38.2 cd/m
2
) and Grey (35.4/37.7 cd/m
2
). At the same
time, the side walls in the Yellow room are brighter (55.7/57.9
cd/m
2
) than in the Grey room (51.5/54.9 cd/m
2
). Therefore the
Grey room can be considered the darkest model. This is consis-
tent with the subject`s answers. The brightest room, according
to the luminance patterns and values of false-colour pictures,
is the Red model (40.7/41.4 cd/m
2
on the window wall and
65.1/63.5 cd/m
2
on the side walls). According to these results,
the Red room, scored by 25% of the subjects, was the brightest
room.This was equivalent to the Yellow room (25%).
DISCUSSION
This paper investigates the dependence of the perceived light
level of space on unsaturated equiluminant colours and its
combinations (colour contrast). The capability of luminance
maps (as a possible method for lighting design) to reproduce
this information is also a subject of interest.
During the above described experiment eight scale models of
the rooms were studied. All the scale models were identical
except for the hue of the colours of the walls and a slight dif-
ference in saturation of these colours (see Table 1). Geometry,
openings, types of surfaces, luminance and reflectance were
equal for all the models. Nevertheless, some differences in
light level perception of the scale models were found.
Consequently, for the models painted in a striped pattern, the
observed and subsequently statistically calculated difference
was considerably significant (Fr = 12.9677, critical value = 5.99,
a = 0.05, df = 2). Here, the room painted in a striped Yellow/Blue
pattern was chosen as the brightest one, the room painted in a
striped Red/Green pattern was chosen as the medium-bright
room, and the room painted in a striped Blue/Grey pattern
was ascertained as the darkest model. This outcome can be
partly explained by chosen colour composition of the patterns.
Yellow/Blue and Red/Green patterns are examples of strong
colour contrast, while Blue/Grey is a composition of achroma-
tic grey colour and poorly saturated blue colour. The derived
NCS chromaticness is 10. Chromaticness of Yellow/Blue is
20/10 and Red/Green is 15/20. Chromaticness, as one of the
variables of the colour in the NCS system, defines the portion
of the chroma relative to white and black components of the
chosen colour (Arnkil et al. 2012). It means that Yellow/Blue
high chromaticness makes the hue difference more visible. It
leads to a higher hue contrast that can also be a factor which
affects perception. It should be noted that illuminance mea-
sured inside the Yellow/Blue model (see Figure 23) was the
lowest among the striped models and therefore hardly had an
impact on the subjective light level perception. Meanwhile, it
was expected that the Red/Green room would be chosen as the
brightest room but the final results were different. The reason
that the Yellow/Blue model rather than the Red/Green model
was scored as the room with highest light level could be the
highest contrast in chromaticness of blue and yellow colour. At
the same time, the size of the stripes of the patterns can play a
great role in brightness evaluation due to fact that at high spa-
tial frequencies the luminance contrast is prevailing (Valberg
2005). Luminance maps of the striped models corresponded
with measured results (see Figure 23).
However, for the one-coloured models, the situation is more
complicated. The difference in the perception of the light level
was observed indeed, but it was not statistically significant.
Participants also commented verbally on the difficulties in
ranking one-coloured scale models. Luminance maps obser-
vation of these models do not contradict the survey results.
However, it was problematic for the subjects to point out the
most bright or least bright rooms due to the minimal differen-
ces in luminance values. . Perhaps, in this case, the difference
in stimuli was too low, almost at the threshold. According to
luminance values obtained from luminance maps (see Figure
25), the Red room has the highest luminance values on the side
walls (63.5/65.1 cd/m
2
) and this room was scored by 25% of
the subjects as the brightest room, which is equivalent to the
Yellow room. Meanwhile, the Yellow room has lower luminance
values – 55.7/57.9 cd/m
2
on the side walls and 37.2/38.2 cd/m
2
on the window wall. Even the Blue room has higher luminance
values – 55/58.7 cd/m
2
on the side walls and 40.4/43.3 cd/m
2
on the window wall, but it was scored as a relatively dark room.
It can be explained for two reasons. First, there is a possible
error of luminance values in representation of the cool hue of
blue colour (Anaokar & Moeck 2005) For low saturated colours