Printer Friendly

User preferences in office lighting: a case study comparing Led and T5 lighting.

1 INTRODUCTION

Lighting consumes about 19 percent of all produced electricity worldwide [IEA 2006] and represents about 35 percent of total primary energy consumption within buildings, like offices and schools [Roisin and others, 2006]. Therefore, the energy efficiency of lighting is important, especially when the price of the electricity has risen constantly during recent years [Statistics Finland, 2011; Eurostat, 2012] while at the same time the energy efficiency regulations have become more restrictive [European Commission, 2010].

Although energy efficiency and savings are important, the most important function of lighting is to provide comfortable visual conditions for the users of the space and help them to perform visual tasks as easily as possible. Therefore, besides the quantity of light and how efficiently it is produced, the quality of the light must also be considered.

Villa and Labayrade [2012] conducted a multiobjective optimization study of lighting installations that focused on energy objectives and user preferences. They gathered user observations and fed them into a computer model that calculated trade-offs between user preferences and energy savings. In that study the room was illuminated with two recessed T5 ceiling luminaires and a compact fluorescent task area luminaire placed on the table. The optimal solution between power consumption and user preferences was found to be 436 lx for task area illuminance, with illuminance uniformity of 0.52.

Logadottir, Christoffersen, and Fotios [2011a] investigated preferred illuminances on workplace by using an adjustment task. Fluorescent lamps with three different CCTs were used: 3000 K, 4000 K and 6500 K. They concluded that the preferred illuminance was affected by the range of illuminances available, with larger adjustment range resulting in higher preferred illuminances. Depending on the adjustment range, the preferred illuminances were 337 lx, 523 lx, or 645 lx. The variances of the preferred illuminances between individual subjects were substantial. The starting point of the adjustment, the so called anchor, also affected the results. They concluded that a single interval adjustment task is not an appropriate research method to determine the preferred illuminance.

Logadottir, Christoffersen, and Fotios [2011b] also investigated the use of an adjustment task to set the preferred CCT for ambient illumination using three different adjustment ranges: 2736-3530 K, 3284-4014 K and 2736-4014 K. All of these ranges included the central value of the adjustment scale, which was 3375 K. The effect of the anchor CCT value and the effect of adaptation time were also studied. Results showed that a higher CCT anchor value resulted in a higher preferred CCT, and a larger stimulus range resulted in larger deviation in the user set CCT. With the whole adjustment range the mean preferred CCT was 3553 K for the nonadapted subjects and 3426 K for subjects who had adapted 5 minutes to the anchor CCT before starting the adjustment. The average of all subjects was 3490 K. However the authors concluded that no universal preferred CCT was found in this study because the results varied between anchor points and the adjustment ranges.

Veitch and Newsham [2000] investigated preferred lighting conditions in open-plan offices. They used 3500 K fluorescent lamps and the lighting consisted of dimmable general lighting and nondimmable task lighting. They reported large variations in preferred desktop illuminances, between 83 and 725 lx. Over 60 percent of the tested participants chose illuminance level lower than 500 lx, with mean value 423 lx and median value 413 lx.

Boyce and others [2006] studied occupant use of switching and dimming controls in offices. Office cubicles were lighted with either a dimmable ceiling mounted T5 luminaire or with a switchable desktop CFL luminaire, that had multiple switching levels for different illuminances, and a T5 ceiling luminaire that provided ambient light at constant level. The maximum available desktop illuminance with these two lighting options was about the same, 1061 lx for the dimmable lighting and 1091 lx for the switchable lighting. The mean preferred desktop illuminance was 458 lx for dimmable lighting and 884 lx for switchable lighting. The findings of this study indicate that there were large deviations in the user set illuminance levels and on average users set the mean desktop illuminance higher with switchable lighting than with dimmable lighting. The researchers believed that this was partially due to differences in the lighting distribution in the cubicle with different luminaires.

The most important aim of this study was to find out what kind of office lighting people prefer in their work place. At the moment it seems that new lighting solutions of the near future will be dominated by light emitting diodes (LEDs) and thus it is vital to know their advantages and disadvantages compared to more traditional light sources. Three luminaire types were chosen for this study: rectangular T5 luminaire (Re_T5), square LED luminaire (Sq_LED) and round LED luminaire Ro_LED). Re_T5 lighting was compared to Sq_LED lighting at three different illuminance levels: 300 lx, 600 lx and 1000 lx. Ro_LED lighting had an adjustable color temperature from 2621 to 6321 K and it was tested at three different CCT settings: 3000 K, 4500 K and 6000 K. This was done to find out how the subjects reacted to the change in CCT at otherwise similar lighting. A case study was performed in which the subjective quality of lighting was evaluated through questionnaire. The tested parameters were glare, reading task easiness, detail distinction from the wall, color naturalness and pleasantness, illuminance level sufficiency and pleasantness, general pleasantness of the lighting, naturalness of the shadows and the appearance of the luminaires.

An additional goal of the study was to determine an optimum illuminance level for office work. Users were asked to adjust the lighting to an optimum level and to a minimum level that they would be willing to work in for a limited time. This was done to determine by how much the illuminance level can be reduced temporarily without causing too much discomfort to office workers in cases where the available electrical power is limited. Users adjusted the optimum level for Re_T5 and Ro_LED lighting, whereas minimum lighting was set only for Re_T5. In the LED adjustment task users also set a preferred optimum CCT in addition to illuminance level.

Energy consumption and related saving potential is also briefly discussed for the tested luminaires.

2 METHODOLOGY

2.1 LUMINAIRES AND TEST ROOMS

The tested luminaires will be referred as follows:

* Ro_LED=Round LED downlight luminaire (Fig. 1a), 27 W, CCT 2621-6321 K, dimmable and color adjustable DALI, recessed ceiling luminaire, [empty set] 200 mm x 120 mm. Adjustment range: 0-834 lx at CCT 2621 K; 0-692 lx at CCT 6321 K, measured from the reference point marked in Fig. 2. CIE color rendering index (CRI) = 94 (CCT 3000 K), 93 (CCT 4500 K), 89 (CCT 6000 K).

* Sq_LED=Square LED panel luminaire (Fig. 1b), 44 W, CCT 4000 K, dimmable DALI, recessed ceiling luminaire, 598 mm x 598 mm x 86 mm. Adjustment range: 0-1075 lx, measured from the reference point marked in Fig. 2. CIE CRI=86.

* Re_T5=Rectangular recessed T5 fluorescent lamp luminaire with low luminance louvre (Fig. 1c), 2 x 28 W, CCT 4000 K, dimmable DALI, recessed ceiling luminaire, 1187 mm x 140 mm x 66 mm. Adjustment range: 0 -1142 lx, measured from the reference point marked in Fig. 2. CIE CRI = 82.

The user study was conducted in office rooms built at the Aalto University Otaniemi campus in Espoo, Finland. Two similar rooms were built. To the first room four Re_T5 luminaires (Fig. 1c) were installed and to the second room six Ro_LED luminaires (Fig. 1a) and four Sq_LED luminaires (Fig. 1b) were installed. The distinct dimensions and photometric distributions of the luminaires resulted in different light distributions and this might have affected to the results of this study.

The dimensions of the test rooms were width 3.51 m x length 4.22 m x height 2.80 m. There were three windows facing west, but these windows were covered with blinds during the measurements to prevent daylight from entering into the room. The two test rooms located side by side and were facing an inner court between two 4-story buildings.

The layout of the rooms and installed luminaires is presented in Fig. 2.

The rooms were furnished with office desks, chairs, books, posters and other objects typically found in offices. On the desk in the middle of the task area there was a laptop, a Macbeth color checker chart (Fig. 4a), a coffee mug and a sheet of printed white paper with black text (Arial, 10 pt).

2.2 LIGHTING CONTROL AND MEASUREMENT EQUIPMENT

Each luminaire type was controlled as a group through digital addressable lighting interface (DALI). This interface was connected to a laptop computer that had Helvar Digidim lighting control software installed. There was also additional dimming control switch (Fig. 3b) for each luminaire group and color control switch for Ro_LED luminaires (Fig. 3a). The DALI-interface does not support color control directly, so the color control was applied through proprietary manual switch from the luminaire manufacturer. The color settings could be saved to programmable scenes of the DALI system, along with the luminaire specific dimming levels.

Illuminances were measured with LMT Pocket Lux 2 illuminance meter and luminances were measured with Nikon Coolpix 8400 and Photolux software. CCTs in the room were measured with Konica Minolta CL-500A illuminance spectrophotometer. Illuminance level measurement error during the user test procedure was between 5 and 35 lx, which was caused by the no-load illuminance in the room. This no-load luminance resulted from the light leaks through the window blinds and the small opaque window opening to the corridor. Highest measured no-load illuminance on the task area was 35 lx during a sunny afternoon but most of the time the background illuminance level was below this. The preset settings were programmed in a situation where the no-load illuminance was 5 to 10 lx.

2.3 USER TEST SUBJECTS

The subjects were staff and students of Aalto University and none of them had previous background in lighting. Before starting the tests, subjects were informed that the study was about office lighting and that test method was a questionnaire, but no further information was given. Forty subjects participated in the test. The age and gender distributions are presented in Table 1.

TABLE 1.
The Age and Gender Distribution of the Test Subjects

Age    Number of Subjects/  Number of Males/ %  Number of Females/
        % of All Subjects     of All Subjects     % of All Subjects

20-29             29/72.5%            24/60.0%             5/12.5%
30-39              5/12.5%              3/7.5%              2/5.0%
40-49               3/7.5%              1/2.5%              2/5.0%
50-59               3/7.5%                0/0%              3/7.5%
Total              40/100%              28/70%              12/30%


If the age of subject is assumed to be in the middle of the given age scale (for example age group 20-29 years old -> age is assumed to be 25), the average age of males was 27 years and the average age of females was 38 years. 16 subjects out of 40 (40 percent) wore glasses normally, so they were advised to wear them also during this test.

2.4 USER TEST PROCEDURE

Nine different preset lighting situations were presented to the subject in a random order and the same questionnaire (Appendix 1) was completed during each situation. The preset lighting situations used in this study are listed in Table 2. First the order of the presented luminaires was randomized and after that the order of individual situations regarding each luminaire was also randomized. Test sequence was for example 3,2,1 - 6,4,5 - 7,9,8 for the first subject and 9,7,8 - 1,2,3 - 4,5,6 to the next person and so on. This was done to eliminate the effect of presentation order.

APPENDIX 1 QUESTIONNAIRE FORMS USED IN THE USER TEST

Indoor lighting questionnaire

Age:     under 20  20-29   30-39  40-49  50-59  over 60 year old
         *         *       *      *      *      *

Gender:  * male
         * female
Mark with an 'X' to the scale below next to a number closest
matching your opinion.

                                             9  8  7  6  5  4  3  2  1
A)  Are you experiencing glare  Not at all
    when you are sitting in
    the armchair?

B)  Are you experiencing glare  Not at all
    when you are sitting
    beside the desk?

C)  Read the first chapter of   Really
    the paper on the desk. How  easy
    easy it is to read in this
    lighting?

D)  How easy it is to see       Really
    details in the measurement  easy
    chart (black letters) on
    the wall in this
    lighting?

E)  How natural are the         Completely
    colours looking in this     natural
    lighting? (Colour checker
    chart, posters on the
    wall, skin colour of your
    own hand)

F)  Is the illuminance level    Completely
    sufficient for typical      sufficient
    office working tasks?

G)  Do you think the light      Way too
    amount is pleasant in the   much light
    room? (5= optimal)

H)  Do you think the colour of  Really
    the lighting is pleasant?   pleasant

I)  In general, do you think    Really
    the lighting is pleasant?   pleasant

J)  Do you think the            Really good
    luminaires themselves look  / stylish
    good/stylish?

K)  How natural the shadows     Completely
    look like in this           natural
    lighting? (Look for
    example the coffee mug on
    the table)

A)  Are you experiencing glare  Disturbingly
    when you are sitting in     much
    the armchair?

B)  Are you experiencing glare  Disturbingly
    when you are sitting        much
    beside the desk?

C)  Read the first chapter of   Really hard
    the paper on the desk. How
    easy it is to read in this
    lighting?

D)  How easy it is to see       Really hard
    details in the measurement
    chart (black letters) on
    the wall in this
    lighting?

E)  How natural are the         Completely
    colours looking in this     unnatural
    lighting? (Colour checker
    chart, posters on the
    wall, skin colour of your
    own hand)

F)  Is the illuminance level    Completely
    sufficient for typical      insufficient
    office working tasks?

G)  Do you think the light      Way too dim
    amount is pleasant in the
    room? (5= optimal)

H)  Do you think the colour of  Really
    the lighting is pleasant?   unpleasant

I)  In general, do you think    Really
    the lighting is pleasant?   unpleasant

J)  Do you think the            Really bad /
    luminaires themselves look  ugly
    good/stylish?

K)  How natural the shadows     Completely
    look like in this           unnatural
    lighting? (Look for
    example the coffee mug on
    the table)

TABLE 2.
Preset Lighting Situations. Presentation Order Was Randomized
between Test Subjects

SituationNo.  Luminaire  Nominal Illuminance  CCT
                         at the Calibration   (K)
                             Point (lx)

1             Sq_LED                     600  4000
2             Sq_LED                     300  4000
3             Sq_LED                    1000  4000
4             Re_T5                      600  4000
5             Re_T5                      300  4000
6             Re_T5                     1000  4000
7             Ro_LED                     600  4500
8             Ro_LED                     600  6000
9             Ro_LED                     600  3000


Before entering a lighting situation, the test subject was asked to wait outside the test room in order to adapt to the illuminance level and color temperature of the corridor that served as waiting area. There was a chair in the corridor, in which the subject was asked to sit until the experimenter gave permission to enter the test room. The horizontal illuminance on the eye level (120 cm from the ground) by this chair was 165 lx, CCT 2900 K. Overall average illuminance of the corridor was 154 lx, standard deviation 88 lx. There were no windows in the corridor, so the illuminance conditions in the waiting area remained constant regardless of the outdoor lighting conditions. After 1 minute waiting time, the test instructor opened the room of the test room and subject stepped in.

When entering the test room, the subject was instructed to sit on an armchair in the corner of the room. After 1 minute adjustment time the test supervisor handed a questionnaire (Appendix 1) to the subject who began filling it. After answering the first question regarding glare the subject was instructed to sit by the desk, located in the middle of the task area and fill the rest of the questionnaire. While sitting by the desk, the subject was first asked to evaluate the glare again. Next task was to read a chapter from the newspaper article (black letters, font Arial, size 10 pt, printed on white A4-size paper) and to evaluate how easy it was to read. Subsequently the subject was asked to look at a Lea-chart (size 20.3 cm x 25.4 cm, distance to the viewer 2.7 m, Fig. 4d) on the wall and to evaluate how easy it was to see details on it. The Lea-chart used in this test was developed for near vision tests, but in this study its purpose was to serve as a "random object" containing a neutral white background and clear black details; the evaluation of the detail distinction was done purely on subjective terms by the user, not measured objective terms.

The subject then evaluated the naturalness of colors, based on the MacBeth color checker chart (Fig. 4a) that was placed on the table in front of the subject, posters on the wall, and the skin color of the subject's own hand. Subject also evaluated how pleasant the illuminance level and color of the lighting was and how well it suited to the typical office working tasks, such as reading. Finally the subject was asked to evaluate the general pleasantness of the lighting, how good the luminaires looked and how natural the shadows looked in the room.

After answering all these questions regarding the presented lighting situation, the subject was asked to wait again in the corridor. The test supervisor changed the lighting situation in the room to a new one and asked the subject back after a 1 minute waiting time. This process was repeated until all nine preset situations had been presented to the subject.

On the second part of the user test, the subject was asked to set his or her own preferred illuminance levels. The starting point for illuminance adjustment was 300 lx in the task area, which was set before the subject entered the room. This illuminance level was chosen for the starting point of the adjustment because it was the lowest illuminance level that was used in the preset situations of the user test. Before instructions for the adjustment tasks were given, the subject waited for 1 minute in the room for the eyes to adjust.

First the subject was asked to set a minimum illuminance level that he would be willing to work in temporarily. The task used in this experiment was reading a chapter from a newspaper article (black letters, font Arial, size 10 pt, printed on white A4-size paper). The purpose of this experiment was to find out how much the light level can be reduced without causing too much discomfort to office workers if the electrical power available for lighting is limited. Such situation may occur, for example, if the building operates on reserve power during power failure.

After this the subject was asked to set the illuminance level to an optimum level, meaning a level that subject felt most comfortable to work in. Because subject performed this adjustment immediately after the minimum level adjustment task, the initial illuminance level was the same as it was set to in that previous assignment. These two tests were done by adjusting the lighting in Room 1. The adjustment range of the Re_T5 lighting used in these tests was 0 to 1142 lx.

Before the final part of the test, subject waited again for 1 minute in the corridor and then entered the Room 2. The illuminance level of this room was set to 300 lx, measured from the reference point on the task area, and CCT was set to one the following settings: 3000 K, 4500 K or 6000 K. This was done to find out if the initial CCT had an effect to the user set CCT. There was again one minute adaptation time and then the subject was given instructions for the adjustment task. The goal of this task was the same as previously, to set the lighting in the room to an optimum level for office working tasks. This time the subject set the illuminance level and the CCT to an optimum level in the room, by adjusting the Ro_LED luminaires. Illuminance level adjustment range was 0 to 738 lx and CCT adjustment range was 2621 to 6321 K. After completing this final adjustment task, the subject was thanked for participating to the test and then dismissed.

In total it took about 1 h for each subject to complete the test series.

3 RESULTS

3.1 LUMINAIRE MEASUREMENTS

Power consumption, luminous flux, correlated color temperature, CIE Color Rendering Index (CIE CRI) and luminous efficacy of the luminaires were measured using an integrating sphere that measures the spectral power distribution in the wavelength range of 350 to 1050 nm. These measurements were performed in the laboratory of Aalto University Lighting unit, Espoo, Finland.

The luminous flux at different dimming levels is shown in Fig. 5. Luminous efficacy of all the luminaires was highest at full power (Fig. 6): Sq_LED 88.2 lm/W; Ro_LED 72.5 lm/W at CCT 2621 K to 63.5 lm/W at CCT 6321 K; Re_T5 61.8 lm/W. Therefore, at full power Sq_LED luminaires produced 43 percent more lumens per watt than Re_T5 luminaires and 26 percent more lumens per watt than Ro_LED luminaires. At s 10 percent DALI dimming level the Sq_LED luminaires produced 128 percent more lumens per watt than Re_T5 luminaires and 40 percent more lumens per watt than Ro_LED luminaires. The LED luminaires maintained their luminous efficacy better when dimmed, in comparison to the Re_T5 luminaires.

The Ro_LED luminaires were equipped with a color adjustment option. Their CCT could be adjusted between 2700 and 6500 K according to the luminaire manufacturer. In the performed measurements the actual CCT scale was 2621 to 6321 K. The dependence of luminous efficacy on the CCT setting is presented in Fig. 7.

The luminous efficacy of the Ro_LED luminaire was highest at its lowest CCT setting. At the highest measured CCT (6321 K) the luminous efficiency was 10 percent lower, when compared to the lowest measured CCT (2621 K).

The CIE CRI of the Ro_LED luminaire was higher at the lowest CCT when compared to the higher CCTs (Table 3). At all CCT settings the Ro_LED luminaire had higher CIE CRI than the Sq_LED and Re_T5 luminaires. The measured CIE CRI of the luminaires remained almost constant regardless of the dimming level. Between full power (DALI=100 percent) and 10 percent power (DALI=10 per-cent) CIE CRI changed < 3 percent on all tested luminaires, being somewhat higher at lower dimming levels than at full power. Between the DALI 100 percent and DALI 5 percent settings the CCT of the tested luminaires changed as follows: Sq_LED luminaire < 0.2 percent; Ro_LED luminaire < 0.3 percent; Re_T5 luminaire < 1.2 percent.

TABLE 3.
Measured General CIE Color Rendering Indices Ra at
Full Power (DALI = 100%)

Luminaire       Ra

Re_T5, 4000 K   82
Sq_LED, 4000 K  86
Ro_LED, 3000 K  94
Ro_LED, 4500 K  93
Ro_LED, 6000 K  89


3.2 OFFICE ROOM MEASUREMENTS

Illuminance levels of the office rooms were measured when the luminaires were at full power. Horizontal illuminances were measured in a 60 cm x 60 cm grid, from a total 42 discrete points around the room, 70 cm above the ground. These point measurement results were inserted into Matlab software and plotted in Fig. 8. Table 4 shows statistics of the illuminance measurements. In the case of Ro_LED luminaires, the illuminances were measured at lowest and highest CCTs, 2621 and 6321 K. Illuminance was measured only from the middle of the task area (Fig. 2) during user tests, because light distribution of the luminaires remained constant regardless of the dimming level.

TABLE 4.
Measured Mean, Minimum and Maximum Illuminances
of the Room. [E.sub.m] = Mean Horizontal Illuminance,
[E.sub.min] = Minimum Horizontal Illuminance, [E.sub.max]
= Maximum Horizontal Illuminance

Luminaire Type    [E.sub.m]  [E.sub.mln]  [E.sub.max]
                    (lx)         (lx)           (lx)
Whole area,
luminaires at
full power (DALI
= 100%)

Sq_LED (4000 K)         878          436         1157

Ro_LED (2621 K)         655          421          827

Ro_LED (6321 K)         552          337          697

Re_T5 (4000 K)         1011          676         1386

Working area,
luminaires at
full power (DALI
= 100%)

Sq_LED (4000 K)        1066          905         1157

Ro_LED (2621 K)         757          650          827

Ro_LED (6321 K)         647          558          697

Re_T5 (4000 K)         1153          914         1386

Luminaire Type    [E.sub.min]/[E.sub.m]  [E.sub.min]/[E.sub.max]
Whole area,
luminaires at
full power (DALI
= 100%)

Sq_LED (4000 K)                    0.50                     0.38

Ro_LED (2621 K)                    0.64                     0.51

Ro_LED (6321 K)                    0.61                     0.48

Re_T5 (4000 K)                     0.67                     0.49
Working area,
luminaires at
full power (DALI
= 100%)

Sq_LED (4000 K)                    0.85                     0.78

Ro_LED (2621 K)                    0.86                     0.79

Ro_LED (6321 K)                    0.86                     0.80

Re_T5 (4000 K)                     0.79                     0.66


Luminance measurements were conducted for the lighting settings used in the preset situations. These measurements were conducted from the corner of the room and at the desk (Fig. 9), because these were the positions where the subjects were seated during the user tests. A Nikon Coolpix 8400 imaging luminance photometer was positioned 120 cm above the floor and its imaging axis was adjusted to the horizontal level. The average luminances of the viewing angle and UGR are presented in Table 5. Full view of the imaging luminance spectrometer was used to calculate the presented values.

TABLE 5.
Measured Average Luminances and UGR-Values of the Test Rooms

                Task Area         Corner of
                                   the Room

Luminaire Type  Average Lv  UGR   Average Lv  UGR
                  (cd/m2)           (cd/m2)

Sq_LED 1000 lx         175  15.1         214  15.1

Sq_LED 600 lx          140  16.3         190  16.4

Sq_LED 300 lx           86  12.6         115  12.7

Re_T5 1000 lx          216  18.8         202  19.1

Re_T5 600 lx           125  16.2         121  17.2

Re_T5 300 lx            67  13.8          61  14.6

Ro_LED 600 lx,         117  18.6          69    20
6000 K

Ro_LED 600 lx,         134  18.6          72  20.2
4500 K

Ro_LED 600 lx,         131  19.2          75  20.3
3000 K


At the same illuminance levels the Ro_LED lighting had the highest UGR-values, followed by the Re_T5 lighting, whereas Sq_LED luminaires had the lowest UGR-values. According to the standard EN 12464-1:2002 [SESKO Standardization in Finland, 2003] the UGR values for typical office work tasks should be [less than or equal to] 19. At 600 lx Ro_LED lighting was above the recommendation, when measured from the corner of the room. Otherwise the measured UGR-values of all tested lighting settings were within the recommendation, although the UGR-values of Ro_LED lighting at all settings and Re_T5 lighting at 1000 lx were close to the limit of the recommendation.

These findings cannot be generalized beyond the scope of this research, because the luminaires were not installed in identical places (Fig. 2) and because their geometry (Fig. 3) and light distribution were not identical. They are, however, of use when interpreting the user study results.

3.3 USER STUDY RESULTS: PRESET SITUATIONS

Re_T5 and Sq_LED lightings were compared to each other at different illuminance levels. Different CCT settings of Ro_LED were compared to each other. Subject gave answers on a nine point scale (1-9) as presented on Appendix 1. The results were transformed to a 0-1 scale so they could be more easily compared with each other. This scale transformation was done using (1):

[R.sub.2] = ([R.sub.1] - 1)/8 (1)

Where [R.sub.1] is the original rating in a 1-9 scale and [R.sub.2] is the rating transformed to a 0-1 scale.

The normality of the results was studied with Kolmogorov-Smirnov-test in SPSS. Only in the question I the answer distributions of 'Re_T5, 4000K, 300 lx' and 'Ro_LED, 600 lx, 6000 K' followed the normal distribution, in every other case the data was not normally distributed. Therefore nonparametric test methods were chosen for further analysis of the data. Mean values of different situations were compared using Kruskal-Wallis and Mann-Whitney tests and correlations between answers were analyzed with Pearson's method. Statistical significance level on these tests was 0.05. Statistical significances between lighting situations are marked to tables after the mean values with an asterisk * and the number after the asterisk tells which two lightings are statistically different.

For example, if lighting A has a mean value 0.87 and lighting B has a mean value 0.70 and these values are significantly different, they are marked 0.87*1 and 0.70*1. If lighting situation is statistically different from multiple other lightings, all the relations are marked after the asterisk. Tables 6 through 16 list all the statistically significant differences between tested lightings, but lighting situations regarding Ro_LED lighting are not further compared to Re_T5 and Sq_LED lighting situations because the Ro_LED lighting was tested only at one illuminance level. Different CCT settings of the Ro_LED lighting have been compared to each other.

3.3.1 GLARE PERCEPTION

Glare was evaluated by asking the subject to first sit in the armchair in the corner of the room and answer the question A: "Are you experiencing glare when you are sitting in the armchair?" Next the subject sat by the office desk and answered the question B "Are you experiencing glare when you are sitting beside the desk?" The results are presented Tables 6 and 7.

Generally the answers imply that subjects perceived slightly more glare as the illuminance level was raised. This difference was statistically significant between the 300 and 1000 lx illuminance levels of the Re_T5 and Sq_LED lighting. In addition the difference between 600 lx and 1000 lx was large enough to be significant with Re_T5 lighting, but not with Sq_LED lighting. The values at the same illuminance levels between Re_T5 and Sq_LED lighting were too close to each other to be significantly different. Mean values of the Ro_LED lighting answers were not significantly different between the tested CCT settings.

The range of the answers varied somewhat between the tested lighting settings because the minimum values of Sq_LED lighting at 300 lx and 600 lx levels were higher than the rest of the tested lighting settings. These two lighting settings had also smaller standard deviation than rest of the tested settings. Median values at 300 lx settings were the same as the maximum values and at other illuminance levels the median around the second highest grade of the 9 point scale. This means that most of the subjects were not experiencing glare when sitting in the armchair. Overall, the results show that there was a slight increase in user perceived glare with increased illuminance levels, whereas the change of CCT did not significantly affect the perception of glare.

Glare observations the subjects made when sitting at the desk followed similar trends as the ones they made in the corner of the room, although here the changes between illuminance levels were more subtle. There was a statistically significant correlation between the glare answers subject gave in the corner of the room and at the desk. The only difference in mean values that was large enough to be statistically significant was between Sq_LED at 300 lx and Re_T5 at 1000 lx. Again the change of CCT did not significantly affect the mean values of Ro_LED lighting and median values of all tested lightings were close to the maximum value. This means that most of the subjects did not experience glare with any of the tested lightings when sitting at the table.

The subjective glare evaluations corresponded partially to UGR-measurements (Table 5). Ro_LED luminaire installations had higher UGR-values at same illuminance levels, when compared to other luminaire installations. This was in line with the subjective evaluations, although in the subjective evaluations the consistency of the answer ratings was too low to be statistically significant. In contrary to the subjective evaluations, a bit lower UGR-values were measured with highest CCT (6000 K) setting of the Ro_LED lighting when compared to the lower CCT settings (3000 K, 4500 K). Differences in subjective glare ratings at different CCTs of Ro_LED lighting were however too small to reach statistical significance.

3.3.2 READING TASK AND DETAIL DISTINCTION

When sitting at the desk in the middle of the task area, subjects read a chapter from a newspaper article printed on white A4-size paper (black letters on a white background) and evaluated how easy it was to read. In addition, the subject looked at a Lea-numbers test chart (black numbers on white background), located on the opposite wall, and evaluated how easy it was to see details. The results of these user observations are presented in Tables 8 and 9.

When subjects were asked "How easy it is to read in this lighting?", the mean values of Re_T5 and Sq_LED lightings at 300 lx and 1000 lx were significantly different from each other, with 1000 lx having higher rating. The results of these lightings at 600 lx and at 1000 lx were similar. At the same illuminance level there were no significant differences between the Re_T5 and Sq_LED lighting. Ranges and standard deviations of answers were a bit smaller at 600 lx illuminance level when compared to the 300 lx and 1000 lx illuminance levels with Re_T5 and Sq_LED lightings. The change in CCT setting had no statistically significant effect to the mean values of Ro_LED lighting. Range of answers was smaller with Ro_LED lighting answers at 4500 K, when compared to the 3000 K and 6000 K settings. The median values of all the tested lighting settings were either on the second or the third highest grade of the 9 point answer scale, which indicates that most subjects found all the tested lighting settings quite suitable for reading task.

When the subjects were asked "How easy it is to see details in the measurement chart on the wall in this lighting?", the mean values of Re_T5 lighting and Sq_LED lighting answers at 300 lx were significantly lower than at 600 lx and 1000 lx illuminance levels. There was also significant, although small, difference between 600 lx and 1000 lx settings of Re_T5 lighting, with 1000 lx answers having higher mean value, whereas with Sq_LED the answers regarding these illuminance levels were almost identical. At the same illuminance level the mean values between Re_T5 and Sq_LED lightings were too close to each other to be significant. The mean values of Ro_LED answers were almost identical at different CCT settings. The ranges of the subject given ratings between the tested lighting settings were almost the same. The median values of the subject ratings were either in the highest or in the second highest grade of the 9 point scale with tested 600 lx and 1000 lx settings of Re_T5 and Sq_LED lighting, which means that most of the subjects found it very easy to distinguish details on the wall in these lighting conditions. At 300 lx the median values were one grade lower on Sq_LED lighting and two grades lower on Re_T5 lighting than at 600 lx, which means that at this illuminance level the detail distinction was usually considered slightly harder.

There was a statistically significant correlation (Pearson) between the answers to questions C and D. Therefore, when subject found it easy to read on the task area, it was also easy to distinguish details on the Lea-numbers test chart on the wall. Overall all tested lighting settings were usually found adequate for reading task and for distinguishing details from the wall, although the large range of the user given ratings indicates large individual differences.

3.3.3 NATURALNESS AND PLEASANTNESS OF COLORS

The naturalness of colors was evaluated by asking the subject to view the Macbeth Color Checker Chart (Fig. 4a) placed on the table in front of them, posters (Figs. 4b and 4c) on the wall and the skin color of subject's own hand. Color pleasantness was evaluated by asking the subject to rate the general pleasantness of colors in the room. The results are presented in Tables 10 and 11.

The mean values of subject answers to question E: "How natural are the colors looking in this lighting?" varied too little between tested lightings to be statistically significant, regardless of the illuminance level or the CCT. Only statistically significant difference was between Sq_LED lighting at 300 lx and at 600 lx, in which case 600 lx was considered about 9 percent more natural on average. Ro_LED lighting had larger range and standard deviation at CCT 6000 K than at CCTs 3000 K and 4500 K but the mean values of these settings were not significantly different.

When subjects were asked "Do you think the color of the lighting is pleasant?", the mean values of Re_T5 and Sq_LED lightings at the same illuminance level were not significantly different from each other. 600 lx and 1000 lx levels were considered almost equally pleasant with Re_T5 and Sq_LED lighting. Subjects considered the 300 lx significantly less pleasant than higher illuminance levels with Re_T5 lighting, whereas this difference was too small with Sq_LED lighting to be significant. Ro_LED lighting at CCT 6000 K was considered significantly less pleasant than at CCT's 3000 K and 6000 K.

There was a statistically significant correlation (Pearson) between the mean values of color naturalness and color pleasantness answers. Therefore, if the colors looked natural in the lighting then the color of the lighting was also considered pleasant. Subjective ratings of the color naturalness did not correlate with the measured color rendering indices of the luminaires. The Ro_LED luminaire had highest measured CIE CRI (Ra=89 - 94, depending on CCT) of the tested luminaires, but subjects ranked the color of Ro_LED lighting to be less natural than Re_T5 (Ra=82) and Sq_LED lightings (Ra=86), regardless of the CCT setting of the Ro_LED lighting.

3.3.4 AMOUNT OF LIGHT AND ITS SUFFICIENCY FOR WORKING TASKS

Sufficiency of illuminance level was evaluated by asking the subject to look at the paper on the table and the task area in general and asking whether the illuminance level was sufficient for typical office work tasks. Pleasantness of light amount was evaluated by asking the subject to look at the task area and the room in general and asking them whether the light amount was pleasant in the room. In this question the optimum level was in the middle of the scale, in contrary to other questions of the questionnaire. Results are presented in Tables 12 and 13.

The subject ratings to question F: "Is the illuminance level sufficient for typical office working tasks?" varied more between the tested lightings than in any other question. There was a direct relation between the illuminance level and its perceived sufficiency: the higher the illuminance level, the more sufficient it was considered. The differences between the illuminance levels were significant in all cases, but at the same illuminance level the Re_T5 and Sq_LED lightings were considered equally sufficient. The ranges of the answers were larger with 300 lx lightings and the median values lower than those of the higher illuminance levels. There were no significant differences in the mean values between different CCT settings of Ro_LED lighting. Overall the findings of this question can be simplified as: the more there was light, the more sufficient the subject considered it to be for working tasks.

Overall the 600 lx was considered to be the most pleasant of the three tested illuminance levels. Tested 600 lx lightings deviated 1-6 percent towards the "Way too much light", when compared to the optimal value 0.5. At 1000 lx the Re_T5 and Sq_LED lightings deviated 30 percent towards "Way too much light" end of the scale. At 300 lx T5 lighting deviated 28 percent and Sq_LED 19 percent towards the "Way too dim" end of the scale. The differences between illuminance levels were statistically significant in each case. There were no significant differences between Re_T5 and Sq_LED lightings at the same illuminance level. The change of CCT setting did not affect significantly to the mean values of the Ro_LED lighting. The deviation and the range of the answers were almost the same with all tested lightings.

3.3.5 SHADOW FORMATION

Shadow formation was evaluated by asking the subject to evaluate the naturalness of the shadows in the room, especially the shadows that the coffee mug casted on the table. This mug was place in front of the subject to table, in middle of the task area. Luminaires used in this study differ in their shape and light distribution, which affects to the shadow formation. Therefore the results cannot be generalized to apply to the light source types used here, but only to the tested luminaires. Positions of the luminaires can be seen in Fig. 2 and the illuminance distribution of the room in Fig. 8. Results of the subject tests are presented in Table 14.

TABLE 14.
Mean Values, Medians, and Standard Deviations of
Shadow Naturalness Results.

Question K: How Natural the Shadows Look Like in This Lighting?
0 = Completely Unnatural, 1 = Completely Natural

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000
                                                        lx

Mean       0.72     0.70     0.70     0.66     0.62     0.68

Median     0.75     0.75     0.75     0.75     0.63     0.75

Std.       0.20     0.21     0.23     0.24     0.25     0.23
dev.

Min        0.25     0.25     0.25     0.25     0.13     0.25

Max        1.00     1.00     1.00     1.00     1.00     1.00

Range      0.75     0.75     0.75     0.75     0.88     0.75

        Ro_LED,  Ro_LED,  Ro_LED,
        4500 K,  6000 K,  3000 K,
        600 lx   600 lx   600 lx

Mean       0.67     0.69     0.73

Median     0.75     0.75     0.75

Std.       0.29     0.27     0.24
dev.

Min        0.00     0.00     0.00

Max        1.00     1.00     1.00

Range      1.00     1.00     1.00

* Statistically significant difference (P < 0.05) between mean
values with same suffix.


Although the mean values of the answers imply some differences between lightings, the differences were too small to be statistically significant. The median values of the answers were identical in every case, except in Re_T5 lighting at 300 lx. Most apparent finding of the subject evaluation of the shadow naturalness is that the range of answers regarding Ro_LED lightings was slightly larger, when compared to Re_T5 and Sq_LED lightings. Overall there was no clear pattern between the perceived shadow naturalness and the illuminance level or CCT of the lighting.

3.3.6 GENERAL PLEASANTNESS OF THE LIGHTING AND LUMINAIRE APPEARANCE

The subjects were asked to evaluate the general pleasantness of the lighting and the appearance of luminaires. The results are presented in Tables 15 and 16.

At the same illuminance levels the tested Re_T5 and Sq_LED lightings were considered to be equally pleasant and there were no differences between subject answers at 600 lx and 1000 lx. At 300 lx both of these lightings were considered to be less pleasant than at higher illuminance levels. Ro_LED lighting was considered to be significantly less pleasant at CCT 6000 K than at CCT 3000 K.

Subjects preferred the appearance of Sq_LED luminaires most, followed by the Ro_LED luminaires. Re_T5 fluorescent lamp luminaires were considered to be least good looking. The differences in the user given ratings were however relatively small. Illuminance level or color temperature of the light had no significant effect to the user ratings.

Although the appearance of luminaires used for general lighting in offices may often be considered to be of secondary importance, it is worthwhile to notice that subjects preferred the less conventional appearance of the tested LED luminaires over the more traditional "box" design of the tested Re_T5 luminaires.

3.4 GENDER AND AGE RELATED DIFFERENCES IN THE USER RATINGS

Since the data was not normally distributed, effect of gender to answers was studied with Mann-Whitney tests; results are listed on Table 17. Significance level on these tests was assumed to be p < 0.05.

TABLE 17.
Effect of Gender to Results. Table Contains p-Values of
Mann-Whitney Test Results

        Question          1        2        3         4
                        Sq_LED,  Sq_LED,  Sq_LED,    Re_T5,
                        4000 K,  4000 K,  4000 K,    4000 K,
                        600 lx   300 lx   1000 lx     600 lx

A     Are you              .171     .216   .012 *       .064
      experiencing
      glare when you
      are sitting in
      the armchair?

B     Are you              .135     .149     .159     .042 *
      experiencing
      glare when you
      are sitting
      beside
      the desk?

C     How easy it is       .486     .420     .283       .057
      to read in this
      lighting?

D     How easy it is       .107     .976   .031 *       .186
      to see details
      in the
      measurement
      chart on the
      wall in this
      lighting?

E     How natural are      .063     .856     .063       .350
      the colors
      looking in this
      lighting?

F     Is the               .975     .436     .722       .612
      illuminance
      level sufficient
      for typical
      office working
      tasks?

G     Do you think the     .563     .475     .886       .507
      light amount is
      pleasant in the
      room? (5 =
      optimal)

H     Do you think the     .237     .869   .005 *       .080
      color of the
      lighting is
      pleasant?

I     In general, do       .104     .566   .024 *       .461
      you think the
      lighting is
      pleasant?

J     Do you think the     .207     .774     .166       .098
      luminaires
      themselves look
      good/stylish?

K     How natural the      .928     .230     .905       .155
      shadows look
      like in this
      lighting?

        Question          5        6        7         8
                        Sq_LED,  Sq_LED,  Sq_LED,    Re_T5,
                        4000 K,  4000 K,  4000 K,    4000 K,
                        300 lx   1000 lx  600 lx      600 lx

A     Are you              .144   .012 *     .424     .234
      experiencing
      glare when you
      are sitting in
      the armchair?

B     Are you              .136     .248     .082     .093
      experiencing
      glare when you
      are sitting
      beside
      the desk?

C     How easy it is     .017 *     .186     .748   .004 *
      to read in this
      lighting?

D     How easy it is       .284     .987     .158     .134
      to see details
      in the
      measurement
      chart on the
      wall in this
      lighting?

E     How natural are      .706     .112   .017 *   .025 *
      the colors
      looking in this
      lighting?

F     Is the               .455     .476     .891     .515
      illuminance
      level sufficient
      for typical
      office working
      tasks?

G     Do you think the     .780     .165     .370   .024 *
      light amount is
      pleasant in the
      room? (5 =
      optimal)

H     Do you think the     .845   .033 *   .028 *   .000 *
      color of the
      lighting is
      pleasant?

I     In general, do       .436     .148   .011 *   .000 *
      you think the
      lighting is
      pleasant?

J     Do you think the     .307   .009 *     .209     .823
      luminaires
      themselves look
      good/stylish?

K     How natural the      .917     .410     .054   .045 *
      shadows look
      like in this
      lighting?

        Question          9
                        Sq_LED,
                        3000 K,
                        600 lx

A     Are you              .903
      experiencing
      glare when you
      are sitting in
      the armchair?

B     Are you              .403
      experiencing
      glare when you
      are sitting
      beside
      the desk?

C     How easy it is       .212
      to read in this
      lighting?

D     How easy it is       .693
      to see details
      in the
      measurement
      chart on the
      wall in this
      lighting?

E     How natural are      .821
      the colors
      looking in this
      lighting?

F     Is the               .253
      illuminance
      level sufficient
      for typical
      office working
      tasks?

G     Do you think the     .079
      light amount is
      pleasant in the
      room? (5 =
      optimal)

H     Do you think the     .589
      color of the
      lighting is
      pleasant?

I     In general, do       .753
      you think the
      lighting is
      pleasant?

J     Do you think the     .123
      luminaires
      themselves look
      good/stylish?

K     How natural the      .350
      shadows look
      like in this
      lighting?

* p-Value < 0.05, statistically significant difference.


As can be seen from the Table 17, there were several situations during the test procedure where the gender had a statistically significant effect to the answers. These situations are listed in Table 18, where the letter represents the asked question and number tells which lighting situation was in question.

TABLE 18.
Mean Values of Answers to Questions with Statistically
Significant Difference between Genders

Situation  Male  Female
A3         0.84    0.59
A6         0.78    0.52
A8         0.78    0.63
B4         0.88    0.72
C5         0.80    0.59
C8         0.82    0.67
D3         0.90    0.77
E7         0.78    0.60
E8         0.71    0.49
G8         0.48    0.64
H3         0.84    0.60
H6         0.83    0.65
H7         0.74    0.50
H8         0.66    0.27
I3         0.79    0.59
I7         0.71    0.48
I8         0.67    0.26
J6         0.68    0.44
K8         0.77    0.51


In all the statistically significant ratings between genders, the average ratings given by females were lower than those given by males, which means that males were on average more satisfied to the lighting in question than females. The average ages of females and males were however different; females=38 years old, males=27 years old; and this might have affected to the results. The age of subjects was distributed unevenly: 72.5 percent of subjects (N = 29) belonged to age group 20-29 years old, 12.5 percent (N = 5) to 30-39 years old, 7.5 percent (N = 3) to 40-49 years old and 7.5 percent (N = 3) to 50-59 years old. Therefore the sample sizes from other age groups than 20-29 are too small to draw conclusions about the effect of age to results.

Due to unequal age and gender distribution of the subjects, no reliable conclusions can be drawn about the age or gender related differences in user ratings.

3.5 USER STUDY RESULTS: IDEAL LEVELS SET BY THE SUBJECTS

After the subjects had answered to questions about the preset lighting situations, they were asked to set their own ideal lighting using the controls on the table (Fig. 3). One of the test subjects had no time to participate to this part of the test, so the answers are based on opinions of 39 subjects instead of 40 that answered to the preset situations. The results of these subject made adjustments are presented in Table 19. Adjustment range was from 0 to 1142 lx with Re_T5 lighting and 0 to 738 lx with Ro_LED lighting. CCT adjustment range of the Ro_LED lightings was 2621 to 6321 K.

TABLE 19. Statistics of User Set Illuminance Levels and Color
Temperatures on the Work Plane

        User Set  User Set  User Set  Preset Color    User Set
        Minimum   Optimum   Optimum   Temperature     Optimum
          Level     Level     Level                      Color
                                                    Temperature

         Re_T5,    Re_T5,    Ro LED    Ro LED (K)   Ro LED (K)
         4000 K    4000 K     (lx)
          (lx)      (lx)

Mean         326       648       517          4462         4151

Median       309       643       517          4500         4047

Std.         140       215       115          1216          727
dev.

Min           74       121       245          3000         2957

Max          600      1142       738          6000         6037

Range        526      1021       493          3000         3080

        Difference  Difference
          between       between
         Preset T   Optimal Re
         and User   T5 and Ro
          Set T         LED
                     Settings

          Ro LED

Mean           -1%        -13%

Median         -5%        -23%

Std.           29%         28%
dev.

Min           -51%       - 48%

Max           101%        102%
Range



The subjects first set the Re_T5 lighting to a minimum illuminance level that they would be willing to work under for short periods of time (1 h). This was meant to emulate a situation where the electricity consumption of the building must be limited for some reason, for example due to energy regulations or limited power availability. The preset illuminance level was 300 lx before the adjustment.

On average the users set the minimum illuminance level for temporary office working to around 300 lx with Re_T5 lighting. On the next stage test subjects were asked to set the Re_T5 lighting to an optimum illuminance level for long time office work. This time the average illuminance level approximately doubled when compared to the minimum level setting, being above 600 lx on the work plane. The deviation of the answers was also larger in this case. At the moment the recommendation for minimum work area illuminance for most office tasks is 500 lx [SESKO Standardization in Finland, 2003], so the tested subjects preferred to have slightly higher illuminance level on the task area than is the minimum recommendation. The range of the answers was wide in both cases, which shows that the individual preferences between people can be very different from the average.

On the final part of the user test, the subject was asked to set the optimal illuminance level and CCT for long time office work using the Ro_LED luminaires. Preset color temperature of the lighting was varied between 3000 K, 4500 K and 6000 K settings and it was set before the subject entered the room. Preset illuminance level on the room was 300 lx. Subjects set illuminance level to about 130 lx lower on average with Ro_LED luminaires than with Re_T5 luminaires. To counteract the lower adjustment range of Ro_LED lighting, subjects were asked whether they were satisfied with the illuminance level after they had adjusted the Ro_LED lighting (Appendix 1, form on the lower right). A similar question was asked about the color of the lighting. The average deviation from optimal rating was -0.25 percent for illuminance level and 0.25 percent for color of the light, so subjects did not apparently feel that they needed a wider illuminance or color adjustment range.

However, Logadottir and others [2011a] showed that the adjustment range may significantly affect user set illuminance. Therefore the difference in user set optimum illuminance between Re_T5 and Ro_LED lightings may be a result of the different adjustment range, rather than result of the different light distribution and CCT of the Re_T5 and Ro_LED lightings.

The preferred mean color temperature was 4151 K. There were deviations in the answers, but on average subjects tended to adjust the CCT more towards the preset CCT that was set before the beginning of adjustment task, when compared to the general average (Table 20). This finding is similar to that found by the Logadottir and others [2011b]. Pearson correlation between the preset CCT and the user set CCT was 0.313 and its significance was 0.052, which is just below the statistical significance level.

TABLE 20.
Variation of User Set Color Temperature by the Preset
Color Temperature

Preset   Average   N   Std.
T(K)     User Set      Dev.
           T(K)

3000 K       3841  13   823

4500 K       4226  14   526

6000 K       4400  12   756

Overall      4151  39   727


3.6 ENERGY SAVING RELATION TO SUBJECTIVE USER COMFORT

According to luminaire measurements described in 3.1, the luminous efficacy of all the three measured luminaire types reduced when the luminaires were dimmed. However the LED luminaires maintained relatively better their luminous efficacy when compared to Re_T5 fluorescent lamp luminaire and also the absolute luminous efficacy of LED luminaires was better at the same dimming levels.

With Re_T5 lighting, subject rating to question F: "Is the illuminance level sufficient for typical office working tasks?" at 300 lx was 32 percent lower when compared to 600 lx rating, whereas the rating at 1000 lx was 13 percent higher when compared to 600 lx rating. The same trend was visible with Sq_LED lighting, as can be seen from the Table 21. This indicates that increasing the illuminance level from 600 to 1000 lx has less significant effect on perceived sensation of lighting sufficiency than reducing it from 600 to 300 lx. The mean values of illuminance level sufficiency answers were almost similar with Re_T5 and Sq_LED lightings at the same illuminance levels. Power consumption relation to

task area illuminance (P/100 lx) remained almost constant with Sq_LED luminaires whereas it continuously decreased with Re_T5 luminaires, when the luminaires were dimmed. This means that although the decrease in subjective illuminance level sufficiency was the same with both luminaires, more energy was saved by dimming Sq_LED luminaires than Re_T5 luminaires.

TABLE 21.
Changes in Power Consumption, Dimming Level, and
Subjective Lighting Sufficiency in Relation to Illuminance Level
on Task Area

Task Area      DALI     P (W)  P/P_600  P(W)/100lx  Illuminance
Illuminance  Dimming           lx                    Level
              Level                                 Sufficiency
               (%)                                    Based on
                                                          User
                                                        Answers

                                           Re_T5

300 lx            27     92      64%          31         0.56

600 lx            56    144     100%          24         0.83

1000 lx           92    200     139%          20         0.93
                                          Sq_LED

300 lx            27     58      56%          19         0.59

600 lx            53    104     100%          17         0.86

1000 lx           95    176     169%          18         0.92

Task Area      Sufficiency/
Illuminance  Sufficiency_600
                   lx

300 lx                   68%

600 lx                  100%

1000 lx                 113%

300 lx                   68%

600 lx                  100%

1000 lx                 107%


Ro_LED luminaires were not compared to the other luminaire types here because they were evaluated only at 600 lx illuminance level during user test.

4 DISCUSSION AND CONCLUSIONS

The luminous efficacy of all tested luminaires was highest at full power. Sq_LED and Ro_LED luminaires had higher luminous efficacy than Re_T5 luminaire, and they also maintained their luminous efficacy better when the luminaires were dimmed. The CCT of the Ro_LED luminaire was adjustable between 2621 and 6321 K and its luminous efficacy was highest at the lowest CCT. Ro_LED luminaires had higher measured values of CIE CRI than the Sq_LED and Re_T5 luminaires. The UGR-values of Re_T5 and Sq_LED lightings were below the recommended maximum value for office work at all tested illuminance levels. UGR-values of the Ro_LED lighting were just above the recommendation at all CCTs when measured from the corner of the room, but below the recommendation when measured at the desk.

Subjects perceived more glare as the illuminance level was raised, but there were no significant differences between tested Re_T5 and Sq_LED lightings at the same illuminance levels. The CCT setting did not affect significantly to the user perceived glare with tested Ro_LED lighting. Overall, most of the subjects experienced no glare or only small amounts of it with all tested lighting settings. The glare ratings given by subjects in the corner of the room and at the desk had statistically significant correlation.

The 300 lx illuminance level was considered significantly harder for reading task than 1000 lx with Re_T5 and Sq_LED lightings, but there were no differences in the user ratings between 600 lx and 1000 lx. At the same illuminance level there were no significant differences between Re_T5 and the Sq_LED lighting conditions. Results of the Ro_LED lighting were similar at all tested CCT settings. The median values of the user ratings were close to the maximum value with all tested lighting settings, which implies that most subjects found it easy to read in these lighting conditions. The same kind of trend in the user ratings was visible when the subjects were asked how easy it was to see details in the measurement chart on the wall: an increase in illuminance resulted in an increase in ratings. There was a statistically significant correlation between the answers of reading task easiness and detail distinction from the wall.

There were no significant differences in the mean subject ratings of color naturalness at different illuminance levels or between Re_T5 and Sq_LED lightings. User ratings of Ro_LED lighting at different CCTs were not significantly different, but at CCT 6000 K the user ratings varied more than at 3000 K or 4500 K.

Color of the lighting was about equally pleasant at 600 lx and 1000 lx illuminance levels and there were no significant differences between Re_T5 and Sq_LED lighting conditions at the same illuminance levels. At 300 lx the color of the lighting was evaluated to be less pleasant than the same lighting at higher illuminance levels and with Re_T5 lighting this difference was large enough to be statistically significant. Subjects evaluated the CCT 6000 K setting of the Ro_LED lighting to be less pleasant than the CCTs 3000 K and 4500 K, which were considered equally pleasant. There was a significant correlation between the color naturalness and pleasantness answers.

The illuminance level had direct relation to how sufficient subjects felt it was for office working tasks: the higher the illuminance level, the more sufficient it was according to user ratings. The change in CCT value on the other hand had no significant effect to the mean ratings of Ro_LED lighting.

600 lx was considered to be closest to the optimal, when the subjects were asked how pleasant the light amount was in the room. 300 lx was considered to give less than optimal amount of light and 1000 lx was considered to give too much light. The differences between the illuminance levels were statistically significant, but there were no significant differences in the mean values of the user answers between the tested color temperatures of the Ro_LED lighting or between the Re_T5 and Sq_LED lightings at the same illuminance level.

When subjects were asked to evaluate how natural the shadows looked in the room, the mean values of the answers did not reveal any significant differences between lightings, regardless of illuminance level or CCT setting.

Sq_LED lighting and Re_T5 lighting were rated equally pleasant at similar illuminance levels. 300 lx was considered to be significantly less pleasant with both of these lightings than 600 lx and 1000 lx levels, which were rated equally pleasant. Ro_LED lighting was rated significantly less pleasant at CCT 6000 K than at 3000 K or 4500 K.

Subjects preferred the appearance of both tested LED luminaires over the Re_T5 luminaire, which represented more conventional office luminaire design than the LED luminaires.

There were some cases in which the mean ratings varied significantly according to the gender of the subject. The average age of the female subjects was however over 10 years higher than the average age of male subjects, and therefore the differences in the average values may be age related rather than gender related. However too few subjects from the older age groups than 20-29 years old attended to this test to draw any definite conclusions about this matter.

Further research would benefit from larger number of subjects, or at least age groups should be more equal sized. Additional research topics could also involve the testing of Ro_LED lighting at CCT 4000 K at different illuminance levels, so it could be compared more directly to the Re_T5 and Sq_LED lighting. There could be also more CCT settings of the Ro_LED involved in this study.

When subjects were given opportunity to set their own ideal illuminance levels, on average they set the minimum acceptable illuminance level for brief duration office work to 326 lx with Re_T5 lighting. In this context 'brief duration' means 1 h temporary situation when the available electricity is limited, for example due to power shortage. The average optimum user set illuminance level was 648 lx with Re_T5 lighting and 517 lx with Ro_LED lighting. These illuminance levels are bit higher than those found by Boyce and others [2006], Villa and Labayrade [2012] and Veitch and Newsham [2000] but close to those found by Logadottir and others [2011a]. The deviation in the user set levels was large with both of the tested lightings and there were indications that adjustment range affected to user set illuminances. This was found also in the study of Logadottir and others [2011a]. Therefore the larger adjustment range of the Re_T5 lighting may have influenced to the results and the difference in the average preferred illuminance levels cannot be explicitly explained by the lighting type. Thus the natural direction for further adjustment study research would be to compose a setup where different lighting types would have equal adjustment ranges. This would help to pinpoint whether the luminaire type has an effect to the user preferred illuminance, or are the differences more related to the available adjustment range. Sq_LED lighting could be involved in this study also.

The optimum user set CCT was 4151 K and it seemed that the preset CCT of the room affected to the user set CCT. This finding is congruent with the results of the previous study done by Logadottir and others [2011b]. The correlation between the preset CCT and user set CCT was however just below the statistical significance level in our study.

Subjective perception of illuminance level sufficiency changed similarly with LED and Re_T5 luminaires and the reduction from 600 lx to 300 lx illuminance level affected more to the perceived illuminance sufficiency than the increase from 600 lx to 1000 lx. Subjects evaluated 600 lx to be equally pleasant to 1000 lx and also the reading task and detail distinction from the wall was evaluated to be equally easy with these two illuminance levels. However, more glare was observed at 1000 lx than at 600 lx. 300 lx was considered to be less pleasant than 600 lx and 1000 lx and the reading task and detail distinction from the wall was considered to be harder at 300 lx than at higher illuminance levels. Subjects evaluated that at 600 lx the light amount in the room was closer to optimal than at 300 lx or at 1000 lx. Subjects also set the optimal illuminance levels nearby 600 lx during the adjustment tasks. Based on these findings, 600 lx can be considered more recommendable illuminance level for office work than 300 lx or 1000 lx. However no distinct differences in the subjective lighting quality of the tested Re_T5 and Sq_LED lightings at equal illuminance levels were found in this study. If T5 luminaires are used, it must be considered that their luminous efficacy decreases significantly when dimmed. LED luminaires maintain their luminous efficacy better when dimmed, and therefore they are more energy efficient options for lighting systems where multiple illuminance levels are required.

The most apparent finding regarding CCT in the user test was that subjects considered the color of the Ro_LED lighting at CCT 6000 K less pleasant than at 3000 K and 4500 K. Similar results were found when subjects were asked about the general pleasantness of lighting. The user ratings at different CCTs were however not significantly different from each other at other tested questions.

There were large deviations in the user set CCTs and illuminance levels, which indicates that the individual preferences of different people can be very different. This has been confirmed also in the previous research of the topic [Logadottir and others, 2011a; Villa and Labayrade 2012; Veitch and Newsham, 2000; Boyce and others, 2006]. Therefore additional CCT and dimming control can increase user comfort in the space by allowing each user to set the lighting according to their own preferences.

ACKNOWLEDGMENTS

This study was done as part of RYM Indoor Environment research program that is funded by TEKES, Aalto University and several Finnish companies. More information about this program can be found at http://www.rym.fi/en/pro-grams/indoorenvironmentprogram/

In the future it might be necessary to limit the lighting levels due to the energy optimization of the building. Adjust the lighting to a minimum level that you would still be reasonable comfortable to work for an hour, by using the slide- switch on the table (LIGHT LEVEL). When you are ready, please inform the supervisor of the test before moving to the next question.

Supervisor comments (do not write here anything by yourself):

Ix:

%:

Assuming that the energy consumption is not a limiting factor, what do you think would be the optimal lighting level for long term office work? Adjust the lighting to a level that you would feel most comfortable to work in office environment, by using the slide-switch on the table (LIGHT LEVEL). When you are ready, please inform the supervisor of the test.

Supervisor comments (do not write here anything by yourself):

Ix:

%:

THANK YOU FOR YOUR ANSWERS!

REFERENCES

Boyce PR, Veitch JA, Newsham GR, Jones CC, Heerwagen J, Myer M, Hunter CM. 2006. Occupant use of switching and dimming controls in offices. Lighting Res Technol. 38(4):358-376.

European Commission. 2010. The EU climate and energy package. http://ec.europa.eu/clima/policies/package/index_en.htm [Accessed: November 6, 2012].

Eurostat. 2012. Electricity and natural gas price statistics. http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Electricity_and_natural_gas_price_statistics [June 27, 2012].

[IEA] International Energy Agency. 2006. Light's Labour's Lost. [Online] http://www.iea.org/papers/2008/cd_energy_efficiency_policy/4-Lighting/4-light2006.pdf. 560 p.

Logadottir A, Christoffersen J, Fotios S. 2011a. Investigating the use of an adjustment task to set the preferred illuminance in a workplace environment. Lighting Res Techn. 43(4):403-422.

Logadottir A, Christoffersen J, Fotios S. 2011b. Investigating the use of an adjustment task to set preferred colour of ambient illumination. Color Res Appl. Published online before print. http://doi.wiley.com/10.1002/col.20714 [Accessed: November 23, 2012].

Roisin B, Deneyer A, Eugene C. 2006. Optimization of lighting power consumption in offices. In: International Lighting Symposium "Modern Quality Solutions for an Efficient Lighting". Sinaia (Romania).

SESKO Standardization in Finland, 2003. SFS-EN 12464-1 Light and lighting. Lighting of work places. Part 1: Indoor work places.

Statistics Finland. 2011. Appendix figure 5. Price of electricity by type of consumer, c/kWh. Statistics: Energy prices [e-publication]. http://www.stat.fi/til/ehi/2011/04/ehi_2011_04_2012-03-20_kuv_005_en.html [Accessed: May 31, 2012].

Veitch JA, Newsham GR, 2000. Preferred luminous conditions in open-plan offices: research and practice recommendations. Lighting Res Technol. 32(4):199-212.

Villa C, Labayrade R, 2012. Multi-objective optimisation of lighting installations taking into account user perferences--a pilot study. Lighting Res Technol. Published online before print. DOI: 10.1177/1477153511435629. [Accessed: November 23, 2012]

TABLE 6.
Mean Values, Medians, and Standard Deviations of
Glare Perception Results. Subject Was Sitting in the Armchair
in the Corner of the Room

Question A: Are You Experiencing Glare When You Are Sitting
in the Armchair?

0 = Disturbingly Much, 1 = Not at All

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000
                                                        lx

Mean     0.87 *   0.89 *   0.77 *   0.82 *   0.86 *   0.70 *
           (1),     (3),      (6)      (8)     (9),     (2),
            (2)     (4),                       (10)     (7),
                    (5),                                (8),
                    (6),                                 (9)
                     (7)

Median     0.94     1.00     0.88     0.88     1.00     0.81

Std.       0.17     0.16     0.27     0.22     0.21     0.27
dev.

Min        0.38     0.38     0.13     0.25     0.13     0.00

Max        1.00     1.00     1.00     1.00     1.00     1.00

Range      0.63     0.63     0.88     0.75     0.88     1.00

        Ro_LED,  Ro_LED,  Ro_LED,
        4500 K,  6000 K,  3000 K,
        600 lx   600 lx   600 lx

Mean     0.76 *   0.72 *   0.78 *
            (3)     (1),      (5)
                    (4),
                    (10)

Median     0.88     0.81     0.88

Std.       0.26     0.27     0.25
dev.

Min        0.25     0.13     0.00

Max        1.00     1.00     1.00

Range      0.75     0.88     1.00

* Statistically significant difference (P < 0.05) between mean
values with same suffix.

TABLE 7.
Mean Values, Medians, and Standard Deviations of Glare
Perception Results. Subject Sitting at the Desk

Question B: Are You Experiencing Glare When You Are Sitting beside
the Desk?

0 = Disturbingly Much, 1 = Not at All

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,   Ro_LED,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4500 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx    1000     600 lx
                                                        lx

Mean     0.85 *   0.89 *     0.82   0.83 *   0.84 *   0.75 *   0.77 *
            (1)     (2),               (6)      (7)      (5)      (2)
                    (3),
                    (4),
                     (5)

Median     0.88     0.94     0.88     0.88     1.00     0.88     0.88

Std.       0.19     0.16     0.23     0.20     0.22     0.26     0.23
dev.

Min        0.38     0.25     0.13     0.38     0.13     0.13     0.25

Max        1.00     1.00     1.00     1.00     1.00     1.00     1.00

Range      0.63     0.75     0.88     0.63     0.88     0.88     0.75

        Ro_LED,  Ro_LED,
        6000 K,  3000 K,
        600 lx   600 lx

Mean              0.79 *
                     (4)

Median     0.75     0.88

Std.       0.26     0.23
dev.

Min        0.13     0.00

Max        1.00     1.00

Range      0.88     1.00

* Statistically significant difference (P < 0.05) between mean
values with same suffix.

TABLE 8.
Mean Values, Medians, and Standard Deviations of Reading
Task Results

Question C: How Easy It Is to Read in This Lighting?
0 = Really Hard, 1 = Really Easy

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx   1000 lx

Mean     0.86 *   0.76 *   0.86 *     0.83   0.74 *   0.84 *
           (1),     (1),     (4),              (2),     (5),
           (2),     (4),     (6),              (6),     (9),
            (3)      (5)     (7),               (9)     (10)
                              (8)

Median     0.88     0.88     0.88     0.88     0.75     0.88

Std.       0.14     0.21     0.20     0.18     0.23     0.21
dev.

Min        0.38     0.13     0.25     0.38     0.25     0.25

Max        1.00     1.00     1.00     1.00     1.00     1.00

Range      0.63     0.88     0.75     0.63     0.75     0.75

        Ro_LED,  Ro_LED,  Ro_LED,
        4500 K,  6000 K,  3000 K,
        600 lx   600 lx   600 lx

Mean       0.82   0.78 *   0.78 *
                    (3),      (8)
                    (7),
                    (10)

Median     0.88     0.75     0.88

Std.       0.19     0.21     0.22
dev.

Min        0.38     0.00     0.13

Max        1.00     1.00     1.00

Range      0.63     1.00     0.88

* Statistically significant difference (P < 0.05) between
mean values with same suffix.

TABLE 9.
Mean Values, Medians, and Standard Deviations of
Lea-Numbers Test Chart Viewing Task Results

Question D: How Easy It Is to See Details in the Measurement
Chart on the Wall in This Lighting?

0 = Really Hard, 1 = Really Easy

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000
                                                        lx

Mean     0.84 *   0.73 *   0.86 *   0.82 *   0.69 *   0 88 *
           (1),     (2),     (4),     (3),     (1),     (5),
            (2)     (3),     (6),    (10),     (6),    (11),
                    (4),     (7),     (11)    (10),    (12),
                     (5)     (8),              (12)    (13),
                              (9)                      (14),
                                                        (15)

Median     0.88     0.75     0.88     0.88     0.63     1 00

Std.       0.15     0.18     0.17     0.17     0.22     0 18
dev.

Min        0.38     0.38     0.38     0.38     0.25     0 25

Max        1.00     1.00     1.00     1.00     1.00     1 00

Range      0.63     0.63     0.63     0.63     0.75     0 75

        Ro_LED,  Ro_LED,  Ro_LED,
        4500 K,  6000 K,  3000 K,
        600 lx   600 lx   600 lx

Mean     0 78 *   0 79 *   0.79 *
           (7),     (8),     (9),
           (13)     (14)     (15)

Median     0 88     0 81     0 88

Std.       0 20     0 19     0.19
dev.

Min        0 38     0 25     0.25

Max        1 00     1 00     1.00

Range      0 63     0 75     0.75

* Statistically significant difference (P < 0.05) between mean
values with same suffix.

TABLE 10.
Mean Values, Medians, and Standard Deviations of Color
Naturalness Results

Question E: How Natural Are the Colors Looking in This Lighting?

0 = Completely Unnatural, 1 = Completely Natural

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000
                                                        lx

Mean     0.84 *   0.77 *   0.81 *   0.83 *     0.75   0.82 *
           (1),      (1)     (5),     (7),             (10),
           (2),               (6)     (8),              (11)
           (3),                        (9)
            (4)


Median     0.88     0.75     0.88     0.88     0.81     0.88

Std.       0.14     0.17     0.20     0.16     0.21     0.18
dev.

Min        0.38     0.38     0.38     0.50     0.25     0.38

Max        1.00     1.00     1.00     1.00     1.00     1.00

Range      0.63     0.63     0.63     0.50     0.75     0.63

        Ro_LED,  Ro_LED,  Ro_LED,
        4500 K,  6000 K,  3000 K,
        600 lx   600 lx   600 lx

Mean     0 73 *   0.64 *   0.74 *
           (2),     (3),     (4),
           (5),     (6),      (9)
           (7),     (8),
           (10)     (11)

Median     0 75     0 75     0.75

Std.       0 20     0.26     0.20
dev.

Min        0 25     0.00     0.25

Max        1 00     1.00     1.00

Range      0 75     1.00     0.75

* Statistically significant difference (P < 0.05)
between mean values with same suffix.

TABLE 11.
Mean Values, Medians, and Standard Deviations of Color
Pleasantness Results Question H: Do You Think the Color of the
Lighting Is Pleasant? 0 = Really Unpleasant, 1 = Really Pleasant

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,   Ro_LED,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4500 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000     600 lx
                                                         lx

Mean     0.74 *   0.69 *   0.77 *   0.78 *   0.62 *   0.78 *   0.67 *
           (1),     (3),     (6),     (3),     (1),     (5),     (12)
            (2)     (4),      (7)     (8),     (6),    (10),
                     (5)               (9)     (8),     (11)
                                               (10)

Median     0.75     0.75     0.81     0.88     0.63     0.88     0.75

Std.       0.22     0.23     0.22     0.22     0.20     0.23     0.26
dev.

Min        0.25     0.25     0.25     0.25     0.25     0.13     0.13

Max        1.00     1.00     1.00     1.00     1.00     1.00     1.00

Range      0.75     0.75     0.75     0.75     0.75     0.88     0.88

        Ro_LED,  Ro_LED,
        6000 K,  3000 K,
        600 lx   600 lx

Mean     0.54 *   0.68 *
           (2),     (13)
           (4),
           (7),
           (9),
          (11),
          (12),
           (13)

Median     0.56     0.75

Std.       0.29     0.26
dev.

Min        0.00     0.13

Max        1.00     1.00

Range      1.00     0.88

* Statistically significant difference (P < 0.05) between
mean values with same suffix.

TABLE 12.
Mean Values, Medians, and Standard Deviations of
Illuminance Level Sufficiency Results

Question F: Is the Illuminance Level Sufficient for Typical
Office Working Tasks?

0 = Completely Insufficient, 1 = Completely Sufficient

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000
                                                        lx

Mean     0.86 *   0.59 *   0.92 *   0.83 *   0.56 *   0.93 *
           (1),     (1),     (5),     (7),     (2),     (6),
           (2),     (7),    (11),    (14),    (13),    (12),
           (3),     (8),    (13),    (18),    (18),    (19),
           (4),     (9),    (14),     (19)    (19),    (22),
           (5),    (10),    (15),             (20),    (23),
            (6)    (11),    (16),             (21),    (24),
                    (12)     (17)              (22)     (25)

Median     0.88     0.63     1.00     0.88     0.63     1.00

Std.       0.17     0.26     0.14     0.20     0.26     0.15
dev.

Min        0.13     0.00     0.38     0.25     0.13     0.25

Max        1.00     1.00     1.00     1.00     1.00     1.00

Range      0.88     1.00     0.63     0.75     0.88     0.75

        Ro_LED,  Ro_LED,  Ro_LED,
        4500 K,  6000 K,  3000 K,
        600 lx   600 lx   600 lx

Mean     0.79 *   0.74 *   0.76 *
           (8),     (3),     (4),
          (15),     (9),    (10),
          (19),    (16),    (17),
           (23)    (20),    (21),
                    (24)     (25)

Median     0.88     0.88     0.75

Std.       0.23     0.25     0.21
dev.

Min        0.13     0.00     0.25

Max        1.00     1.00     1.00

Range      0.88     1.00     0.75

* Statistically significant difference (P < 0.05) between
mean values with same suffix.

TABLE 13.
Mean Values, Medians, and Standard Deviations of Light
Amount Pleasantness Results
Question G: Do You Think the Light Amount Is Pleasant in the Room?
0 = Way Too Dim, 1 = Way Too Much Light, 0.50 = Optimal

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000
                                                        lx

Mean     0.55 *   0.42 *   0.67 *   0.55 *   0.37 *   0.67 *
           (1),     (1),     (3),     (5),     (2),     (4),
           (2),     (5),     (9),    (12),    (11),    (10),
           (3),     (6),    (11),    (16),    (16),    (17),
            (4)     (7),    (12),     (17)    (18),    (21),
                    (8),    (13),             (19),    (22),
                    (9),    (14),             (20),    (23),
                    (10)     (15)              (21)     (24)


Median     0.50     0.38     0.63     0.50     0.38     0.63

Std.       0.19     0.21     0.20     0.18     0.15     0.20
dev.

Min        0.13     0.00     0.38     0.25     0.13     0.25

Max        1.00     0.88     1.00     1.00     0.88     1.00

Range      0.88     0.88     0.63     0.75     0.75     0.75

        Ro_LED,  Ro_LED,  Ro_LED,
        4500 K,  6000 K,  3000 K,
        600 lx   600 lx   600 lx

Mean     0.55 *   0.52 *   0.54 *
           (6),     (7),     (8),
          (13),    (14),    (15),
          (18),    (19),    (20),
           (22)     (23)     (24)

Median     0.50     0.50     0.50

Std.       0.19     0.20     0.21
dev.

Min        0.25     0.13     0.13

Max        1.00     0.88     1.00

Range      0.75     0.75     0.88

* Statistically significant difference (P < 0.05) between
mean values with same suffix.

TABLE 15.
Mean Values, Medians, and Standard Deviations of Lighting
Pleasantness Results

Question I: In General, Do You Think the Lighting Is Pleasant?
0 = Really Unpleasant, 1 = Really Pleasant

        Sq_LED,  Sq_LED,  Sq_LED,  Re_T5,   Re_T5,   Re_T5,
        4000 K,  4000 K,  4000 K,  4000 K,  4000 K,  4000 K,
        600 lx   300 lx   1000 lx  600 lx   300 lx     1000
                                                        lx

Mean     0.74 *   0.63 *   0.73 *   0.74 *   0.58 *   0.75 *
           (1),     (1),     (5),     (7),     (2),     (4),
           (2),      (4)      (6)      (8)     (5),     (9),
            (3)                                (7),    (10),
                                                (9)     (11)

Median     0.81     0.63     0.75     0.75     0.63     0.88

Std.       0.23     0.24     0.24     0.21     0.23     0.24
dev.

Min        0.13     0.00     0.25     0.38     0.25     0.13

Max        1.00     1.00     1.00     1.00     1.00     1.00

Range      0.88     1.00     0.75     0.63     0.75     0.88

        Ro LED,    Ro    Ro_LED,
        4500 K,   LED,   3000 K,
        600 lx    6000   600 lx
                   K, 600
                     lx

Mean     0.64 *  0 54 *   0.66 *
           (10)    (3),     (12)
                   (6),
                   (8),
                  (11),
                   (12)

Median     0.69    0.63     0.75

Std.       0.25    0.28     0.23
dev.

Min        0.25    0.00     0.13

Max        1.00    1.00     1.00

Range      0.75    1.00     0.88

*Statistically significant difference (P < 0.05) between
mean values with same suffix.

TABLE 16.
Mean Values, Medians, and Standard Deviations of
Luminaire Appearance Results

Question J: Do You Think the Luminaires Themselves Look
Good/Stylish?

0 = Really Bad/Ugly, 1 = Really Good/Stylish

          Sq      Sq      Sq    Re T5,  Re T5,   Re T5,    Ro
         LED,    LED,    LED,    4000   4000 K,   4000    LED,
         4000    4000    4000   K, 600  300 lx     K,     4500
        K, 600  K, 300    K,      lx                1000   K, 600
          lx      lx     1000                        lx      lx
                          lx

Mean    0.78 *  0.75 *  0.77 *  0.62 *   0.58 *  0.61 *  0.70 *
          (1),    (4),    (7),    (2),     (1),    (3),    (11)
          (2),    (5),     (8)    (5),     (4),    (6),
           (3)     (6)            (7),    (10),    (8),
                                   (9)     (11)    (12)

Median    0.75    0.75    0.75    0.63     0.63    0.69    0.75

Std.      0.19    0.20    0.19    0.23     0.21    0.27    0.24
dev.

Min       0.13    0.13    0.13    0.13     0.13    0.00    0.25

Max       1.00    1.00    1.00    1.00     0.88    1.00    1.00

Range     0.88    0.88    0.88    0.88     0.75    1.00    0.75

          Ro    Ro LED,
         LED,   3000 K,
         6000   600 lx
        K, 600
          lx

Mean      0.67    074 *
                   (9),
                  (10),
                   (12)

Median    0.75     0.88

Std.      0.24     0.23
dev.

Min       0.13     0.13

Max       1.00     1.00

Range     0.88     0.88

* Statistically significant difference (P < 0.05) between mean
values with same suffix.

TABLE 19.
Statistics of User Set Illuminance Levels and Color
Temperatures on the Work Plane

        User Set  User Set  User Set  Preset Color    User Set
        Minimum   Optimum   Optimum   Temperature     Optimum
          Level     Level     Level                      Color
                                                    Temperature

         Re_T5,    Re_T5,    Ro LED    Ro LED (K)   Ro LED (K)
         4000 K    4000 K     (lx)
          (lx)      (lx)

Mean         326       648       517          4462         4151

Median       309       643       517          4500         4047

Std.         140       215       115          1216          727
dev.

Min           74       121       245          3000         2957

Max          600      1142       738          6000         6037

Range        526      1021       493          3000         3080

        Difference  Difference
          between       between
         Preset T   Optimal Re
         and User   T5 and Ro
          Set T         LED
                     Settings

          Ro LED

Mean           -1%        -13%

Median         -5%        -23%

Std.           29%         28%
dev.

Min           -51%       - 48%

Max           101%        102%
Range

User created lighting settings

Age:     under 20  20-29   30-39  40-49  50-59  over 60 year old
         *         *       *      *      *      *

Gender:  * male
         * female

User created lighting settings

Age:     under 20  20-29   30-39  40-49  50-59  over 60 year old
         *         *       *      *      *      *

Gender:  * male
         * female

Please set your own "ideal" lighting setting with the control
sliders on the table (COLOR, LIGHT LEVEL).

After setting your own "ideal" lighting, please answer the
following questions.

                              9  8  7  6  5  4  3  2  1

1)  Light level     Much too                             Much
    after your own  bright                               too
    adjustment (5=                                       dim
    ideal)

2)  Colour of the   Much too                             Much
    light (5 =      cold                                 too
    ideal)                                               warm

If you could not set the lighting to your liking, how would you
improve it if there was possibility?


Janne Viitanen (1) *, Jorma Lehtovaara (1), Eino Tetri D.Sc (1), and Liisa Halonen D.Sc (1)

(1) Aalto University School of Electrical Engineering, Lighting Unit, Espoo, Finland.

* Corresponding author: Janne Viitanen, E-mail: janne.viitanen@aalto.fi
COPYRIGHT 2013 Illuminating Engineering Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Viitanen, Janne; Lehtovaara, Jorma; Tetri, Eino; Halonen, Liisa
Publication:Leukos
Article Type:Report
Geographic Code:4EUFI
Date:Apr 1, 2013
Words:14596
Previous Article:Using forced choice discrimination to measure the perceptual response to light of different characteristics.
Next Article:Self-report diary: a method to measure use of office lighting.
Topics:

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters