To investigate the problem of light penetration and distribution in moderate sized public library reading rooms, five tests, theoretical and practical, were conducted.
The first test, using the sky projection concept, computed the percentage of the sky component received in the interior for the various seasons and times of day in the North temperate zone. This was an attempt to satisfy the need for a useful tool capable of giving a quick and reasonably accurate result while the design is still in its early stages. At this stage in the design process, the intricate technicality of the illuminating engineers' tables and formulas have proven to be of little use. This study attempts to solve this problem by simplifying the methodology.
Utilizing the data of the C.I.E. simplified table for standard overcast sky, the second test measured the relative influence of the window proportion, position, and area upon the light received at the work plane.
In the third test, a model of a room was constructed to detect the surfaces influencing the amount of the internally reflected component from windows located at one side wall. Under these conditions, the contribution of the ceiling and the rear wall were the decisive factors in the internally reflected component. This knowledge provided the basis for a graphical method for determining, for lighting purposes, the optimum shape of ceiling and back wall in all types of rooms. This method proved to give more accurate, as well as more useful, data for the designer than the theory of integrated sphere or the split-flux concept.
One of the reasons for eye strain and visual distraction is to read in an environment of unbalanced brightness grading beyond the adaptive tolerance of the eye. For this reason, a fourth test was undertaken to find a formula to be used in testing the efficiency of windows in providing balanced brightness in relation to the room size and the depth of the work plane. Data for this formula were obtained from recording the brightness distribution of both the natural landscape and the sky luminance distribution.
The fifth test dealt with the influence of the window reveal shape upon the quantity and quality of light in the interior. The objective was to find a means of improving the window performance in cases where the freedom of enlarging the window is restricted. Windows with three different reveal shapes were chose: reveals beveled to the exterior, square reveals, and reveals beveled to the interior.
Based on the five above tests, the following conclusions are:
1) that tall windows are more efficient in term of light penetration and quality than either square window or the horizontal window of the same glazing area,
2) that the ceiling and the back wall combined have the decisive impact upon the amount of the internally reflected component,
3) that it was possible to obtain a formula for brightness grading by observation and measurement of the distribution of light in the natural landscape. For the three zones the relationship is 1:0.3:0.10 This confirms a previous datum presented by Hopkinson for brightness grading, and
4) that the window with reveals beveled to the exterior introduces a great deal more light than when the reveals are square or beveled toward the interior.