For all the plausible attempts of the investigators in evaluating the color rendering properties of fluorescent lamps on the bases of colorimetric specifications, it is as yet undeniable that they are to be found mistaken in neglecting the variable of luminance, and apparent luminance changes of objects by light sources, we believe, are very important factors determining the pratical acceptability of the light sources. So the authors propose a new method of appraising these properties dealing with the distances between color points of defined 20 color papers under the sample light and of those under the standard light in a 3-dimentional metric color space. Munsell renotation system is adopted as the metric color space, and the interrelation of the steps of the three attributs is assumed to be 0.1V=0.4C≠1H (at C=6). In one of the three methods adopted to calculate the color distances, chromatisity differences of the light sources are compensable by appropriate parallel displacement of the color solid ; and it is concluded that this method gives the most reasonable measure of the color rendering properties of lamps for each color of objects.
It is apparent from the results that the considerable color distortion occurs in colors in the neighbourhood of red under the standard type lamps while the defects are markedly reduced by the de luxe type lamps and a mixed light of fluorescent lamps and incandescent lamps. The arithmetical mean of these distances of color displacements, which gives a single merit figure of the over all color rendition by a lamp, leads to the general conclusion that the de luxe white fluorescent lamp is superior not only to the standard white fluorescent lamp but also to the special de luxe white fluorescent lamp which applies a kind of deep red phosphor.

View full abstract

Generally said, the light distribution from the fluorescent lamp may be assumed as that of a perfectly diffused linear light source with uniform brightness, at least for the end of lighting design. However, in the photometry of the fluorescent lamp, this assumption does not suffice. The brightness distribution of the elementary portion of the lamp is no more uniform but slightly pointed. H. D. Einhorn and J. D. Sauerman, 1946, proposed for this, such a relation as following:
B (α)=B (o) {k+ (1-k) cosα}
where α stands for the angle of emission, B (α) the brightness against α and k some constant proper fraction. The light distribution from a fluorescent lamp, basing on this assumption, may be written as
I (θ)=I_n sinθ {k+ (1-k) sinθ}, [Equation (7) in the text].
Now, in performing the photometry of a fluorescent lamp, because of its large luminous area and its low brightness, we are usually obliged to do it at a so short distance as the inverse square law is not yet valid. According to these two causes, the light distribution observed at the usually available distance should be by far different from the simple approximation
I (θ)=I_n sinθ
The present authors, at first, calculated this apparent light distribution from the same assumption as that of Einhorn and Sauermann, and then expanded it into series, such as
I (ρθ) =2I₁h {kΣ∞0a₁lx^2l+1 (-k) ∞Σ0a_2lx^2l}, [Equation (14) in the text]
where ρ: photometric distance,
I₁: horizontal intensity per unit length of lamp, h: half length of the lamp, x=h/ρ, and aη: some functions of θ only, given in the table 1 in the text. Io is a matter of course, that this expression tends to the equation (7), as ρ becomes infinitely large. The authors, being informed only with the assumption and the conclusion but nothing with the details, concerning to the papers of Einhorn and Sauerman, worked out on them. Afterwards they could read the original papers and found that the authors' work consisted with that of Einhorn and Sauerman in conclusion but had gone into more detail than it.
Comparing the calculated data with the one observed, though the accuracy of the measurement was not sufficient, the present authors concluded that the assumption on the brightness distribution by Einhorn and Sauerman was valid, only as a practical approximation but a fairly good one.
Some authorities offer the relation between total flux and horizontal intensity. The American Standard gives 9.25 at the photometric distance as long as five times the lamp length, while the Japanese Industrial Standard 9.3 at the distance more than four times the lamp length. The present study calls us an attention for using such a rule, that is, the ratio varies not only with the value of k in the above equations but also with the photometric distance. The present authors give an approximate formula, such as the equation (20) in the text, for the factor as a function not only of the value of k but also of the photometric distance. The formula is traced in Fig. 2, and the observed data are plotted in it, as shown in Fig. 4. There are some discrepancies between the calculation and the observation, which require us further studies. Comparison was done with the measurement by C. E. Horn, W. F. Little and E. H. Slater, ETL in U.S.A., 1952, too.
Some considerations on the effects of the occultation and the interflection between lamps, in the case of multiple lamp measurement, are added.

View full abstract