The Effective Luminous Intensity (ELI) of two successive flashes for various durations and the intervals between the two flashes have been investigated. As the flash source, electodeless lamps and a circular bright dot with an angular diameter of 1'50" were used. The durations of each flash were varied between 5 and 500 milli-seconds for 6 steps and the intervals between the two flashes were changed between 1 and 500 milli-seconds for 10 steps. Four observers aged from 24 to 32 participated in the observations. The brightness of the flash lights observed were with the threshold level or one of three suprathreshold levels.

As a result, it was found that:

1) For flashes with a duration of 5 milli-seconds with a threshold brightness, the threshold value started to increase as the intervals increased from about 5 milli-seconds. This did not agree with Bouman's results (1952) which started to vary from about 60 milli-seconds.

2) For flashes with one of three suprathreshold brightnesses, the ELI started to decrease as the intervals increased from about 10 milli-seconds irrespective of the durations. The ELI increased to a stable level at a duration of longer than about 500 milli-seconds, irrespective of the brightness. The two flashes were perceived as a single flash for intervals shorter than about 100 milli-seconds, irrespective of the brightness

Vector quantization is used as a method to compress information such as many points in a given space, into reduced representative points.  Each of the representative points has the least distortion from the other points.  Further, Voronoi partition technique is used to divide the space into segregated small spaces.  Representative points represent the points in the segregated small spaces.  This report proposes a new method for improving the vector quantization so as to make an approximately equal number of points to comprise each area.  One of the applications of this vector quantization method is the evaluation of illumination environment.  In this case, points represent positions where human beings (occupant: in this report, they are students ) exist at a sampling time in an illuminated space.  The general evaluation of illumination value is obtained from the measured values of uniformly divided points in the space.  However, this proposed method is comprised of specific measured points corresponding to the positions where human beings exist.  Therefore, this evaluation is a better method for the evaluation of illuminating environments for human beings rather than more traditional uniformly spaced illumination measurement methods.

Use of the Monte Carlo method to simulate sky luminance and daylight illuminance has led to the following findings.

(1) In the case of a sky with a uniform cloud cover, the sky luminance distribution approaches that in the CIE standard overcast sky as the optical thickness of the clouds grows larger. If the sky luminance (Lm) is given by the equation with clear sky luminance (Lc) and overcast sky luminance (Lo), that is, Lm=e^-AxĄLC+(1-e^-Axt)LO (where t is the optical thickness of the clouds), coefficient A will have a value of 0.16-0.28 when the sun altitude is 30-60 degrees. On the other hand, if the daylight illuminance is given by equation Sr=S0'e^-Bxt, coefficient B will have a value of 0.012 - 0.027. The optical thickness of clouds in a heavily clouded sky is 52-82 when the sun altitude is 30-60 degrees.

(2) In the case of an intermediate sky with separate clouds, the sky luminance Lm is represented by equation Lm=(1-C/10)LC+(C/10){e^-AxtLC+(1-e^-Axt)LO},  where C is the total cloud amount. With an increase in the total cloud amount, the daylight illuminance averaged on the overall surface decreases, but this is not always so for the maximum and the minimum.

It is well known that perceived color is not determined by the photometric values alone but instead depends on the surround pattern and the adaptation condition of the eye. It is important that we know how color in complex patterns appears because the scenes we usually see are complicated.  In this study, we estimated the surround effect for the appearance of achromatic colors with two and four surrounds using a new matching method that gives us the equivalent luminance of the simple uniform surround. The magnitude of the surround effects of plural stimuli can be qualitatively described by the equivalent luminance.  The results show that the average luminance of surround stimuli cannot explain the total surround effects of plural stimuli, and that the spatial additivity of surround effects depends on the stimulus conditions.  In addition, some characteristics of the surround effects differ between conditions with a gap and conditions without a gap.  The spatial additivity of the surround effects in the present study can be expressed by the summation of the luminance values of each surround weighted by a function of the luminance of the central patch.

We have established a visualization system for spatial illuminance, in order to visually recognize the three-dimensional luminous condition in a lighting space. It calculates illumination vector and mean spherical illuminance at every point on a regular grid, and displays them graphically with perspective. This paper describes the simulation and analysis of the luminous environment of a soccer ground and an interior, including the spherical presentation of mean spherical illumninance distribution, color contours of mean spherical illuminance distribution on a certain 2-D section, perspective display of pointing arrows, and isosurface of mean spherical illuminance. These readily understandable presentations enables us to lay the foundations for applying spatial illuminance to the analysis of lighting conditions.

A visual search task, such as the detection, discrimination, and separation of a target in a display, requires a display layout that has been properly designed, taking account of color and luminance discrimination abilities. Previous investigations on visual search, however, have focused on fundamental analysis rather than practical application. This paper describes (1) color matching properties from the viewpoint of the practical use of complex backgrounds, and (2) conspicuity expressed in quantitative terms of visual search times and the color differences between the target and the background. Such an expression could help determine the maximum number of colors that could be simultaneously presented to get the best conspicuity in a visual display. This paper shows that complex backgrounds decrease the precision of color matching considerably. Color and luminance discrimination data obtained in earlier fundamental investigations may therefore not be applicable to practical applications. This paper also shows that because the presence of a large number of colors surrounding a target increases the noise level and slows a visual search, the maximum number of colors that can be simultaneously shown on a visual display ranges (depending on the purpose of the display) from 15 to 30.

KEYWORDS: visual search, color conspicuity, color difference, color discrimination, color matching