The observer is presented with two alternating fields: a reference field (frequently a white one) and a variable chromatic test field. It is critical that both fields are the same shape and size. Circular fields are usually the easiest to obtain. The reference and test fields need to occupy precisely the same location and are presented alternately one at a time.
When the reference and test fields are identical it is critically important that when they are alternated they appear as a single and steady field. That means the spatial attributes of each field need to be identical. The reason for these criteria will become obvious in a moment.
The perceptual criterion in heterochromatic flicker photometry is a just noticeable flicker. If the two fields alternate slowly, e.g., around 2 Hz (cycles per second) the observer will first see one color and then the other. As the frequency of alternation increases the colors will begin to fuse. For example, if a red and white field are being alternated they will fuse to a pulsating pink. If the frequency gets very high, e.g., around 35-40 Hz then the pulsating flicker will disappear unless the radiance differences between the test and reference fields is very large. However, when the frequency of alternation between the test and reference fields is about 15-20 Hz the perception of flicker will disappear for a very narrow range of chromatic field radiance. The observer is asked to adjust the amount of chromatic light until a threshold amount of flicker is perceived. There are two settings that will yield this threshold flicker: 1. when the chromatic field is slightly greater than the reference and 2. when it is slight less. The experimenter typically averages this small range of settings to determine the criterion response.
If there there are significant differences between the test and reference fields other than spectral and brightness differences the minimum flicker judgement can be quite difficult. For example suppose that there are differential spatial inhomogeneities in the two fields. Suppose that there is a finger print or spot of dust in field but not the other. Such spatial inhomogeneities will cause the perception of flicker that has nothing to do with the criterion response which is supposed to be solely dependent on equating the amount of light in the test and reference fields.
Let us now assume that the fields are both absolutely homogeneous, of identical size and shape. Also assume that when one field is turned off at precisely that moment the other is turned on. It would seem we have set the stage for being able to adequately perform heterochromatic flicker photometry with no artifacts to complicate things. But, indeed there is one factor we have not yet considered, chromatic aberration. The link to chromatic aberration shows the problem with regard to glass lenses. In fact, the human eye lens also exhibits chromatic aberration. Consequently, for example when flicker photometry is done between a white reference field and a blue test field, these two fields will not be imaged identically on the retina. Consequently the observer will perceive these two fields as differing in size and an artifacts flicker will result at the outer edge of the stimuli. For the time being let us assume that chromatic aberration is handled with an achromatizing lens. We will deal with this concept in another place.
The Actual Experiment
Let us assume that we alternate the test and reference stimuli at 17 Hz. The reference field luminance is set to about 100 candles (an amount of light to which the cone receptors respond very well). The observer's task will be to adjust the amount of chromatic light so that he (she) perceives a threshold amount of flicker. The experimenter will note the radiance of the chromatic light that produces this flicker threshold and places a new wavelength in the test field. And so goes the experiment until all wavelengths have been tested an adequate number of times.
Plotting the Data
Since we want to plot a spectral sensitivity function derived by heterochromatic flicker photometry the dependent variable (radiance of the chromatic field that gave minimum flicker) has to be divided into 1.0. This gives us the reciprocal of the radiance which is the standard way of representing psychophysically derived sensitivity. When this reciprocal of radiance is plotted as a function of wavelength one has a photopic spectral sensitivity function.
Table of Contents
Table of Contents [When not using frames]