"The Blind Photometer"
Or, Slugging towards a True “Photo” Meter
By Gary Regester, Edit: 2003.02.02
Your photometer reads “How much light there is, right?” Wrong, more or less. The photometer measures only a third of the visual spectrum and is blind to the remaining two/thirds. Now ask any 100 pro photographers whether the sensitivity of their $500 light meter is based on the sensitivity of film or on the sensitivity of the human brain? Careful now, it could be a trick question. And don’t expect any help from your neighborhood meter salesperson either. Answer please?! Every single photographer, all 100 or 100,000, would say, “Film sensitivity, of course!” And they would be 100% wrong! Correct answer is: “The human brain.” The truth is that the “photo” meter is color blind and “sees” only the GREENS of visual light. And, conversely, only the green layer or green channel determines the speed of film or CCD. What, only green?? What lies?? Conspiracy? Start the class action suit! Throw the meter out with the baby?! Bring my Polaroid back here!! Well, have there been lies?? Let us just call it a “commercial oversight” on the part of the lightmeter manufacturers and sellers. Here are some facts about photometrics and, maybe, a solution that would end some of the myths about your photometer.
The facts: The sensitivities of your so-called “photo” meter which you bought to quantify light measurements for photographic exposures (or in terms of foot candles, lumens or lux), does NOT, as its name seems to suggest, have any relationship to the sensitivities of photographic materials used to make photographs. Instead these photometers follow photometric standards which have their basis in the monochromatic human judgment of brightness that occurs only in the narrow spectrum of green (peaks at 555 nanometers between 500-600 nm, properly “Yellow” as in “CYMK”, but Green in the final analysis of the positive photo, as in “RGB”). This means that photometers are color blind and have no sensitivity in the broader visual light spectrum (380 – 720nm) meaning that meters do not see the “R” or the “B” (red and the blue). The modern “photometer” is unable to sense the red or blue components when averaging your subject’s color or the relative red and blue components of your light source needed to correctly expose color imaging materials. Meters, film speed and lumen ratings of light sources - all are based on green, green and green. (See establishing references below.)
Photometric standards were established in 1931 by the Commission Internationale de I’Eclairage (CIE), the international association of lighting engineers, and by turn, are used by manufacturers of your present photometer. (The CIE also set continuous light standards for tungsten at 2800K; sunlight at 4900K and daylight at 6500K – where and why do we have the 3200K and 5500K standards?? Kodak?) The CIE photometric standards are solely based the sensitivity of the human retinal cone vision using a group of humans (a jury decision) whom are asked to make monochromatic brightness matching judgments. In no manner do the so-called photometric standards follow the sensitivities of photographic film. This cone or “photopic” vision peaks in the center of the visual spectrum (555 nm). Illus. 1 shows both the brightness sensitivity of the human retinal cones (photopic vision – the smaller curve) upon which photometrics are based and the brightness sensitivities of the human retinal rod (the higher and more narrow curve, called scotopic vision – note the shift towards the blue [of night?]. This scotopic rod vision is basis of your peripheral vision. Note that scotopic vision is much more sensitive to light, and, curiously, women have a greater rod density than men). Compare illus. 1 with illus.2 which shows the three color stimulus of human color matching functions. Photopic vision utilizes only the central “green” curve and does not include the blue and red curves used by humans to make color matching judgments. Present photometers quantify solely the “photopic” green of human brightness judgment. However, it is the interrelations of the human brain’s color matching functions which are the basis of what most of us experience as vision and these color matching curves are engineering basis of all color imaging materials - be they film or electronic. So why do we use a “photo” meter that sees only 1/3rd of the picture?
Illus. 1 -Brightness Matching Human Illus. 2 -Color Matching Human Sensitivities Sensitivities
The science of photometrics: Overview- the references that follow demonstrate that photometrics is based on the brightness matching functions of human cone vision. This photopic vision is the basis for the definition of the light output called “lumens”, the measurable unit of luminous flux. And it is this lumen curve - a brightness matching monochromatic human factor - upon which all photometers have their engineering basis (or perhaps better, bias).
a. Human brightness matching is the basis of “photopic vision”:
“The Commission Internationale de I’Eclairage (CIE) is the international authority responsible for standardizing the relative spectral sensitivity functions for human vision and the units of light measurement. Two primary curves for human spectral sensitivity have been standardized by means of the psychophysical techniques of brightness equivalence matching as a function of radiance and wavelength. The CIE adopted the CIE standard photopic observer in 1931 for the fully light-adapted human observer and the CIE standard scotopic observed in 1951 for the dark-adapted human observer. The derived functions are the relative photopic luminous efficiency function (Vl) and the relative scotopic luminous efficiency function (V’l). These two functions form the basis for radiometric to photometric unit conversions. These spectral functions (Vl and V’l) have significant intersubject and intrasubject variability depending upon factors such as age, health, and the adaptation state of the human visual detector. The physiological basis for these internationally adopted functions are the two broad classes of photoreceptors in the human retina, the rods and cones. The photopic (Vl) curve has maximum sensitivity at 555 nm in the yellow-green region of the visible spectrum.”
-Philip Hughes et al., “Optical Radiation”, Pineal Research Review, No. 5, 1987, pg. 8.
b. For photometric purposes, photopic vision is the basis of “lumen” measurement:
“Luminous flux (luminous power) is the quantity derived from radiant flux (radiant power) by evaluating the radiant energy according to its action upon a selective receptor, the spectral sensitivity of which is defined by a standard luminous efficiency function.
“Unless otherwise indicated, the luminous flux relates to photopic vision.
“The unit of luminous flux is the lumen (lm).”
-Table I (4.1) Basic Photometric Quantities, page 787, Wyszecki & Stiles, “Color Science”, Wiley, New York 2000 Definition adopted by CIE in 1970
c. That “lumen” is monochromatic:
“The lumen is the luminous flux of monochromatic radiant energy whose radiant flux is 1/683W and whose frequency is 540x1012 Hz. This frequency corresponds to wavelength ld = 555.016 nm, in standard air.”
-page 256-7, Wyszecki & Stiles, “Color Science”, Wiley, New York 2000
d. Photometric meters are “corrected” to photopic vision:
“Broad-band photometry . . . involves as its main hardware component a photometer which consists of either a thermal detector or (more commonly) a photon detector whose relative spectral responsivity has been modified, often referred to as “corrected,” to approximate the [photopic] V(l) function. Typically, this approximation is accomplished by means of a specially designed combination of absorption filters with appropriate spectral transmittance characteristics. The procedure of designing a [photopic] V(l) filter is similar to that used in designing tristimulus-filters for colorimeters.
-page 256, Wyszecki & Stiles, “Color Science”, Wiley, New York 2000
e. the accuracy of photometers are limited when different light sources are compared:
“The detector output signals are usually calibrated by means of lamps that have been standardized in terms of the quantity the photometer is designed to measure. Modern photometers can be very precise and reproducible, but their accuracy is often limited when lights of dissimilar spectral radiant power distributions are intercompared. The deviations from accurate measurements are usually caused by residual differences between the actual spectral responsivity of the photometer and the standard [photopic] function.
-page 257, Wyszecki & Stiles, “Color Science”, Wiley, New York 2000
Confirmation of the points above in Minolta sales brochure for Industrial Light Meter T-10
As shown above, Minolta has “corrected” its photo sensors to match the photopic vision curve to “within 8%”. Certainly this is good enough for the comparison of brightness in retinal cones of human beings, but not particularly useful for precise determination of exposures of color imaging materials.
Photopic vs. photo. So what’s wrong with measuring light with a human sensitivity standard rather than an imaging sensitivity standard? The first thought is that as the photopic standard represents only center third of the visual spectrum, then a predominately blue or red subject would be incorrectly measured with a green biased meter. True, but more importantly, the unbalanced light sources which may have plenty of green, but are significantly too weak or too stronger in blue or red, are incorrectly measured – prime example is tungsten. (fyi- 5500K has evenly balanced Red, Green and Blue components.) An exposure meter would not reveal that tungsten light has nearly no blue and too much red (and IR) and that overcast sky has more blue with less red than the balance needed to correctly expose “daylight” film. Thirdly, films and electronic medias have their own peculiar sensitivity curves which have no direct relationship to the human photopic standards– for example, the CCD has a very weak blue sensitivity while tungsten balanced films have cheated up the blue sensitivities to compensate for the lack of blue in tungsten light. You may have wondered why the ASA/DIN was always less than Daylight film. Or why the blue digital channel always has comparative more “noise”. Compare tungsten with overcast sky; daylight film and CCD sensitivities-
Note in the spectral distribution graphs below, that at the green peak of photopic vison (555nm), the two light sources are identical, but that the amounts of comparative reds and blues are vastly different. The photometer would have seen only the similarity of the green and is blind to the differences between the reds and blues.
Note in the following two spectral sensitivity graphs, daylight film and the CCD, that again the meter reference would be only the green sensitivity (ASA speed), but would not take into account the CCD’s comparatively weak response (on the right), especially to blue and also red.
Towards a better PHOTOGRAPHIC Meter: A PHOTO meter should have applied three color sensing technologies (at this moment, probably CCD or CMOS) to quantify light in a way to be more similar to the three colors components of the accepted models of human color vision. Basically, an RGB color meter joining to an exposure meter. Such an “photochroma” meter would be able to measure and indicate the relative differences in the red, green and blue components, both the light reflected from the image subject and the incident light from varying light sources, especially spectrally discontinuous discharge lamps such as fluorescent, mercury vapor and carbon arc (HMI). Then, these interpolation of the red, green and blue components of an incidental light source with subject reflectivity can be compared with the known values of red, green and blue sensitivity of various imaging materials (such as photographic films, CCDs and future materials), this invention would determine optimal exposure, color correction and other necessary information to optimize the imaging process.
The results of existing photometers are based on a simple three point relationship - the amount of the green output of the light source, with the amount of the green reflectivity of the subject, with the green sensitivity of the imaging material. The improved meter would increase the interrelationships by a factor of 9 times (illus.6). The green (yellow) bias of existing meters is acceptable only if the relative amounts of the red and blue components are similar to that of the amount of green (yellow), but such meters cannot make interrelated measurements between unequal light source, predominate subject color or unbalanced sensitivity of imaging medias (such as CCD and tungsten film).
varying color requirements of photo sensitive materials - As a specific example of color sensitive materials, the improved meter could be programmed to provide the optimal exposure by considering each of the three layers of any of the great variety of photographic films made by different companies for different purposes or electronic medias.
varying light sources- Again, the use of the CCD (or similar electronic imaging device) allows a quantitative analysis of each of the three color outputs of light which then translate into imaging use. Such a meter could also be used to validate color judgments (CRI) or health evaluation (blue light is shown to be more “alerting” for sleep disorders and winter depression) of competing light sources. In Appendix B, find differences of exposure for a digital camera in milliseconds for each of the red, green and blue channels based on differences in light output of eight apparently similar fluorescent lamps of equal wattage. This is an exact example of the use, method and purpose of said invention- the ability to measure three colors, not one.
By correlation of media and light source, an improved meter would compute perfect exposure with color corrections (filters), thus matching a specific light source to the color sensitivity curves of a specific film or electronic media.
example of use: A image maker (photographer, cinematographer, videographer) selects or imputs the emulsion or digital sensitivities. Possible this imput can be upgraded as new films and electronic medium evolve. As example, the following chart is Kodak’s Ektachrom Daylight Balanced Film, ASA 100:
Second, an improved photometer would determine the relative output of the light source. In the following illustration, note the differences between natural Tungstun light and two fluorescent tungsten “equivalents”. Note that the first two examples below would be identical to a conventional “photopic” meter, seeing green only, but are in fact quite different in their relative red and blue output – meaning that reds and blues in the photo would be relatively over or under exposed with a conventional light meter that only “sees” green.
Finally, a third reading is made of the subject, which may be primarily blue or red, neither red or blue is seen by conventional photometers. An improved “photochroma meter would quickly combine all three readings/ setting to make a determination of best exposure and filters for a given medium, the light source and color balance of the subject.
TOO LATE!- Is this not the photometer we thought we invested in? Only to learn it has been a Quixotic dream? But good thinking a little too late. Are photometers not a relic of the past. Bring on “white balance”!
|close this window|