Julian Loke wrote:
> I have read some interesting arguments about this topic, but haven't
> yet formed an opinion. Did you see these links:
> http://www.richardhess.com/photo/18no.htm
> http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=000hry
Because the calibrating light source for a reflected-light meter normally
is a transilluminated surface (e.g., a Koritsu tester), it's debatable
whether a reflected-light meter is "calibrated" to any particular
reflectance. It is possible to examine the calibration equations for
incident- and reflected-light meters and predict how readings from the two
might compare.
Meter calibration equations are given in ANSI/ISO 2720-1974 (R1994),
General Purpose Photographic Exposure Meters (Photoelectric Type)--Guide to
Product Specification [which replaced ANSI PH3.49-1971]:
A^2 LS ES
2^EV = --- = -- = --
t K C
L = (rho) E
For a flat, diffuse reflector,
(xi) = PI (rho)
= PI L/E
Where
EV = Exposure Value (from the APEX system)
A = Relative aperture (f-number)
t = Shutter time
L = Luminance
E = Illuminance
S = Arithmetic ISO film speed
C = Incident-light meter calibration constant
K = Reflected-light meter calibration constant
rho = Luminance coefficient
xi = Reflectance
PI = 3.14159...
ANSI/ISO 2720-1974 assumes that all quantities are given SI units. For an
incident-light meter with a flat (cosine) receptor, recommended values for
C are 240 to 400; 250 seems to be the choice for most manufacturers. For
reflected-light meters, recommended values for K are 10.6 to 13.4; Canon,
Nikon, and Sekonic use 12.5, while Minolta and Pentax use 14 [the
difference is about 1/6 step]. Using C=250 and K=12.5 (this IS the EOS
list ...), the "assumed" reflectance is
(xi) = 3.14159 * 12.5 / 250 = 0.157,
so that we might expect a reflected-light reading of flat surface with a
reflectance of 15.7% to agree with an incident-light reading in the same
light. Does this suggest that a Canon meter (excluding evaluative mode) is
calibrated to 16%? In a sense, perhaps. I'm not sure that knowing this is
of any great value, however.
Note that this is NOT necessarily what might be expected from an
incident-light meter with a hemispherical (cardioid) sensor. Though most
meters incorporate a hemispherical sensor, and the hemispherical sensor
probably is better for readings of 3-dimensional objects, those readings
are not easy to relate to readings made with a flat sensor (at least I
don't know how to do it ...)
The recommended values were developed years ago, using broad-angle
(equivalent to full-frame averaging) reflected-light meters.
Several publications suggest that the average overall reflectance of an
outdoor scene is closer to 13% than to 18%, although Kodak insist that
their gray cards still are 18%.
None of the above says anything about spectral sensitivity, which does seem
to vary among manufacturers. The issue isn't directly addressed by the ISO
standard, although the standard does call for filtration to correct a CIE A
illuminant (2856 K) to 4700 K. Most meter and camera manufacturers seem to
use an uncorrected A illuminant, however.
Jeff Conrad
*
****
*******
***********************************************************
* For list instructions, including unsubscribe, see:
* http://www.a1.nl/phomepag/markerink/eos_list.htm
***********************************************************
