Originally posted by dms A cloudless stretch of blue sky with the sun above 45degrees and not near the area measured, should match/be very close to the sunny 16 rule. Indeed this is the usual way to calibrate a (reflected light) light meter. The reason it works is because the blue sky is a midtone.
I am not meaning to directly contradict what you said here, but it is important to understand that "Sunny 16" is a method of estimating exposure and is not directly related to conventions for metering, though the two definitely draw from the same well.
For the sake of most readers, the differences may be summed with a set of bullet points:
- Film/sensor sensitivity (ISO or EI) numbers are determined by the response of the medium to light at various intensities
- Meters measure light and translate the measurement to suggest EV settings for a particular film or sensor sensitivity
- Meter calibration is related to the response curves used to assign ISO speeds
- Sunny 16 assumes light at a particular intensity to derive useful EV settings in the absence of a meter
- Meter readings under of certain known light values (18% gray, green grass, northern sky) under Sunny 16 conditions will return an EV similar to that produced by Sunny 16
The Wikipedia article on Film Speed has an extensive discussion summarized in the section that deals with both metering and Sunny 16 and continues to a discussion of Exposure Index, the practical alternative to ISO sensitivity.
Film speed - Wikipedia
Below is additional discussion from my understanding of this stuff.
I find it useful to think of "Sunny 16" as simply assuming that a scene will be well-served by an EV series of shutter speeds for f/16 where the shutter speed is 1/(ISO film speed). The convention is based a particular light regime and the expected response (characteristic curve) for commonly available films for pictorial photography. That it works is a happy coincidence. Note that it starts with two assumptions and makes a third regarding results on the negative. It depends heavily on the second (film response) and even heavier on the fourth that I have not yet mentioned. The subject is assumed to not include large areas at the extremes of the film's dynamic range. For things like snow scenes and white beaches for example, Sunny 22 might be more appropriate. As noted earlier in this thread, there are tables for those special cases.
Metering, OTOH is simply measuring either reflected light from the subject or incident light striking the subject and calculating an EV based on the film speed. The only assumptions are that the film will cooperate by capturing the full range of light values present on the subject for that EV and that the meter was reading an appropriate sampling of light.
So why does a reflected meter reading of a green lawn or the northern sky at midday give an EV similar to Sunny 16? Probably because an incident reading of midday sun on a clear day satisfies the first of the basic assumptions of Sunny 16 only this time we actually measured it. The tonal light values of the green lawn or northern sky are about the same, perceptually, and the film response is designed to build similar densities on the negative for each. Whether that value is near black or near white for a given EV depends on the film speed. The expected values for grass and the northern sky are similar to the value for an 18% gray card, hence the tie-in for EV calculated from incident readings.
Yes, I said films because it was with film that "Sunny 16" was worked out and it is the response to light from those films that allows it to work. Aren't digital sensors the same? The answer is complex, but the short and happy version is that yes, they are basically the same for digital cameras whose sensor response sensitivity is 18% gray traceable (CIPA Standard Output Sensitivity or SOS). Those cameras include those bearing the Pentax brand label. For cameras using other methods, YMMV.
I don't know about the rest of you, but my brain is overfilled.
Steve