I have now carried out some own, more detailed tests.
First, some of the info I had provided before is wrong. I apologize for this.
(1) EffectiveLV
The first correction is for the EXIF value reported by PhotoME as "EffectiveLV".
It is defined in
Exiv2 - Image metadata library and tools as "
Exif.Pentax.EffectiveLV: Camera calculated light value, includes exposure compensation"
I carried out some exposure brackets and used EV compensation, too. This value doesn't change and is different from the LV used when taking the image!
For instance, study the fourth attachment to this post (I hope that EXIFs stayed intact...):
Taken at ISO100, 1/500, f/8 -> 15.0 LV
(Sometimes, the maker notes would say 1/751.9 rather than 1/750. It seems the shutter speed is measured and deviations are recorded!)
However, the EXIF EffectiveLV is 16.8 LV. Why?
Now, look at AE-Shutter Speed = 1/442.7
This is what the AE recommend and the next rounded available speed was 1/500.
So, sometimes, you must just fall in between speeds 1/2 stop apart and two consecutive images are exposed differently because of this...
AE incl. compensations says at ISO100, 1/442.7, f/8 -> 14.82 LV
Now take +2EV exposure compensation into account, +1EV from EV compensation and +1EV from the bracketing series this is a part of.
AE without compensations -> 16.8 LV
So, now we know what EffectiveLV means:
It is the LV the camera electronics thinks it would be best to take the image at, ignoring EV compensation advice if given
EffectiveLV doesn't change with EV compensation or during a bracketing.
(2) Calibration of metering
I have done a test shot of a sheet of paper (fourth attachment).
As we already know, LV taken with is 15.0 LV.
The white background is at 90% (79% gamma-corrected), or -0.34EV from full white. So, it should be 14.7 LV.
But big question and my second mistake:
Should a subject with LV value N taken at LV N be recorded ar RGB=255?
From
Light meter - Wikipedia, the free encyclopedia I note the sentence: "
ISO 2720:1974 recommends a range for K of 10.6 to 13.4 with luminance in cd/mē. Two values for K are in common use: 12.5 (Canon, Nikon, and Sekonic) and 14 (Kenko and Pentax); the difference between the two values is approximately 1/6 EV."
So, the issue is more complicated and my take at it was pretty naive.
Anyway, the matrix field readings are:
15.6LV (middle) and ranging between 15.1 and 16.1LV. The EffectiveLV of 16.8LV reflects the +1.1EV rule for spot metering still seen to be valid.
The average is 15.6 which is +0.9EV compared to the scene luminance computed above (14.7).
So what it all does mean?
Well, first a short departure into gray cards here:
Meters Don't See 18% Gray by Thom Hogan
It suggests that a 12% gray card, rather than a 18% gray card, should show up in the middle of a histogram. We don't need the argument here.
But the camera meters cannot know the difference between black, gray or white. So, whatever subject they meter, they should always try to take the picture such that it ends up in the middle of the histogram.
So, a subject with light meter reading of X, when taken at x -1EV exposure, should end up near the right edge of the histogram.
So, if we take this into account, our histogram measure of 14.7 should be corrected by 1EV and read 15.7 LV -- pretty close to the actualy meter readings.
Having said all this, we need a break and I provide three more photos for your enjoyment, taken at
- matrix
- mutlifield
- spot (no EV correction)
- the sample discussed above
- a 100% crop from it (because it is so f**g sharp