Originally posted by JimJohnson I'm having trouble understanding your explanation Lowell. With the A pin shorted the camera should have no idea whether the lens is in the A position or on a manually set f/stop. If the body and the lens have the same f/stop using aperture priority, how would the camera know to provide any kind of different exposure compensation?
Looking at the actual exposures, my guess is the lens' manual f/stop settings are simply more accurate than the camera's ability to know how far to move the aperture lever to reach a specific f/stop. The 'A' setting shots, at least to me, seem to be slightly underexposed.
Jim, My bad, I was thinking K lens with a pin shorted. What i found in this case was max and min aperture correct, but in the middle it over exposed due to the difference (limiting by the camera) of aperture before the mechanical stop of the lens. with an A lens as you state, clearly it should be correct across the whole range regardless.
As for the way in which an M ( or K ) lens over exposes, i found it to be quite non linear. However it depends on exactly how you are measuring exposure. I do not use a scene, but a uniformly lit block wall, and measure exposure by greyscale value. It is quite accurate, and shows the exposure increasing from wide open to about the 1/2 way point on the aperture range, and then decreasing. it is relitively uniform at about +1 1/2 stops between the middle 3 stops on the aperture scale, so one could propose to set EV comp to -.5 or -.7 and as you suggest fix the errors in post. Its a hard call to make, with the obvious benefit that you do get some measure of automation, plus P-TTL flash.
Flash is to me the more important thing. I don;t really understand why they cut TTL support after the *istD series.