... real world photographic tests are more interesting than studio tests at ƒ 5.6 with a set sensitivity, letting the camera choose its own shutter speed. Different formats get different effects from different lens designs.
Seeing as camera manufacturers are cheating with the sensitivity boosts at lower ƒ-numbers, maybe it's time to check the shutter speeds as well.
The issue affects all formats, and is a function of the discrepancy between the Numerical Aperture:
http://en.wikipedia.org/wiki/Numerical_aperture
and the entrance pupil diameter in terms of light gathering ability, which I worked out here:
http://forums.dpreview.com/forums/read.asp?forum=1018&message=36808367
Other issues
add to this
unavoidable fact of life, such as the microlens efficiency, stack height, etc.
As for sensor sensitivity being boosted, the sensor sensitivity is fixed, and is not a function of ISO. Raising the ISO simply applies an analog gain to the signal, which, in some cameras, is simply more efficient than a digital push:
http://www.josephjamesphotography.com/equivalence/#iso
In other words, an analog gain may be being applied to negate the light loss, but this will simply result in more noise than would otherwise be expected.
So, what is going on is that opening up from f/2 to f/1.4, for example, will not result in a full extra stop of light gathering ability. But this effect also exists from f/2.8 to f/4, albeit not as strongly. So it's not like there's no benefit to using f/1.4 over f/2, or f/2 over f/2.8, etc.
However, this is true
for all systems . In terms of Equivalence, it works in favor of the larger sensor systems at large apertures, since the effect is far less, for example, from f/4 to f/2.8 on FF than it is from f/2 to f/1.4 on 4/3.
