There does [do] seem to be some things left out of the EOS 1D X that photographers are worried about. The biggest one I have seen is the omission of f/8 AF sensors. This means that a 500 f/4L IS with a 2X teleconverter wouldnâ€™t have AF at the center point, like the current 1D and 1Ds cameras. I am still waiting for official confirmation from Canon about this issue. Is it a hardware thing? or could it be addressed with software?
The minimum aperture required for a particular AF sensor pair to function depends on the spacing of the "AF sub-exit pupils" for the two detectors.
At the smallest workable aperture for a particular detector pair, these are located (by design) against the edges of the actual lens exit pupil (as it is during AF, at the lens' maximum available aperture). Their separation is in effect the "baseline" of a rangefinder, which is what a phase detection AF detector system actually is. Each detector of a pair works from light coming only through its own sub-exit pupil.
Thus, the spacing is different for the AF detector pairs with different "minimum aperture" requirements.
This siting is determined by the little optical system that feeds each AF detector pair, which among other things includes a pair of physical apertures. The sub-exit pupils are images of these.
The greater the spacing of the sub-exit pupils, the greater the "precision" of focus detection, which is why the detector pairs that require an f/2.8 minimum aperture are more precise, and are desirable to use (when they will work).
But if the sub-exit pupils are located outside the boundaries of the exit pupil of the lens proper (as when the lens aperture during AF is smaller than that which supports the sub-exit pupil spacing of the particular detector pair of interest), then the AF detectors get no light at all, and thus cannot function.
And, prudently, the camera body declines to include various classes of AF detectors from the AF scenario based on knowledge, reported by the lens, of its maximum aperture.
This is thus a hardware consideration, and cannot be overcome by any firmware fiddling.
Note finally that this is in no way a photometric issue; having "enough light" (in the exposure sense) is not the critical issue here. This is why AF detectors requiring a minimum aperture of, say, f/5.6 cannot be put into play merely by limiting ourselves to "brighter" scenes.
When a 2X extender is placed on a lens which, by itself, has a maximum aperture of f/4, the result is a new lens, with more elements, longer and weighing more, having a maximum aperture of f/8. It is popular to say that the "effective
maximum aperture" is now f/8. But that is the real
aperture of our "new lens". Thus it is the exit pupil produced by that actual new aperture which is the limiting factor on which AF detector pairs will work with that augmented lens in place.