I don't know about No.1, but No.2 is easy to answer. I pointed the camera at a scene where there was a lit table lamp at the center, zoomed to 3.6. The shutter speed on auto was 1/13. When I zoomed in to 7.2, the shutter speed changed to 1/30. Conclusion: in CIZ, the exposure is based on the zoomed-in image.
Bob
Bob,thanks for the reply and for doing the experiment. I was actually wondering which "portion" of the zoomed-in image it uses for exposure: the entire area within the dotted lines, or only the exposure area selected in the menu. In your experiment, when you zoomed to 7.2x, I assume the overall brightness of the frame increased since more of the lamp was filling the frame than at 3.6x zoom. Thus the faster shutter speed when zoomed to 7.2x than when zoomed to 3.6x. But was that because it used the exposure area that you had set in the menu, or because it used the entire area within the dotted lines? Hard to tell - either way would lead to a lower exposure at the higher zoom level. Do you agree, or is my logic faulty?
I used your lamp idea to experiment. I set the exposure area to "Center" in the menu, then zoomed into CIZ territory, composed with the lamp in the center of the frame, depressed the shutter half-way and noted the exposure info. Then repeated this with the lamp at the edge of the dotted line area, (still fully within it - just not in the center). Then repeated again with the lamp just outside the dotted lines.
It was clear that with the lamp in the center, it was using a lower exposure (lower ISO in this case) than with the lamp at the edge of the dotted area, or outside the dotted area. That could have been because it was using the center area for exposure (per my menu selection), and the lamp in the center was bright. Or it could have been disregarding my menu setting and using the entire area between the dotted lines, but that entire area was brighter when the lamp was in the center than when it was at the edge of the dotted lines, or outside the dotted lines.
So I was unable to draw any conclusions. At the moment, I can't think of a better way to test this, so it remains a mystery to me. I'd be interested in any suggestions you (or anyone else) can offer on how to properly test this.