Meanwhile, I'm going through my photos from the trip Patrick and I took to Disney World in January. The goal for that trip was PACK LIGHT. We shared one suitcase, one laptop case, and one camera (my point and shoot -- more on that later).
And I've discovered something interesting: nearly all of the uncorrected photos have no real blacks. Observe:
Namely, the left side of the histogram; I've had to drag the black level over 5% or so on nearly all of the photos (it actually makes a huge difference in the overall appearance of the photo).
Now, what I think -- correct me if I'm wrong, someone -- this means is that the camera tends to overexpose, as I don't generally have the same drop-off on the white side of the histogram.
I have no idea if that's common with this camera or not -- it got decent reviews when I bought it, though the SD 900 had been released at that point. In point of fact, the SD 900 was not much more expensive, but the 800 had a nice selling feature for me -- a slightly wider-angle lens.
Real wide-angle lenses for SLRs cost a bundle. My dad had some, but of course not for Canon digital SLRs (digital photography was in its infancy in early 2000), so even if I could find an adapter, they wouldn't manually focus -- and with my eyesight, manual focus is difficult.
(Not impossible, mind you, but difficult.)
Anyhow, real wide-angle lenses cost an arm and a leg, but the point and shoot covered a bit more of the spectrum than either the kit lens that came with my Rebel (18 - 55 mm, I think) and, of course, the zoom lens (60ish - 200ish, I think). So I figured that if I was ever in a "oh, I just need a bit more screen real estate" pinch, the P&S would help.
Anyway, I really don't want to have to manually correct every single image that comes out of the camera -- right now, I'm just correcting potential 4 and 5 star images. But I'm wondering if there isn't an exposure compensation trick I can figure out with the point and shoot.
(All of which is moot if I'm interpreting the histogram wrong, but it was a fun thought experiment.)