Best Shot Selector, does it work?
There is controversy over whether the BSS feature on the CP950 actually works. Sure, sometimes a sharp picture results when you use it, but is it just random chance? I'm trying to think of experiments to prove this one way or the other, here is what I can conclude so far. I am open to suggestions for other experiments to try! These are all clues to the operation of this feature.
(I now think that BSS DOES work, but not in the way that people might expect. Read on!)
1) The feature does not simply just select the "sharpest" shot from the entire series.
For this test, I pointed the camera at a picture in a frame against a white wall. I adjusted the lighting so that a 1/4 second exposure was used, so that I could get some "blurring" by shaking the camera. I took series of 10 shots with BSS, only holding the camera still for one shot. Here are the results. In every case, the BSS feature picked a blurred shot instead of the sharpest one. Out of 10 shots, I held the camera still for the 4th, 5th, 5th, 4th, 7th, 1st, 5th. The camera would never pick the "sharp" shot if the sharp shot was close to the front of the series. It was my impression that it was selecting one of the last shots.
However, if the single sharp shot is one of the last ones, it will pick it, so the camera does have the ability to select a sharp shot..
#9 sharp, picked it, #10 sharp, picked it. #9 sharp, picked it. #9 sharp, picked it. #10 sharp, picked it.
2) The BSS feature is not random selection.
For the next test, I covered the lens during a series of the last shots. Here are the results. Covered #6-#10 it picked one of the covered shots. Covered 6-10, picked a covered shot. Covered 7-10, it picked a covered shot. Covered 9 and 10, it picked a non-covered shot. Covered 9 and 10, it picked a non-covered shot. Covered 7 to 10, it picked a covered shot. Covered 8-10, it picked #8. Covered 8-10, it picked a non-covered shot. Covered 8-10, it picked a covered shot. Covered 10, it picked a non-covered shot. Covered 10, it picked a non-covered shot. Covered 10, picked a non covered shot.
So you can see that it does seem to pick some shots based on some criteria, and that these shots are not always the last one. The more shots that are covered near the end, the more likely that it will pick one of them. If you just cover the last shot, it will never select the last shot. If you cover the last two, it will usually avoid both of them. If you cover the last three, it will pick one of them about 50% of the time. If you cover the last 4, it will probably pick one of them, etc. I think that this implies that there is a short running series of shots that are compared, probably three.
3) The BSS feature will never select the first shot.
I could not make the BSS feature select the first of the series of shots. For example, I took a series of 4 shots, holding the camera still for the first shot only. The camera didn't select the first shot in 10 trials. I also tried this with just three shots in the series with the same results.
I think that the BSS feature does work, but not the way that you might expect. From the experiments, it looks as if three things are happening.
1) Contrary to what I expected, it does seem to have the ability to select the sharpest shot!
2) It will never select the first shot for some reason, even if only three shots are taken. (used as a baseline for comparison?).
3) It actually does not appear use all of the shots in a long series when making a comparison. Evidence here points to only somewhere between the last 4 and 2 shots being used.
(Re-cap the Proof for #3)
* In a long 10 shot series, it will only pick a sharp shot if it is #10, #9 or sometimes #8. The camera simply would not select sharp shots in positions 1 to 7.
* In covered shot tests, it will avoid covered shots if they are the last 2 or three shots, but will pick a covered shot if the last 4 or more shots are covered).
In other words, the camera seems to keep a running log of the last 2 to 4 shots that you have taken. It may be comparing them against the first shot as a baseline, which is why the first shot is never selected. Nikon expects you to be watching the screen and stop pressing the release when a good shot is captured. When you let up on the button, the sharpest of the shots is selected. My personal feeling is that only the last three shots are used (plus maybe the first one as a baseline).
If you hold the shutter down for all 10 shots that are available in Normal mode, not all are actually used used. In Fine mode, only 5 to 7 shots are available, and the situation isn't quite so bad.
This means, that if you are using BSS, there does not appear to be any point in holding the shutter down for more than 4 (or maybe 5) shots if you can't see what you are taking a picture of on the screen. If you can see, then you should stop soon after you suspect that you saw a "good" shot on the screen. Because the camera will never select the very first shot, you need to let the camera take at least three shots for BSS to have any effect.
(NOTE that the BSS feature seems to act differently on the CP990, it will pick the first shot with firmware 1.1, for example. Don't extrapolate these CP950 results to the CP990).