I miss-read this exercise at the start and so have really only done the second part properly although I have chosen pictures which I thought would be darker or lighter than normal to see whether I got the results I expected. It was a really grey day so I used the natural light in the conservatory, mounted my camera on a tripod and used a remote shutter release so that my pictures would be sharp. In each case I used matrix metering and no exposure compensation for the initial shots, adjusting the exposure up or down depending on the results. My first subject was a lily, which I thought was predominantly white. I took a series of close up shots looking into the flower, using aperture priority mode and checking the histogram after each shot.
I had expected the pictures to be underexposed because I thought that the camera would compensate for the light subject. What I found though was that when you actually look a these pictures, there is a lot of pink, green and even brown there, so the subject was not as light as I had originally assumed. I also wonder whether, the fact that I was shooting into the flower diffused the light so it was not as bright.
The only picture that came out too dark was one taken in shutter priority mode on 1/60 sec. This resulted in an aperture of f/3.2 and I can only imagine that this was as wide as the lens would go and it was not enough to expose the image properly.
My second subject was a dark pink gerbera flower against a violet background and was chosen because I thought it would be darker than normal. I pretty much got the results I expected with this one and although several pictures fell within an acceptable exposure range, when I compared the pictures with the actual flower and background, the one which most closely matched the colours was underexposed by 2.67 stops. This is quite dark though, the one I prefer is the centre one which is still under exposed by 2 stops.
I then chose a white lisianthus flower and, like the lily, took some pictures looking right inside the flower. Unlike the lily though, the initial picture, without exposure compensation, and shooting at f/36 and matrix metering, did come out quite dark and I had to compensate the exposure by +.67 stop to get an image I was happy with. Exposure compensation set at +1.33 stops produced an image that was just about ok but there were clear signs of clipping when exposure compensation was increased to +2 stops.
My last two subjects, a dark tray sitting on a darkish cushion and a dark cat ornament, sitting on the same cushion both produced shots which were over exposed when photographed initial, using aperture priority and f/8. In both cases I had to reduce the exposure to -1.33 stops to get acceptable pictures and even then the black tray looks slightly over exposed.
The lesson here is that even when using matrix, or ‘smart predictive’ metering, I do need to check the exposure and adjust accordingly, particularly when the subject is lighter or darker than average. In this initial set of pictures, only those of the lily were effectively exposed using the camera’s own matrix metering system with no exposure compensation.