Originally posted by vinzer To minimize differences between the different exposure modes, I set it to spot metering and pointed at the exact same spot on my test subject. Here are the results.
I don't think your composition is precise enough for a valid comparison. By my eye, the underexposed picture may have the white belly in the center and the 'overexposed' picture may have the black 'arm' in the center. This might cause what you're seeing.
Originally posted by vinzer I noticed that with the shooting modes I tried (P and M), the camera would not give a setting other than shutter=4 and aperture=5.6, even when using the green button in the M mode. For reference, the old 300D I switched from would give a setting with a super long shutter speed, but it did give the proper exposure.
I would expect 'M' with green button to give the same results as P, though there is probably a wrench in the works here. What you may be seing but not paying attention to is Auto-ISO in action. When in 'P', your sensitivity is probably being manipulated by Auto ISO because you're already at max aperture (for tele end of kit lens) and a very slow shutter speed. There is no Auto-ISO at 'M' mode--IIRC, it is fixed to the low end of the auto range. You can verify this at shooting time with the OK button and review it later in the EXIF. So if you had the same fixed ISO setting in P and M, I would expect the same exposure if you use the green button.
As for your 300D or some other older DSLR's, I would expect that even if Auto ISO is enabled, it probably doesn't move in 1/2 or 1/3 EV steps and instead only offers 1EV steps (100, 200, 400, 800, 1600, 3200) so you would be less likely to see the shutter & aperture remaining constant with only ISO adjustments. As for metering differences, it would not necessarily be surprising that given a tricky scene like in your test, different make/model bodies may meter a little differently. Not only are their meters (and possibly lenses) different, but the bodies may implement different program lines, etc.
Originally posted by vinzer I love my K10D, and I'd rather have this fixed than turn back on Pentax altogether, as some would probably do. I just need opinions first if I really did get a dud camera before I make use of the warranty.
Sad that I had to be given a rough entry to the Pentax world, but it happens with any brand. I'm not leaving just yet. Haha.
Your help would be much appreciated.
I am guessing the fault here is a tricky scene and dubious testing techniques. No camera is going to be perfect, and I suspect that in terms of metering, Pentax might be a little less consistent than some others--but not so much that I would consider changing for that reason; any camera will be fooled sometimes. I think you will get more consistent results using the matrix metering and learn to apply compensation or use AE-L as necessary. Spot metering has its purpose, but its use is more of an advanced technique that requires special attention.