Originally posted by wlachan Since the iso, shutter & aperture were set manually, the metering is a non-issue here. There are 5 possible explanations that I can think of:
1) The actual wide open t-stop of this lens is way smaller than 1.4 (not rare for fast lenses).
2) Mechanial aperture is not linear (and often they aren't).
3) The aperture blade position was not probably calibrated. You can tell by turning the aperture from f2 to f1.4 slowly. The aperture blades should just disappear completely at 1.4, but not too soon.
4) The rear aperture arm was not properly calibrated. You can tell by flipping the rear aperture arm slowly and the aperture blades should disappear just when the arm hits the end, but not too soon.
5) The stop down mechanism of the camera was not properly calibrated. This will affect other lenses too.
PS: Pentax lenses with aperture rings & A setting often disagree with each other slightly. For instance, you can take an identical shot at certain aperture, say f4, one set from the lens, one through the camera. The actual exposure could be different (by up to 1/2 stop based on experience) due to the reasons above. This is not neceassily faulty, just the limitation of mechanical aperture and loose QC for mass production. Fully electronic apertures of EOS lenses can be made more precise for the same reason.
I would have to agree here also with respect to possible causes, BUT.....
I looked at the images and put them side by side,, then did a screen capture to evaluate the overall image in PSP X3
what I can tell you is the following.
There is appearance of more depth of field in the F2.8 shot, than in the F1.4 shot, although to tell if the aperture really stopped down It would have been nice to have one distant Out of Focus light, that would take the shape of the aperture blades, when stopped down.
If we look at the greyscale value of the carpet next to the bottle, in the first shot it is 68 and in the second it is 87. This represents about a 1/2 stop difference. I don't know what you metered on to get the exposure but it does show that there is some error in the aperture.
If you look at the edges of the photos, and again the greyscale value of the carpet the wide open shot shows a value of 51 in the lower left, the F2.8 shot shows 84. This is vignetting and is typical of all lenses wide open, and at F2.8 the difference center to edge is less than 1/10 of a stop or pretty good.
I think the lens is functioning properly, within the tolorances.
What you should do, is take the lens and using spot metering, shoot only the carpet, uniformly lit, and in AV mode, run a set of shots at each aperture setting and plot the center and edge greyscale value on a spread sheet. this will give you your system's exposure accuracy and also the lens performace in terms of vignetting. I have found nearly all lenses will have some exposure errors, for example my tamron 28-75 F2.8 has a drift upwards which ius just about uniformly increasing with aperture,starting with perfect exposure wide open to +1 stop at F32. It is always food to map the performance of your kit so you know how it all works together. The worst time to learn about exposure errors on any part of your system is after you screw up a once in a lifetime shot.