Originally posted by paulh or 3rd party screens such as Katz-eye, etc.
Lowell has reported good results with the *ist screen. I have evaluated my K10D against the Katz Eye (both with and without Optibrite) and can confidently report that it does not help the exposure issue with stop-down metering and that screen.
Like Lowell, I have also done extensive testing of my aperture ring lenses. Unfortunately I lost the spreadsheets in a hard-drive failure. I can summarize in a few bullet points:
- The problem exists at apertures wider than f/4 and is characterized by underexposure of 1-2 stops
- The severity of the problem varies by lens and selected aperture. Of those in my "stable" the Jupiter-9 85/2 is the worse with 3 stops underexposure at f/2.
- Exposure compensation is no help unless you always shoot at the same aperture with the same lens since the offset is non-linear and varies by lens
- There is a similar lack of meter linearity at narrow apertures in dim light, but I have assigned that one to lack of meter sensitivity at those light levels. This was a known issue with stop-down meter cameras back-in-the-day and not unexpected.
Solution? Be familiar with your lenses and apply appropriate compensation as needed based on the chosen aperture. Chimp, adjust, and re-shoot, if needed.
Cause? From what I have been able to glean and from the correction of the issue in one mode (M) on the K-5, K-7, and K-r; the K-series digital bodies apply a mathematical correction to the meter reading to correct for behavior of the focus screens at wider apertures. This correction is based on the maximum aperture as communicated through the contacts on the lens base and from the aperture to be set by the body. Both are available when an A-contact lens is detected.
That data is missing when using the aperture ring to control the iris with subsequent default to a baseline program that may or may not work too well. This behavior dates back to the *ist film camera and is documented in the user manuals for affected models. The newer bodies have a modified algorithm that is applied in M mode for stop-down metering and which yields essentially accurate readings with those cameras in that mode only. Use Av mode with your M42 lenses and you are back to the same old issues.
All that being said, the general picture is one of underexposure with the K10D in stop-down mode at wider apertures. The OP indicates the opposite with a general pattern of overexposure. Overexposure such as described generally indicates a lens issue (lens is not stopping down to the set aperture as expected when doing an actual exposure). Regardless, it is usually good to rule out the meter as the source of error and go forward from there.
Steve