A test of meter/exposure calibration between *ist DS and K10D
Text at
http://homepage.mac.com/godders/mCAL/calibration-K10-DS.txt
I've read on DPReview.com many people saying that metering with pre-A
series lenses was off in the K10D. I only have one such lens and
don't use it often, but I decided to give a try to see if I could
find a big calibration difference by using one repeatable lighting
setup and comparing results between the *ist DS and K10D bodies. This
is far from an exhaustive test but was interesting ...
I used an FA35/2 AL and a Zenitar 16/2.8 FE. The FA lens will act
exactly like a pre-A series lens if I unlock the aperture ring from
the A position and set it manually to a specific f/opening, so the
metering should be the same as it would be for the Zenitar. There is
a difference in the way pre-A and A-later series lenses iris
mechanisms are calibrated, however, so I wanted to explore that a
little bit too.
I chose a simple target .. the white wall in my office .. and set up
the tripod and remote release. For each body, I set the exposure to
Manual mode, set the focus to Manual mode, and the ISO to 200. I did
one body at a time, running through the whole sequence of exposures.
The sequence went like this for each body:
For a baseline check, I used a Sekonic L328 incident light meter to
obtain a reference exposure. I started with the FA35 on "A" aperture
ring setting and made the exposures at f/[EMAIL PROTECTED] sec and f/[EMAIL
PROTECTED] sec. I
then switched to f/4 and f/16 with the aperture ring and repeated the
same sequence, to test the alternative way the iris mechanisms are
regulated in the lens' two modes of operation. Next I allowed the
camera body to meter the exposure with the lens set at A for f/4 and
again for f/16. Next again with the same lens, f/4 and f/16 setting
on the lens. And then finally I switched to the Zenitar and did the
same two exposures at f/4 and f/16, allowing the body to do the
metering.
The capture was all in RAW mode and I didn't want RAW processing to
influence exposure results. Files were downloaded to my computer and
opened with Camera Raw 3.6. The same specific point in the image was
set for each frame and the white balance adjusted with the eyedropper
tool to ensure even color rendering. (The ACR defaults render K10D
RAW exposures in this light with a lot more yellow than DS RAW
exposures, but all adjust beautifully to a neutral tone.) I then
captured the histogram readings with all other settings at the defaults.
The results are the histograms shown in this chart:
http://homepage.mac.com/godders/mCAL/calibration-K10-DS.jpg
The left column is the DS, the right is the K10D.
The first four rows show the FA35, first with aperture controlled by
the body setting at f/4 and f/16, and then with the lens' aperture
ring setting f/4 and f/16. These are the manually set exposures from
the external incident light meter. As you can see, they are pretty
regular and even.
The remaining four rows show what the in-camera reflective meter nets
at the same aperture settings ... first two using the body controlled
aperture, the next two using the lens' aperture ring. Note that there
is more variation here between f/4 and f/16, even with the body set
aperture setting, because you're hitting the bottom of the meter's
sensitivity range. Then, in the next two rows, you see that at f/16
the Manual-semi automatic stop down setting really does bottom out
and the f/16 exposure is well underexposed.
The last two rows show the Zenitar 16. Note that the same
characteristic between f/4 and f/16 holds, the latter being well
beyond the Manual-semiautomatic stopdown range, and that the shape of
the spikes is a little different (due to the lens' much wider field
of view).
Looking at the full range of tests and comparing the DS vs K10D
metering with both lenses, I'd have to say that the two bodies are
pretty consistent in how they respond and they're all certainly
within a reasonable range of correct, on target exposure relative to
the reference incident exposures on top. The four f/16 exposures that
are well underexposed are simply beyond the metering range, the other
drop from the reference shows the difference between an incident
light reading and a reflected light reading. Proper exposure with a
reflected light meter on this target would require approximately a
+0.7 to 1.0 EV exposure compensation setting.
There is some variation, but I feel it's all well within reasonable
calibration accuracy.
Godfrey
--
PDML Pentax-Discuss Mail List
[email protected]
http://pdml.net/mailman/listinfo/pdml_pdml.net
