When I first got my EOS-5d, I had a little trouble
with dynamic range - mostly with blown-out highlights but also some
problems with blank shadows. But as my skills with exposure
improved (mostly, adjusting my habits to 'expose to the right' and
careful checking for clipping), many of the problems went away.
Even more, when I switched from using the camera raw conversion
software provided by Canon to Adobe Camera Raw, the problems
vanished almost completely. There are still images where I
struggle a little bit, but they only occur in high contrast
situations.
But recent discussions on Mike Johnston's
The Online
Photographer blog got me thinking about dynamic range again, so
I decided to test my EOS-5d and see what the actual, usable dynamic
range was.
Whenever I'm about to test a camera, or film, or lens, I seem to
turn to just a few handy tools in my testing toolbox. In this
case, based on Mike Johnston's report that the Leaf Aptus22 back
boasts ten stops of dynamic range, I nearly order a new tool - a 41
step Stouffer step wedge with steps of .1 logD. That step
wedge, when illuminated evenly from behind, provides a range of
brightnesses that spans 13 2/3 stops - easily enough for any camera
I'm likely to encounter.
But before I did that, I broke out an old standby - my 31 step
wedge, which provides a range of 10 stops. I figured it would
be enough, so I used it and my flat panel light box for the testing.
All I did was put the step wedge on the light box, turn the
lightbox on, and photograph the step wedge using the EOS-5d and my
100mm f/2.8 macro lens. I adjusted the exposure so that the
first step of the step wedge fell just sort of clipping on the
highlight end of the histogram. ISO was set to 100.
The following image is from that exposure, run through Adobe
Camera Raw with the shadow and highlight settings adjusted to get
the maximum range from the image.
Every step is 0.1 logD, or one third of a stop, and to my eye,
the range runs from step 1 to step 25, so I'll call that 24 steps or
8 stops of range. That's with some adjustment in ACR,
including setting the exposure and shadow sliders and a little bit
of curve tweaking to even things out. Levels has been used in
Photoshop to set the white and black points to full range.
This next image is what I get with some digging with curves
layer. Here, I'd say the range runs from step 1 through step
27 or 28. Call it nine stops - a pretty amazing result.
This isn't a very sophisticated test. It took me about
twenty minutes to do the test, a bit more to write this article up.
The result, though, is pretty much real world. The software
used is the software that's available to most photographers working
digitally. It's possible that I could eke out a bit wider
range by using other, more sophisticated Camera Raw software.
But this is a real world, honest way to test the dynamic range of a
camera - it's an artificial scene but the range of values is
accurate. I haven't used camera settings that are different
from what I'd use in the field.
And I'm more than a bit surprised, actually. If you figure
you can get more than 8 stops of detailed range, and closer to nine
when you're willing to live with a bit more noise, and I'm having a
hard time feeling like this is a limit I can't live with.
