There has been much discussion of the "inaccuracy" of TTL Auto flash metering.
Some of the problems may not be truly inaccurate metering. Instead, they may be
at least partly a lack of understanding of the physcial limitations of on-camera
flash.
One needs to consider some of the physics and other limitations of this kind of
flash metering. Remember the inverse square law? It applies with a vengeance
here. If you shoot a subject from, say, 4 meters against a background that is
another 4 meters behind the subject the light to the background will fall off by
two stops. In other words, whatever the flash output, the light on the subject
and background will differ by 2 stops! (Ignoring, of course any contribution by
the ambient lighting? more on that later.) Since for all OMs featuring TTL
flash
metering, the exposure measurement is centerweighted, the actual exposure
relationship of the subject and the background will largely depend on the
relative
area of each image within the centerweighted portion of the frame. If the
subject
is smaller than the area of the background, the metering circuits will tend to
pump out enough light to "properly" expose the background while overexposing the
nearer subject by as much as 2 stops. This is what usually happens. You can
help
a lot with a 1 to 1-1/2 stop minus exposure compensation dialed in if you are
willing to have a dark background
What else can you do about this? First off, recognize when it will happen.
Then
there are several approaches to mitigate the problem. Try moving the subject
closer to the background. Look for a more reflective background. Get closer to
the subject (fill more frame with the subject). Move farther away from the
subject. Wait a minute, I just told you to get closer! How does moving farther
away work?
If the subject is still 4 meters in front of the background but you move to 8
meters away from the subject (and use a longer lens to maintain the framing of
the
subject), you are now 12 meters from the background. The background is 12/8 of
the distance to the subject away so the light falls off at the background by 1
over the square root of 12/8 (this is not tough math!). Now the background gets
82% as much light as the subject, easily with in the exposure latitude of print
film and, for most uses, of transparency film. Another approach moves JUST the
flash farther away - by bouncing it off a ceiling or wall. The total distance
traveled by the light from the flash to the subject and background is what
matters
- not the distance the light travels from the subject to the camera. (Isn't
physics fun?)
You can also open up the aperture so that more ambient light is used (still
keeping the shutter speed with in the synchronization range). This helps
because
it reduces the proportion of the exposure due to flash.
When exposure is really critical (Gary Reese's endangered species example is
great) manual flash and guide number calculations (or an incident flash meter)
are
the way to go. Either guide number or flash meters MUST be calibrated by test.
Don't trust spec sheets. For snapshooting and photojournalism with print film,
careful composition and TTL Auto flash can work very well, and is a lot quicker.
For other uses between these extremes, you be the judge.
Gary Edwards
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|