hkr4627@acf4.UUCP (Hedley K. J. Rainnie) (02/20/85)
An opinion about automatic cameras:
I bought a Nikon FG some time ago and blasted happily away on Program Mode
for several rolls of film. After this, I realized that program mode (as
well as aperture and shutter and whatever) is good for snapshots but little
else. The pictures that automatic cameras will screw up on are usually
those which are interesting--typically because of unusual lighting. The FG
also has TTL (through the lens) flash with the SB16B flash. This screws up
more or less like program mode despite the fact that the SB16 is a great
strobe.
The real problem behind both is that through the lens metering is inherently
innacurate. The camera will select different exposures for a subject wearing
a white blouse and one wearing a black blouse. In the first case, the face
will probably be underexposed and overexposed in the latter case.
The reason behind this is that exposure should vary according to light, not
reflectivity of the subject. What one should measure is the light falling
on the subject rather than the light being reflected. So skip the in
Camera meter and use an incident light meter. My Sekonic 518 Digipro also
measures flash. All of my pictures are now taken on full manual and there
has never been one bad exposure.
One bad thing is that you now depend on the aperture markings on the lens.
On variable aperture zoom, or some macro lenses, you will never know for
sure what the aperture is. Filters also present a problem.
Rei Shinozuka
ihnp4!cmcl2!acf4!hkr4627
"The best camera is an FM2 without a battery. "
p.s. The Sekonic really eats through batteries
p.p.s. The only rotten thing about the FM is that shutter speeds do not change
smoothly, unlike the FG.
ron@wjvax.UUCP (Ron Christian) (02/22/85)
I disagree with the statement that incident light measurement
gives a more accurate exposure. This does not take in account
the reflectivity of the subject. If your model's face is in
shadow, it will be consistantly underexposed if you rely only
on measurement of incident light. The best way, in my opinion,
(Get that disclamer in there! :-)) to figure exposure is to meter
off the surface you want to be exposed correctly. In the case
of portrature, this usually means metering on the light reflected
from the face. Even skin tone (race, degree of tanning) can affect
exposure.
It seems to me that incident light measurement has the same problems
you'd have in automatic mode. That is, it's the best possible way to
get an average exposure. I personally am usually not happy with average
exposures.
Now, on camera metering accuracy: When I first got my hand held
meter, I was shocked to find that it disagreed with my camera on
the order of 2 stops or so. But the key here is to make sure you're
metering off what you think you're metering off of. My hand held
meter takes a very average exposure, where the Nikon FE is center
weighted. I found that if I got close enough (without casting a
shadow) to my subject with the hand held, it gave the same reading
as the camera within 1/2 a stop. I suspect that a spot meter would
agree with my camera at the shooting distance.
I also do a lot of shooting in manual mode, but for a one time shot
it's not necessary. Most of the good cameras have what's called
a 'memory lock', which is a quick-and-dirty way to zero in on a
particular surface. You get close to the correct surface, (see
first paragraph) press the memory lock, and hold it in as you back
off to the correct distance and take your shot. I've had good luck
with this, although if I'm taking more than one picture in the same
setting I switch to manual. The Nikon FA, FE and F3 have memory locks.
I don't know about the FM and FG.
I shoot about 20 rolls a month, about 1/4 on aperture priority, the
rest on manual. I use the FE meter exclusively, except when I'm using
my old Nikkor Non-AI lense in bright light. I've had good exposures
and bad exposures. It depends on how much thought you put into the
shot. I don't think there is any one technique for always getting
the 'right' exposure.
--
Ron Christian (Watkins-Johnson Co. San Jose, Calif.)
{pesnta,twg,ios,qubix,turtlevax,tymix}!wjvax!ronhkr4627@acf4.UUCP (Hedley K. J. Rainnie) (02/23/85)
Assuming that you are using a film with any sort of exposure latitude, you should NOT adjust exposure due to the model's complexion. If she has dark skin, you do not want to increase exposure until her skin becomes lighter. If you are taking pictures of a white wall, you do not want to decrease exposure until it looks the same as a black wall. An in camera meter will attempt to do this. Reflected light is a component of at least two variables. Light intensity and reflectivity of the subject. The in camera meter must guess at the values of the components. It does so by assuming that every object it photographs has precisely the reflectivity of an 18% grey card. If the subject has this same reflectivity, fine. Otherwise, you will have to manually compensate... "Hmm does that green dress look a half stop or a full stop brighter than 18% grey?" As for the model being in shadow, the incident meter should be placed in the same shadow as her face, thus yeilding accurate exposure. Most incident meters do average between shadow and highlight by way of a plastic bubble over the sensor. However, replacing the bubble with a disk and pointing it alterately towards the lightsource and away will yeild accurate highlight and shadow readings. Though I am an amateur photographer, all of the professionals I have known rely heavily on the incident light meter--so heavily that most admit that they are completely helpless without it. Now, I am not going to say for a moment you can get perfectly exposed shots using any method unless we are being purely technical. There may be times when you want to under or overexpose a shot deliberately. I happen to feel that an incident light meter will give you a "technically correct" exposure which you may then alter to suit your needs. Incidentally: The FM does not need an exposure lock since it is a completely manual camera. Rei Shinozuka ihnp4!cmcl2!acf4!hkr4627