Archived posting to the Leica Users Group, 2000/11/20

[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]

Subject: [Leica] Accuracy and reproducibility
From: imx <imxputs@knoware.nl>
Date: Mon, 20 Nov 2000 15:25:24 +0100

Recently there was a discussion about the drop in transmission of light when
using an UV filter. But how accurate can we work in BW photography?
Most products and processes allow for  plus/minus 5% margin.  Let us add a
few steps.
We start with the exposure meter: 5% deviation from the indicated value
versus the true light level. Often 1/10 of a stop and more..
The shutter speed of the Leica M may deviate by 10 to 15%.
The mechanism of the aperture stop is allowed to deviate by 5% from the
indicated value. 
The true filmspeed may deviate by 1/6 of a stop from the nominal value.
The development process itself brings again a deviation of 5% to 10%, As
humans we cannot reproduce the process exactly and time, temperature, mix of
chemicals, quality of chemicals are all parameters that may deviate by at
least 10%.

Mostly the plus deviations in this chain will cancel out the minus
deviations, but even then we have to assume that one exposure of 1/15 sec at
1.4 on ISO400 film may deviate from the same exposure on a second film by at
least 25% and in worst case situations, you may be off by 40%, even if you
try to control all parameters and try to be as accurate as you can.
Exposure meter: 5%
Shutter         5% - 10%
Aperture        5%
Filmspeed       5% - 10%
Development     5% - 10%.

With a Leica R8 we are lucky as the shutterspeed is more accurately
controlled and the true aperture may be accounted for at the exposure
metering stage. We are also lucky as the exposure meter can use smaller
increments of shutter speed when we are in aperture priority.
But with an M, we have to work in increments of half a stop. This implies
that whatever figures the exposure meter gives, we have to round off to
these half stops. If the meter reads (as example) EV 10.3, we can use only
EV 10 or EV 10.5, as we can select only half stops (apertures 2.8 and 3.4:
again examples). With EV 10 we overexpose by a 1/3 stop and with EV10.5 we
underexpose by 1/5 stop. But given the margins, both exposures might be
right, or both may be off.
I once shot 36 frames at the same speed and aperture of a grey card and
after development, measured the density of each negative. Any variations
between the densities could only be explained by the error margin of the
shutterspeed. There was a measurable and sometimes visible deviation.
The upshot: deviations in a magnitude of 1/4 to 1/3 stop will occur, even if
we control our process as good as we can. It is hardly possible to use a
narrow margin of 1/6 of a stop, and we should settle for 1/4 stop as a
reasonable margin of accuracy.
But: if all deviations add up, we may differ by a full stop compared to
another person. So if someone remarks that he gets fine results with TriX,
pushed to 800, he may in fact be using his tolerance values and another
person with TriX at 400, might get identical results in identical
situations, depending on his margin of tolerance.
Upshot 2: never follow the specifications, given by others, as their set of
system components may deviate significantly from your own. Establish your
own set of specifications and stick with them. If you fail to get good
results with TRiX at 800 (your system), and another one gets good resuts
with TriX at 800 (his system), do not think you are doing something wrong
and do not try to emulate the other's recommendations. Or if you do, make
comparison shots with both sets of specs in ONE situation.
I do all comparison shots with flash light as this is the only way to make
sure that at least one big variable is under control.

Erwin