Bipod

Thanks, selmslie!

Bipod

Don't you want the "Sunny 16" rule to work?
And don't want the result of widening the aperture by one stop while
reducing the shutter speed by one stop to be the same as the original
exposure?

If f/16 & 1/125 doesn't give the same exposure as f/11 and 1/250,
and you don't know which shutter speed is off, it's going to be
very hard to use a light meter and take pictures.

selmslie Loc: Fernandina Beach, FL, USA

As you read the document you will understand that the method is sound.

1. Once a monitor is warmed up and you display a white page you can expect a constant image brightness. It will be the same for months until your monitor fades a little but if it is calibrated you can maintain the same brightness level until the monitor begins to fail. Since you are not dependent on sunlight the weather does not matter.

2. I don't need to calibrate the scanner. At any given time, what the Epson scanner measures is the relative density from where the negative is not even present to the density of the film edge (film base+fog, FB+fog). This is normally about 65 units on a scale from 0-255 for 35mm film and about 40-45 units for medium and large format film. The maximum density (dmax) I have measured is about 235. If the Epson scanner were not precise in these values it would be useless for scanning film.

3. The method I am using can be applied to all sorts of film formats but a step wedge can only be used with sheet film. Even with sheet film, a step wedge approach would be cumbersome and it would consume a lot of film and developer and time.

4. I was able to test more than 80 separate film and developer combinations over a period of about a month. In some cases, a single test revealed a useful film speed and development time, FB+fog level, dmax and curve shape while others needed two tests. I don't recall needing more than two tests on the same combination. The spreadsheet I created (I posted it earlier) provides a way to adjust the ISO and development time for the next test.

When I was printing in a wet darkroom, I did use the step wedge when I started to use a new paper or changed the contrast setting in my wet darkroom. However, I don't need to use a step wedge any longer since I don't print with an enlarger. I am committed to the hybrid process.

Bipod

Thanks for the explanation. I'm afraid I still haven't gotten around to going though your page with
the care it deserves.

It's great to hear about your film/developer combination tests. That must have greatly improved
your negative developing results.

I don't recall if you said you had seen Phil Davis's Beyond the Zone System book, but you might want to check it out.
He provides special rulers and tools for working with characterist curves, and even a mechanical "Power Dial" calculator:
just photocopy and cut out with scissors. (The calculator is most useful for sheet film, since each sheet can sit in the devleoper
for a different time, but if you want to delvelop and entire roll at, say, pull -1, it will tell you how many seconds that is.)


Yes, provided the supply voltage is constant or the regulator in the monitor's power supply is working (usually true).
I, too, use a monitor as a light box (I wrote a program that can control the luminence and reduce by a certain number of stops).
However, with any light box, it is not enough just to point the camera at it.. You also have to make sure that no lights or bright
surfaces outside the frame cause invisible flare in the lens. A lens hood is the easiest way.

Also, if you want to know the actual luminance of the light box, you'll need a calibrated light meter. Unfortunately,
different brands of meter are calibrated to different levels ("k factors"). They fall into several groups:
K=11.37: Gossen
K=12.50: Sekonic, Canon, Nikon
K=14.00: Pentax, Minolta
Thus on exactly the same scene, luminance meters in different groups will give different readings.
Could a 5.7% diffrence invalidate experimental results? Sure it could. There is no "exposure latitude"
when you're trying to calculate film speed or what time gives "normal development".
http://dpanswers.com/content/tech_kfactor.php


That's quite a leap. The fact that the Epson scanner scans film does not imply that it is calibrated. If it were
compressing the dynamic range half a stop, how would you know? If middle densities were being
underestimated by half a stop, how would you know?

You don't know if the scanner is reporting the relative density accurately, repeatably or with sufficient
sensitivity. To find out, you would have to test it.

Every instrument needs to be calibrated. Calibration involves testing several known samples
to give data points. These samples represent the actual curve of what the instrument is supposed
to measure. The data points reported by the instrument represent it's measured curve.

The measured curve can differ from the actual curve in several ways:
sensitivity, range, linearity, noise (random error), repeatability.

The simplest case is a line. Measured line can differ from actual line in:
slope. X-intercept, lack of linearity, endpoints (range), and repeatability.

Since you are measauring the same way each time, many of your results are valid.
But if all the curves are too steep, too shallow, etc. then your measured film speeds
could all be too fast or too slow.

If experimental science were easy, everybody would be a scientist. But any testing
is better than just taking every specification as gospel and hoping for the best.

The manufacuters could make this a lot easier by publishing test procedures for
their products, and selling the required test instruments and supplies. In fact,
they do that for their industrial equipment. But not for their consumer products.

Thanks again!

selmslie Loc: Fernandina Beach, FL, USA

I have read the book but have lost my hard copy. I think I have a PDF or Kindle version.

I have several light meters including a Gossen and Sekonic incident/reflected meter and Pentax and Minolta spot meters (the latter can be adjusted). I have reconciled the differences between them to the extent that only a couple of exceptions (like the one built into my Rolleiflex 2.8F) need to be compensated for accuracy.

But my faith in metering is not very high. I prefer to expose based on "1/3 stop darker than Sunny 16" for broad daylight and increase the exposure based on Exposure Value tables for darker scenes. This works well for both film and digital.

When I speak of "exposure latitude" I am referring to the tolerance of film to accommodate overexposure and the tolerance of digital sensors to record underexposure. I go into this in more detail in Film vs. Digital Characteristic Curves.
You are probably aware of the distinction between accuracy and precision when it comes to scientific measurements. A meter that is consistently off consistently by the same degree, for example on average +10% +/-1%, may be considered precise but not accurate. One that is off an average of +1% +/-10% would be accurate but not precise.

In this regard, the Epson scanner may be precise without necessarily being accurate. After all, the numbers it shows for film density (0-255) don't need to correlate to specific film log densities. They just need to be consistent from one reading to the next. I need precision and repeatability, not accuracy, so I don't really need it to be calibrated.

As they say, "If wishes were horses, beggars would ride."

Kodak, Ilford, Fuji and Agfa never agreed to a standard way to measure film speed. Even the figures from the Massive Dev Chart are suspect since there is no standard for agitation, etc. We don't really have an alternative to doing our own testing.

Bipod

All great points! But we aren't exactly "beggars" are we? We're paying for these products.

Film speed is a thorny problem because it is difficult to reduce a density-by-exposure curve to a
single number. (The same goes for digital sensor curves, by the way.)

ASA adopted a fractional gradient method in 1947 -- then halved the number twice just to be on the safe side.

Some methods currently in use:
ISO -- an improved version of the sold ASA method.
Kodak Contrast Index (CI) -- only used by Kodak
Ilford -- similar to ISO but with different speed points

However, all film manufactures publish at least one density (Diffie-Hellman) curve for each
film, the one for standard development time. From that, one can calculate any of the above three
speeds. It won't precisely match the standards, but it be pretty close.

Regarding your Espon scanner: it may indeed be "precise but not accurate", but you don't know
because different densities could be off by different amounts. A calibrated reflective step tablet might allow
you to test it (though I have never tried to test a scanner).

As for light meters, you probalby know how those K-factors were choesn: by having a panel look at hundreds of film
prints of "average scenes", and judge which ones were the best photos. I seem to recall this was first done back in the
1940s by Eastman Kodak.

I'm not sure that's still relevant for digital cameras--or ever was for hand-held (as oppsed to built-in) meters.
And it seems really subjective.

A hand-held light meter is like a thermometer: it should tell you the acurate temperature, not somebody's
idea of what's best for you to think is the temperature.. That's the argument Ansel Adams made to Kodak
and I agree with him.

All incident light meters meant for industrial use display the actual illuminance value. For example, I have one
meant for measuring lighting to see if it's bright enough to meet OSHA standards. I use to check my enlarger for
evenness of illumination of the easel (and when I align and adjust enlargers for friends).

Like most photo enthusiasts, I have a bunch of light meters loose and in cameras. But the only one I trust for
critical applications is a Pentax spotmeter.

It's amazingly hard to find a reference for calibrating light meters--except another (dubious) light meter.
I suppose I could send my Sekonic or Gossen back for calibration, but then who knows how they would
calibrate it. Seems like everything is a "trade secret", nowadays.

I wrote a program to figure out solar illumination based on date, latitutde, timezone, time of day, elevation
and atmopheric conditions (haze, smoke, etc). But the latter are too unpredictable and too hard to measure
to make the program good as a referance. It does give me a "sanity check" on meter readings.

Summing up: I used to rely on Fred Picker's "personal film speed" methodology. It worked most of the time
but I kept getting errors I couldn't explain. I was wasting a lot of time trying to track down the sources of
these errors. So I bit the bullet and started measuring shutter speeds, etc., and tagging the equipment.
Gradually, the errors have been going away.

I haven't done as much testing of different films as you have, since I basically only use one film for
outdoor work. But I'd like to do more, if I can find the time.

As complicated as all this is -- it is possible to measure and control it. But digital cameras and computer
printers are "black boxes". Seems like there would be one "open stanadards" manufacturer but there isn't.
It's too much like patent medicines for me.

Unfortunately, there is no independent testing of digital cameras or printers, let alone regulation. But with
film cameras, at least you could stick your favorite film into a new cameas and compare directly with your
old camera.

Consumers liked patent medicines. Lydia Pinkham got rich selling her "Vegetable Compound",
and Samuel J. Carter got rich selling "Carter's Little Liver Pills". Neither was ever proven to cure anything,
except as swollen wallet. The money-making bonanza didn't end until the FDA and FTC stepped in.

The Great American Consumer is not the sharpest knife in the drawer. An the Great American Busineesman
is not the most ethical person around. It's a bad combination.

William J. A. Bailey got rich selling "Radithor" Radium Water -- which surprisingly actually did
contain radium. According to Wikipedia:

Radium was the latest new hot technology. Very hot.

selmslie Loc: Fernandina Beach, FL, USA

One of the reasons for this is that some films have gradual toes and shoulders (a pronounced "S" shape) while others are straighter. There is really no fair way to pick a speed point. What I did was to try to get the curve shapes to align around middle gray minus 1, 2 and 3 steps. In general this produced a film speed that closely agrees with box speed but this approach fails for a couple of the films I tested - Rollei Retro and CMS 20.

At the other extreme my Epson scanner never appears to max out between dMin and dMax in any of my B&W films. My Coolscan 9000 can easily handle very dense film but I have to be careful to not develop to a high contrast.

In general, scanned film seems to work better with film developed to a slightly lower contrast. The good news is that this means slightly more latitude for overexposure. The bad news is that film speed may be a little lower.

I also find it to be my best meter (got mine from Fred Picker) but my digital version reports values to only the nearest 1/3 step. But that's close enough.

In the final analysis I feel that too much attention to precision diverts my attention from the actual goal.

bpulv Loc: Buena Park, CA

Thank you to all you UHH members who have responded to my thread. I started off with a simple question, what is the difference between a number 2 and 3 step tablet. I thought that three or four people would respond, I would have an answer and we would be done. Instead, I have been pleasantly surprised how my simple question has expanded into a stimulating and expanding discussion that is obviously of interest to many more of you then just myself. So keep it up!

User ID

No. Cuz there's only one shutter speed.
No. Cuz there's no aperture [no lens].
No. Cuz thee's no light meter involved.
No. Cuz you don't "take pictures" when
plotting an H&D curve. Have you read
the thread at all ?


`


`