Reciprocity misbehavior.

[Bob F.]

So, after all the graphs and charts are drawn is it not true that a metered exposure of 10 seconds can be be multiplied by 1.62 to become 16.2 seconds and a 100 second exposure becomes 162 seconds? I don't really need a graph or chart for this do I?

No, sorry, that is not the case. Look at the equation again: there is a value (different for each film) called tc,1 which is the adjustment for a 1 second exposure (found by experiment) for that particular film. That value needs to be plugged in to the equation to get the final time.

Cheers, Bob.

 

[gainer]

No, sorry, that is not the case. Look at the equation again: there is a value (different for each film) called tc,1 which is the adjustment for a 1 second exposure (found by experiment) for that particular film. That value needs to be plugged in to the equation to get the final time.

Cheers, Bob.

Furthermore, the ^ symbol means "Raised to the power" of the number that follows it, not multiplied by it. There is no simple multiplicative factor to convert indicated time to corrected time. I am preparing to post a more complete exposition which may make it easier to use and understand.

 

[gainer]

It is not linear -For TRI-X, based on Kodak pub F4017, if the EV indicates a 10 second exposure, the correction is +2 stops - or 40 seconds - with a 20% reduction in development. At an EV indicating 100 seconds the corrections is 3 stops or 13.3 Minutes with a reduction in development of 30%. Even a one second exposure is supposed to be at +1 stop. (I generally don't start correcting till there is an indication for 2 sec or more.)

Kodak has been quoting those numbers for a long time. A problem is that view camera users do not usually consider aperture adjustment a viable option. The aperture is set for depth of field as a rule. Most of the majestic scenery has little need for f 64, but still there is the adage "f64 is where it's at."

Howard Bond's data show much less than 1 f-stop equivalent time extension for 400TX. Nevertheless, the ancient Kodak data I had from before 1950 does fit the equation that I presented quite well. My equation would calculate 41 seconds added exposure at 10 seconds indicated, or 51 seconds altogether.

 

[gainer]

There is an advantage to plotting the correction vs the indicated exposure. You need plot only one line, either on linear or log-log paper, which we will call "Basic correction". On log-log paper it is a straight line with a slope of 1.62 inches (or cm) rise per inch (or cm) run, as they say in surveying. It passes through 1 second correction at 1 second indicated. On linear graph paper, you calculate enough points to define the line for an imaginary film with a reciprocity correction of 1 second at 1 second indicated exposure. This calculation is easy on a TI-30 pocket calculator. Basic correction = indicated time ^ 1.618.

For each film there is a film factor that is multiplied by the basic correction to get the actual correction. This value is added to the indicated time to get the corrected exposure time. Film factors must be defined by experiment.

I have provided a table of values that you can plot on either form of graph paper. If you know any reciprocity correction for any indicated time for any film, you can calculate the film factor by simply dividing the actual correction by the basic correction that you read from the graph for that indicated time. For example, if I know that the exposure time for my film is 20 seconds at 16 seconds indicated, the actual correction is 4 seconds, the basic correction is 88.8 seconds, and the film factor is thus 4/88.8 or 0.045. This film factor then allows you to estimate the reciprocity correction for that film for any desired indicated time. In the example case, an indicated exposure of 32 seconds would have a basic correction of 273 seconds which multiplied by the film factor od 0.045 gives a correction of 12.3 seconds for a total exposure of 44.3 seconds.

Let me restate what we have so far. Given a plot of basic reciprocity corrections in the form of an increment to be added to the time indicated by your light meter, and a film factor for a particular film, you can read the basic correction for the indicated time, multiply it by the film factor and add the result to the indicated time to get the time you should use. For any film, if you know the reciprocity correction for a given indicated time, you can get its film factor by dividing the actual correction by the basic correction from the graph.

All of this is empirical of course. It fits the films for which I have data with practical accuracy. The fact that the same basic curve works with both traditional and tabular grain films of two different manufacturers lends some degree of confidence to the method.

For most users of view cameras, changing aperture to correct for reciprocity is not a desirable option. The exposure time adjustments determined by Howard Bond and represented by these charts and graphs will assure the retention of shadow detail.

CHART FOR ESTIMATING CORRECTIONS FOR RECIPROCITY FAILURE

Indicated Basic For this film: Multiply by
time, seconds correction
1 1 400TX...........................0.17
1.4 1.7 400TMAX, 100TMAX.........0.07
2 3.1 HP5+.............................0.11
2.8 5.4 100Delta........................0.05
4 9.4
5.6 16.5
8 28.9
11 50.6
16 88.8
22 155
32 273
45 477
64 836
90 1465
128 2567
180 4497
256 7880

Multiply the basic correction by the factor for your film and add it to the indicated time.For example, if your light meter says you need 16 seconds for 400TX, the corrected time would be 88.8 X .17 + 16 = 31 seconds.

 

[gainer]

That chart was in proper format when I submitted it. Now I can't even read it. I hope you can see what I was getting at from the words.

 

[gainer]

That chart was in proper format when I submitted it. Now I can't even read it. I hope you can see what I was getting at from the words.

I'm trying again.

CHART FOR ESTIMATING CORRECTIONS FOR RECIPROCITY FAILURE

Indicated.........Basic...........For this film:..................Multiply by
time, seconds...correction
1..................1................400TX............................0.17
1.4................1.7.............400TMAX or 100TMAX.......0.07
2..................3.1..............HP5+.............................0.11
2.8................5.4.............100Delta.........................0.05
4..................9.4
5.6...............16.5
8..................28.9
11................50.6
16................88.8
22................155
32................273
45................477
64................836
90................1465
128...............2567
180...............4497
256...............7880

Multiply the basic correction by the factor for your film and add it to the indicated time.For example, if your light meter says you need 16 seconds for 400TX, the corrected time would be 88.8 X .17 + 16 = 31 seconds

 

[Bob Carnie]

so according to my simple calculations
tri x indicated 90 second exposure would actually be a 265.05 second exposure?

 

[gainer]

so according to my simple calculations
tri x indicated 90 second exposure would actually be a 265.05 second exposure?

I think my table is still not clear. The basic correction for 90 seconds is1452. The film factor is 0.17. The product of those is 247. Add 90 and you get 337.

Whoever said that FP4+ data would be enough, I found that Ilford thinks one size fits all. My equation fits their curve, but Howard's data do not. The difference is that I get a film factor of 0.51 when I use Ilford's curve. The same basic correction that I presented above fits both ends of Ilford's curve when I use that film factor. I would try the same correction for FP4+ as for HP5+, since Ilford seems to think they are all the same, including Delta400.

 

[gainer]

P.S.
I believe Howard's data. I also believe that HP5+ and FP4+ corrections are enough alike that it won't make a significant difference in practical work.

 

[Bruce Osgood]

so according to my simple calculations
tri x indicated 90 second exposure would actually be a 265.05 second exposure?

I get a different number for 400TX:

1465 (correction factor) X .17 + 90 = 339.05. Or 5.65 minutes

 

[Lee L]

You can tell that Mr. Gainer used to design flexible, field-usable aids to navigation and other such things professionally. The attachment is a .pdf file that will make a little card to carry for reciprocity adjustments using Gainer's numbers. (I did a .pdf to make sure it holds formatting and can be printed and resized properly.) The first two columns are metered exposure times expressed as minutes in the first column or as seconds in the second column. These "just happen" to match the time markings on my light meter. The third column shows basic correction times in seconds, calculated by raising the metered time to the power of 1.62 per Gainer. The right hand box indicates Gainer's factor by which to multiply the "Basic Correction", dependent on which film you're using.

You get your final exposure time by metering, reading the corresponding basic correction, multiplying the basic correction by the film's Bond-Gainer reciprocity factor, then adding the result to the metered time. It's very simple if you have a calculator at hand and can do multiplication. An example is on the attached .pdf "card" to carry along as an aid in using the card. This is a preliminary stab at this, and I may work on making it more user friendly as time allows.

Someone with the time and inclination could make a slide rule version of this information for field use, with a chart of film factors or a place to list your own factors on the back. A circular version on card stock could be very packable and handy. Yes, I'm old enough to have learned to use a slide rule, and I still have a couple of them.

As for Ilford FP4+ and it's relation to FP5+, Reeves's data shows that FP4+ has marginally more reciprocity failure at 128 seconds of exposure, a loss of 1.2 stops as opposed to 1.18 for FP5+. This translates to a Schwarzschild exponent of .79 for FP4+ and .83 for FP5+. That points to Mr. Gainer's comments being right on target regarding exposing FP4+ at the adjustments for FP5+.

I also believe that Mr. Bond's data is accurate and Mr. Gainer's analysis is correct. I need to order the back issues so I can have a first-hand look and learn more. I have no way to prove this, but I think the manufacturers assume that anyone using their film for critical work will test under conditions of use, and that they don't think that testing under manufacturers' lab conditions will have a lot of bearing in the always unique circumstances in the field. Therefore, they only provide general guidelines that will get the users close enough on the first try to have something usable, and from which to make their own adjustments.

Lee

 

[Bruce Osgood]

Gadget writes:
"Log(tc) = log(tc,1) + 1.62 log(tm)

where tc,1 is the correction at 1 second indicated time."

So, to find out what a films tc,1 is, you must conduct a test, OK, what is this test?

 

[David]

The data provides the following numbers (assuming the reasonableness of the given factors).

 

[gainer]

When I went to calculate the adjusted time for 400TX at 90 seconds after I opened the message box, my chart scrolled off the screen. I just recalculated it, but I used 1.618 as the magic number whereas I had used 1.62 when I calculated the numbers for the chart. I doubt the 2 seconds or so will ever be missed out of 339. I hope I never have to do an indicated 90 second exposure. I get impatient with a 5 minute developing time.

 

[gainer]

Gadget writes:
"Log(tc) = log(tc,1) + 1.62 log(tm)

where tc,1 is the correction at 1 second indicated time."

So, to find out what a films tc,1 is, you must conduct a test, OK, what is this test?

If you know the correction for ANY indicated time, you can calculate the film factor by reading the basic correction off the curve or from the chart and dividing that number into the amount you had to add to correct for reciprocal trade disagreement. Most film makers give you something, and the values I have seen will more likely give you some over exposure, so you should be able to zero in on a reasonable film factor after 1 or 2 trials.

 

[kaiyen]

yay! I can now populate a roughly accurate chart that has multiple films on it! then I can (hopefully) add in proper film factors after I (hopefully) do the appropriate testing.

I'll add more and more films to it over time, until I have the entire world included! mwaaahahahah...

g'night.
allan

 

[Bob Carnie]

Mr Gainer

I see where I made the mistake and I now do understand the chart
Thank you very much for your patience and time . I am starting to do a series of night exposures that I believe this chart will be invaluable for my project.

 

[Lee L]

A Gainer reciprocity method primer

I've been at home with sick kids for the last few days while my wife is out of town. Working on this between other things has been much more interesting than cleaning house, and gave me something to ponder when I was cooking and cleaning. I've been aware of the Schwarzschild method for calculating reciprocity law failure for years, and was interested in how Gainer's method may relate to it, especially since Gainer's is based on readily available data that was carefully acquired. Every time I read a manufacturer's reciprocity data I get the feeling that I'm looking at a rough guess with Schwarzschild applied.

Robert Reeves, as I have mentioned before, did a lot of film tests prior to 2000, and his method and results can be seen at:
http://www.robertreeves.com/filmtest.htm
with specific B&W results at:
http://www.robertreeves.com/b&w.htm
Look at it when you get a chance. It's worthwhile.

So I've been trying to find a way to convert Reeves' Schwarzschild exponents to Gainer film factors. I can find no direct mathematical conversion that's convincing. However, I did take the Schwarzschild equations and exponents from Reeves and the Gainer formula and factors posted here (I don't have the article at hand), and I took the four B&W films they share in common as a basis for finding out how closely the methods match. I found that I could apply the same corrections for all four films common to both tests and get calculated exposures from the Gainer formulas and from the Reeves test results (with conversions applied to yield Gainer film factors), that all agreed within about 1/6th of a stop. This agreement is across four films at exposure times of 1 sec, 128 seconds, and 900 seconds of metered exposure, and also across probable emulsion changes. I figured that was close enough to make the conversions useful as a starting point when applied to films Reeves tested that are not in the Bond data.

So I've attached the results in a three page .pdf file, with a Gainer basic correction chart (log axes with instructions for reading log axes), a film factor table with 13 B&W films (5 from Bond data and 8 from Reeves data), instructions, and a worked example. In case you want to know if this is immediately applicable to films you use, the films added from the Reeves data that aren't in Bond's data are: Pan F (not Plus), FP4+, Plus-X, Tech Pan (bye-bye), Ilford SFX, XP-2, T400CN, and TMZ.

Lee

 

[cao]

What is the correction preserving?

Forgive me if I seem dense, but what is the correction preserving? Without even considering the developer, we could think in terms of density of development sites for a given light in a given time. Are we preserving zone V? How is this site density curve being changed? What I am shooting at is this; I have the niggling feeling that the curve is changed in a way that isn't fully correctable by changing development. If this is the case, then we can stretch tones around by deliberatly exposing so as to create RF, and hence RF can be used as a creative control with the right films. I'm seeking a correction function mapping an site density curve to a different curve in the face of RF. Do we already know this from Gainer's correction?

 

[Lee L]

Forgive me if I seem dense, but what is the correction preserving? Without even considering the developer, we could think in terms of density of development sites for a given light in a given time. Are we preserving zone V? How is this site density curve being changed? What I am shooting at is this; I have the niggling feeling that the curve is changed in a way that isn't fully correctable by changing development. If this is the case, then we can stretch tones around by deliberatly exposing so as to create RF, and hence RF can be used as a creative control with the right films. I'm seeking a correction function mapping an site density curve to a different curve in the face of RF. Do we already know this from Gainer's correction?

This is always part of the problem with reciprocity failure. The Kodak recommendations for development adjustment earlier in this thread are a good place to start, but I don't have the backissue of the Bond article in Photo Techniques, so I can't say what he did. I can tell you that everything I've seen done in astrophotography appears to be aimed at preserving density on a gray card, or Zone V. I expect that Bond has tried to preserve the entire tonal scale to whatever degree is possible, but can't be sure until I see the article. Look at the Reeves websites in my last post to see the standard astrophotographers technique for determining the Schwarzschild exponent. As always, you'll probably have to test materials for your own workflow. Even then, Covington hints that the variation in film batches is enough to swamp your standard adjustment, especially at the typical 8 minute to hours long exposures in film astrophotography. Astrophotographers aren't typically concerned about preserving a specific dynamic range, they mostly want to go as faint as possible without getting the sky background washed out too much, and there are varying atmospheric conditions that determine that limit, often unpredictably. Astrophotographers bracket exposure times as much as possible.

I know this doesn't help with the moonlit landscape. I need to get the Bond and Gainer articles myself. You're not being dense, it's just that the testing for this is very tedious, and I've never seen anyone do an analysis as complete as the Bond article seems to be. I do recall having read some techniques used by cathedral photographers, including stand development, but that was years ago, and I don't have a ready reference. Read the Reeves web pages to get an idea of how to procede with your own testing using a Wratten #96 or a B+W 110 filter. (If you have and IR sensitive film, the Wratten #96 is not typically a 3.0 density in those wavelengths, so experienced people say to go with the B+W 110.)

There are a couple of books out there on "night photography" that might have useful information, but I haven't seen them.

Lee

 

[gainer]

Forgive me if I seem dense, but what is the correction preserving? Without even considering the developer, we could think in terms of density of development sites for a given light in a given time. Are we preserving zone V? How is this site density curve being changed? What I am shooting at is this; I have the niggling feeling that the curve is changed in a way that isn't fully correctable by changing development. If this is the case, then we can stretch tones around by deliberatly exposing so as to create RF, and hence RF can be used as a creative control with the right films. I'm seeking a correction function mapping an site density curve to a different curve in the face of RF. Do we already know this from Gainer's correction?

The failure of film to obey our elegant law of reciprocity has been known for a long time. The general effect is to change film speed with length of exposure. It is often thought of as having a threshold below which there is complete reciprocity. That is, we can exchange light for time. Less light needs more time in a reciprocal inverse manner. Actually, the failure is there from 0 time on up, but is negligible below a certain exposure time.

It is as if you put a slower film in your camera for longer exposures. The object of the corrections is first to preserve the shadow detail. Call it Zone I if you will. That is a matter of adjusting exposure time to make up for the loss of film speed. Once that is done, normal development may or may not preserve the desired contrast index. Ilford does not mention any change of development in their recommendations for reciprocity correction, but the change of exposure time they recommend for all their films is almost exactly what my table calculates for a film factor of 0.51 . It is fairly certain that by increasing the exposure time by the amount calculated from my table for those films Howard tested you will obtain a printable negative without change in development.

 

[cao]

I am aware of the meaning of RF; I just hoped there might be a good general theoretic model for how film deviates from a simple integrator of light energy. While I have a math background, I have none in physical chemistry, so I can't do anything but ask questions.

 

[gainer]

I am aware of the meaning of RF; I just hoped there might be a good general theoretic model for how film deviates from a simple integrator of light energy. While I have a math background, I have none in physical chemistry, so I can't do anything but ask questions.

Me too. I thought the standard meaning of RF was "radio frequency" and was wondering how that came into play here.

I am an engineer by education and nature. Our traditional job is to put theory into practice. Sometimes we don't even have a theory to work on, but still have to find a way to practice something. Other times, theory does not predict nature very accurately and we use empirical equations to fit data. With some degree of experience, we learn to recognize certain general mathematical forms that seem to fit the available data. Every human being does the same thing without knowing it, as when a baseball player learns to chase and catch flyballs. I cannot claim to have or even to know a theory of why my empirical equation fits data, although it has enough arbitrary coefficients to fit many situations.
The most remarkable thing to me about what I have learned is that the basic RF (have it your way) curve fits all films I have data for with only 1 arbitrary constant for each film. It appears that the only things you need to know are the basic curve or equation which depends only on the indicated exposure but is independent of the film and a film factor which is independent of the indicated exposure time.

 

[Bruce Osgood]

Oh, oh....... You mean it's not Range Finder?

 

[cao]

Me too. I thought the standard meaning of RF was "radio frequency" and was wondering how that came into play here.

I've seen reciprocity failure written as RF and RLF. I meant to save typing rather than mislead. I'll be sure to define stuff more carefully now.