TMY reciprocity compensation

LKT · Jul 5, 2006

I am experimenting with long exposures and have found some very detailed charts on reciprocity compensation for TX400 and TMX, but there is no data on TMY, which is my default film.

<http://seeinglight.com/reciprocity.shtml>

Any suggestions on finding corresponding data on TMY? Have tried Google and Digitaltruth.

TIA.

Ka Tai

Steve Hamley · Jul 5, 2006

Why not check the site of the people who actually make it?

http://www.kodak.com/global/en/professional/support/techPubs/f32/f32b.jhtml#1153016

Steve

Lee L · Jul 5, 2006

Yesterday I looked up the information I have on TMY in two astrophotography books for a friend here on APUG. They give Schwarzschild factors of 0.80 and 0.81 for TMY, both derived from the same test method. I'll not go into the method here unless you request it, but the method relies on exposure tests.

To use this factor, you plug it into the formula T=(t+1)^(1/p)-1 where
T=corrected exposure time in seconds
t=metered exposure time in seconds
p=Schwarzschild factor of the film
(Be sure to follow the "order of operations" rule and raise t+1 to the power of 1/p before subtracting the 1.)

This should get you started.

Lee

P.S. Decided to attach the numbers in minutes for p=0.81. The difference between that and p=0.80 in the following numbers works out to a 5-10% longer exposure across the range included, which is close enough. The tabbed columns will probably run together in HTML, so you'll have to sort that out.

Metered/Corrected
1 2.65
2 6.19
3 10.2
4 14.52
5 19.12
6 23.93
7 28.94
8 34.11
9 39.44
10 44.92
11 50.52
12 56.24
13 62.07
14 68.01
15 74.06
16 80.19
17 86.42
18 92.73
19 99.13
20 105.61

c6h6o3 · Jul 5, 2006

Lee L said:
Yesterday I looked up the information I have on TMY in two astrophotography books for a friend here on APUG. They give Schwarzschild factors of 0.80 and 0.81 for TMY, both derived from the same test method. I'll not go into the method here unless you request it, but the method relies on exposure tests.

To use this factor, you plug it into the formula T=(t+1)^(1/p)-1...etc.

Way too complicated. Just use Kodak's charts. If 1 second is indicated, add 1/3 stop. If 15 seconds is indicated, add 1/2 stop. If 200 seconds indicated, add 1-1/2 stops. For anything in between, interpolate. Easy. Quick.

I generally give TMY a little more light than this (i.e., 1 second:1/3 stop, 10 seconds: 2/3 stop, 100 seconds: 1-1/2 stops), but I contact print onto Azo so I'm comfortable with a little denser negative than I would want for enlargement. Using either table, however, you'll be close. You really don't need to plug into any formula that would require a calculator.

gainer · Jul 5, 2006

There is a problem with using aperture adjustment to correct for reciprocal trade disagreement for view camera users. Most of the time, the aperture controls depth of field and time controls the exposure.

The best fit I have found to experimental data for long exposures is:

T = Tm + a * (Tm ^ 1.42)

where Tm is the exposure time read from the light meter's calculator and a is constant for any given film. For TMY or TMX you may use 0.06.

This is a very simple calculation,for most pocket calculators that photographers are likely to use have the ^ function. The part of the equation to the right of the = plots as a straight line on log-log graph paper.

The experimental data of Howard Bond do not agree well with Kodak, who seem to use the same correction for all films.

Lee L · Jul 5, 2006

gainer said:
The best fit I have found to experimental data for long exposures is:

T = Tm + a * (Tm ^ 1.42)

where Tm is the exposure time read from the light meter's calculator and a is constant for any given film. For TMY or TMX you may use 0.06.

Pat,

I'm just headed out the door, but I wanted to note that you might have a typo here. Isn't the last number 1.62 rather than 1.42? I recall it was the golden mean, Fibonacci ratio, or whatever you want to call 1.618....

Lee

jp80874 · Jul 5, 2006

c6h6o3 said:
Way too complicated. Just use Kodak's charts. .

The problem with just using Kodak's charts is that the Kodak chart ends at a metered exposure of 100 seconds. Lee's list starts at one minute or 60 seconds and goes up in minutes from there.

John Powers
Lee's friend on APUG who needs all the math help he can get.

gainer · Jul 5, 2006

Lee L said:
Pat,

I'm just headed out the door, but I wanted to note that you might have a typo here. Isn't the last number 1.62 rather than 1.42? I recall it was the golden mean, Fibonacci ratio, or whatever you want to call 1.618....

Lee

That wasn't the only one of my boo-boos. The part that plots as a straight line on log-log paper is the right hand side of:

Tc - Tm = a * (Tm ^ 1.62)
where Tc is the corrected exposure.

Encephalitis is a terrible thing.

This equation gives 204.3 seconds as the corrected exposure time for a measured time of 100 seconds, and 918 seconds at measured 300 seconds.

Sorry about that. Now all you need is a meter that will make those readings. It shouldn't be a problem for the f-64 club.

gainer · Jul 5, 2006

Now I remember. 1.42 was my factor for the ratio of stand development to agitated when the same dilution was used for both.

c6h6o3 · Jul 5, 2006

jp80874 said:
The problem with just using Kodak's charts is that the Kodak chart ends at a metered exposure of 100 seconds.

True, but is this really such a problem? For each order of magnitude (10^0, 10^1, 10^2) we add approximately 1/2 stop exposure. (I know, this is an approximation but it's close.) I should think that at 1000 seconds (10^3) the correct compensation would be 2 stops. In other words, you're never going to be off by more than 1/2 stop no matter how bad your interpolation.

gainer · Jul 5, 2006

You still have to be able to correct exposure without changing f-stop. When you change the f-stop, you change the effective measured time. The correction of 2 stops would be unacceptable to the view camera user who sets the f-stop without concern for its effect on exposure, then sets the exposure time to suit the f-stop.

I calculate a corrected exposure time of 5346 seconds when the metered time is 1000 seconds. That is a change from 16.7 minutes to 89.1 minutes. If you open 2 stops, the metered time would drop to 250 seconds. I would calculate the correct exposure time to be 710 seconds, or 11.8 munutes, not 1000 seconds or 16.7 minutes. It would be 1/2 stop overexposed at 1000 seconds. This may or may not be acceptable exposure error, but opening from f/64 to f/32 or from f/32 to f/16 might change the picture quite a bit.

LKT · Jul 6, 2006

gainer said:
You still have to be able to correct exposure without changing f-stop. When you change the f-stop, you change the effective measured time. The correction of 2 stops would be unacceptable to the view camera user who sets the f-stop without concern for its effect on exposure, then sets the exposure time to suit the f-stop.

I calculate a corrected exposure time of 5346 seconds when the metered time is 1000 seconds. That is a change from 16.7 minutes to 89.1 minutes. If you open 2 stops, the metered time would drop to 250 seconds. I would calculate the correct exposure time to be 710 seconds, or 11.8 munutes, not 1000 seconds or 16.7 minutes. It would be 1/2 stop overexposed at 1000 seconds. This may or may not be acceptable exposure error, but opening from f/64 to f/32 or from f/32 to f/16 might change the picture quite a bit.

Well, I certainly get a lot more info than I bargain for!

You are certainly right about not wanting to change aperture when using a view camera. It affects the image to a greater extent than with small format. In fact, with B/W, for which exposure is not so critical, I sometimes take several exposure with different apertures for the same image, changing the shutter speed to keep the exposure the same. In effect, I bracket the depth of field.

Thanks for the formula. I made a few calculations and compared them with the table I got in SeeingLight Gallery. For exposures up to one minute, the recommended time as calculated is shorter than recommended by the table, after that the calculated time is longer. I assume that the data on the table are obtained empirically.

Another interesting bit is that you recommend to use the same constant for TMX and TMY. So both film have the same reciprocity characteristics? That will solve a lot of my problems.

Thanks everybody for your input. What a great forum this is.

Ka Tai

Lee L · Jul 6, 2006

Ka Tai,

I noticed that some of the tables you are using are from Bruce Barnbaum's empirical tests or from Photo Techniques, a magazine that he often writes for. IIRC, Gainer's formula is designed to fit data from Howard Bond, derived empirically and published in Photo Techniques.

The numbers I gave are from empirical tests as well, but show p=0.88 for TMX. These are derived from tests shots of a gray scale with no filter at 1/8 second, then with a 1000x ND filter and exposures of 128 seconds at bracketed f-stops. The Schwarzschild factor is then calculated from the adjustment needed in f-stops to render the gray scale as close to the 1/8 second exposure as possible.

If you don't find that doing exponential equations in your head clears your head and puts you in a receptive, zen-like state, you can always plug the formula I gave (which is a Michael Covington derivation from the standard Schwarzschild formula) into a spreadsheet and work up your own pocket chart that converts metered times to reciprocity corrected times.

There is another long thread on this topic with more information and a number of charts here on APUG. (there was a url link here which no longer exists)

Lee

LKT · Jul 7, 2006

Lee L said:
Ka Tai,

I noticed that some of the tables you are using are from Bruce Barnbaum's empirical tests or from Photo Techniques, a magazine that he often writes for. IIRC, Gainer's formula is designed to fit data from Howard Bond, derived empirically and published in Photo Techniques.

The numbers I gave are from empirical tests as well, but show p=0.88 for TMX. These are derived from tests shots of a gray scale with no filter at 1/8 second, then with a 1000x ND filter and exposures of 128 seconds at bracketed f-stops. The Schwarzschild factor is then calculated from the adjustment needed in f-stops to render the gray scale as close to the 1/8 second exposure as possible.

If you don't find that doing exponential equations in your head clears your head and puts you in a receptive, zen-like state, you can always plug the formula I gave (which is a Michael Covington derivation from the standard Schwarzschild formula) into a spreadsheet and work up your own pocket chart that converts metered times to reciprocity corrected times.

There is another long thread on this topic with more information and a number of charts here on APUG. (there was a url link here which no longer exists)

Lee

Thank you for the link to the previous thread on reciprocity failure. You guys certainly havce gone through it in depth. I shall bookmark it and refer to it when I am in doubt.

For now, I have made a table of corrected exposure times. Seeing that a metered exposure of 30 minutes will end up as over 4 hours when corrected, I have decided not to go further than that. There is a limit to what one is willing to sacrifice for the sake of art.

TMY and TMX both uses the same correction factor. Thank god for small mercies.

Ka Tai

TMY reciprocity compensation

Privacy & Transparency

Privacy & Transparency