Comparative reciprocity failure of available sheet films

[Tom Kershaw]

While I have read that Kodak TMY-2 has good reciprocity characteristics and the Fomapan films show very significant reciprocity failure; the ILFORD data sheets for HP5+, FP4+, and Delta 100 all seem to show the same reciprocity characteristics through a rather imprecise and limited graph.

A photo.net contributor has published a formula to calculate reciprocity failure of Delta 100 'Corrected time = 1.15567 x (Metered time)^1.4379'; however, I have seen little in way of comparison between HP5+, FP4+, and Delta 100. Does a lack of comparison indicate these films all have the same reciprocity characteristics, which seems unlikely...?

Tom.

 

[Thomas Bertilsson]

Howard Bond published an article on this with some very useful information. I believe it was Photo Techniques in 2003. I don't have a copy on hand, unfortunately.
It included Tri-X 400, Tmax 400, Tmax 100, 100 Delta, and HP5+. That was the old version of Tmax 400, of course.
Tmax 100 and 100 Delta were the most linear of the bunch. But TMY is two stops faster... HP5 and Tri-X were the worst performers. Another conclusion was that no compensation in development needed to be done, which is good news.

100 Delta and HP5 definitely did not have the same reciprocity characteristics, according to Bond's test.

You'll have to locate the article somehow.

- Thomas

 

[Ulrich Drolshagen]

I found it here. http://www.phototechmag.com/articles/articles/200705/0403Bond_Reciprocity2.pdf

Ulrich

 

[Ian Grant]

Thanks Ulrich, that's the most useful article I've seen in a long while. I'll make a precis of the data to keep with my lenses.

Ian

 

[An Le-qun]

Tom,
In case you are interested, Fuji's boast that Acros 100 shows very little reciprocity failure seems to have something to it--which I found out quite by accident. I don't recall whether the sheet included with the box of 4x5 provides extensive information for it, though, although that may be available elsewhere.
Le-qun

 

[Ulrich Drolshagen]

I recently suggested in a German forum already to collect data on reciprocity-failure. It seems to me more valuable than information on developing times in respect to certain developers, as it's inherent to the respective film, largely independent from individual influences and moreover difficult to test without the proper equipment.

Ulrich

 

[gainer]

Search www.unblinkingeye.com for "LIRF is Lurking at Your F-stop" by Patrick Gainer. It shows some interesting facts about reciprocity obtained by analyzing Howard Bond's data in an unusual manner.

 

[Todd Niccole]

Tom,
In case you are interested, Fuji's boast that Acros 100 shows very little reciprocity failure seems to have something to it--which I found out quite by accident. I don't recall whether the sheet included with the box of 4x5 provides extensive information for it, though, although that may be available elsewhere.
Le-qun

I've tested this film for astrophotography and it does have very good reciprocity characteristics. I could see that density of nebula was still building against background in test exposures up to 32 minutes. For terrestrial shooting this just goes to show that it does indeed have excellent performance, likely the best on the market with LIRF in mind. Fuji boasts no reciproity adjustment for exposures up to two minutes. I plan further tests but I'm a little concern about how easily it seems to scratch; I'm going to have to track that down.

 

[Ulrich Drolshagen]

Search www.unblinkingeye.com for "LIRF is Lurking at Your F-stop" by Patrick Gainer. It shows some interesting facts about reciprocity obtained by analyzing Howard Bond's data in an unusual manner.

I just did a quick check with some data for HP5 from the Bond article. It really seems to work.

Ulrich

 

[ghost]

Search www.unblinkingeye.com for "LIRF is Lurking at Your F-stop" by Patrick Gainer. It shows some interesting facts about reciprocity obtained by analyzing Howard Bond's data in an unusual manner.

Thanks for the article-

using your graph for all films, the corrected exposure for an indicated time of 1 sec is 1.3 sec, for a time of 10 sec is 14 sec? am I doing this correctly? seems kind of low corrections for Tri-X in my experience- or I have been over-correcting...

 

[gainer]

The formula should have lead you to these numbers:
The time to be added to 10 seconds indicated exposure for a film that has a correction of 0.3 seconds at 1 second indicated is:
tadd = 0.3*(10^1.62) = 12.51. .

Add 12.51 seconds to the indicated 10 seconds to get 22.5 seconds after correction for non-linear reciprocity. Your suspicion was correct. I don't know what you did to get 4 seconds instead of 12.5, but you surely didn't calculate 10^1.62 in the process!

Remember: the coefficient a in the equation is the amount of the reciprocity CORRECTION at the measured exposure time of 1 second, 0.3 seconds in the case at hand.

 

[gainer]

I forgot to emphasize that it appears from my analysis of Howard Bond's data that the value 1.62 is common to all films he tested within the accuracy of measurement. A change from 0.3 to 1.0 in the coefficient "a" is the difference between 621 and 1838 seconds at a measured exposure of 100 seconds.

 

[Lee L]

Thanks to Ulrich for finding the Bond article online. I found it once, but the link to it disappeared and I thought it was gone.

I've been working on Bond's numbers off and on for a while. The best fit depends on the film, and is either a power or log fit. The following are best fits derived with CurveExpert, a MS Windows program that I run under linux. The formulae are presented as they would be typed into a spreadsheet to calculate a corrected time that takes reciprocity failure into account, and all times are in seconds in these formulae.

In practical terms, these would be within a small fraction of a stop of Gainer's method.

Tri-X
corrected time = EXP(1.2147591*LN(metered time)+0.19783161)

HP5+
corrected time = EXP(1.2746481*LN(metered time)-0.18828707)

100 Delta
corrected time = EXP(1.0020463*LN(metered time)^1.0793326)

TMY
corrected time = EXP(1.1577419*LN(metered time)-0.076131411)

TMX
corrected time = EXP(1.0179975*LN(metered time)^1.0959838)

I'll also attach a .pdf chart I just made up before logging onto APUG using these formulae, as I just happened to be working on this again, having some time exposure ideas I want to try. I've also added columns for Plus-X as per the Pinhole Designer software, which is also a very close match for the old Kodak and Ilford generic corrections, which are also in a column on my chart. The format for corrected times in the chart is mmm:ss Please let me know if you see any mistakes, as I just finished this up. I should also mention that the Efke 25 times are derived from an internet post by Andrew O'Neill at photo.net, with his data interpreted as per Robert Reeves and Michael Covington in their respective books on astrophotography.

Lee

 

[Lee L]

While we're on it here are some formulae for Acros. If you plug in the linear and power fits, you'll find no significant difference in the results or the fit to the data, which was taken from an APUG post by André E.C. : (there was a url link here which no longer exists)

Don't bother with corrections at times of 120 seconds or less.

Acros regressions from CurveExpert

Power Fit: y=ax^b
Coefficient Data:
a = 0.72538705
b = 1.06559210
correlation coeff: 0.99999128

Linear Fit: y=a+bx
Coefficient Data:
a = -36.07858400
b = 1.20142490
correlation coeff: 0.9997068

Lee

 

[gainer]

Lee, please calculate the corrected exposure for Acros at a metered exposure of 1 second. I'm missing something here. It's not old age. I have plenty of that. If y is the answer, then the corrected time can never be less than 0. 7254 seconds by the power fit, or greater than 0 until metered time > 30.03 seconds by the linear fit. Something doesn't fit!

 

[Lee L]

Lee, please calculate the corrected exposure for Acros at a metered exposure of 1 second. I'm missing something here. It's not old age. I have plenty of that. If y is the answer, then the corrected time can never be less than 0. 7254 seconds by the power fit, or greater than 0 until metered time > 30.03 seconds by the linear fit. Something doesn't fit!

Pat,

I have. That's exactly why I recommended not using any corrections below 120 seconds, where they aren't necessary with Acros. The fit is only good beyond that time.

I also logged back in to address the original question of Delta 100 reciprocity. Whatever method was used by the photo.net poster certainly doesn't agree with the behavior found by Howard Bond. It's more in line with the generic curves that manufacturers posted, which Bond found to be inaccurate or out of date.

Lee

 

[gainer]

WHOOPS! Maybe it is old age. The computed time at one second metered time is 0.7254 by the power fit and -34.8773 by the linear fit. it still doesn't fit.

 

[Lee L]

Tom,

I found the expression you posted through google, and the post indicates that it was based on Ilford's leaflet information for reciprocity corrections, which Bond found inaccurate in his tests. If you read the Bond article that Ulrich posted earlier, you'll find how his results differ, and why he thinks that's the case. Kodak's and Ilford's generic recommendations appear to be for older films, and have remained basically unchanged for decades.

The expression you posted is also extremely close to that posted by Ilford reps on photo.net, where corrected exposure = metered exposure ^ 1.48 , which is again their generic curve of very long standing.

Lee

 

[Lee L]

WHOOPS! Maybe it is old age. The computed time at one second metered time is 0.7254 by the power fit and -34.8773 by the linear fit. it still doesn't fit.

Pat,

I ran the regressions only on the "data" from Andre's post starting with 120 seconds, the point at which failure kicks in, and longer. So both the power and linear equations only describe behavior beyond 120 seconds, and should only be applied there. Interesting that the power curve does describe behavior pretty accurately below the data to which it was fit.

Lee

 

[Chazzy]

Lee, I don't understand how your chart would be used in the field. I see metered times, and under each film the numbers do not look like corrected times.

 

[Lee L]

Lee, I don't understand how your chart would be used in the field. I see metered times, and under each film the numbers do not look like corrected times.

As I said in an earlier post, #13, kinda buried under the list of equations,

The format for corrected times in the chart is mmm:ss

This is to make it easier to use with a watch or timer. I have a Gossen Digiflash meter that has a countdown timer that goes up to 30m 59s and does countdown beeps for the last 10 seconds with a long ending beep, so the chart was aimed at using that way. This chart is from a spreadsheet, and I didn't bother to type in a formula to convert times over 60 minutes to hh:mm:ss.

The grayed lines are for the full stop times displayed on most meters: 1s, 2s, 4s ..... 1m, 2m, 4m, etc. The others are approximately 1/3 stop steps because my most used meters are marked that way and non-linear interpolation "in my head" between full stops can be a brain bender.

Lee

 

[gainer]

If you have a pocket calculator such as the T1-30XIIS, you need only know 1 number for each film you use and one constant, 1.62, that is good for all films. The sequence of entries is as follows, where Af is the film constant, tm is the metered time, and tc is the corrected time.

tm^1.62*Af+tm = tc

Let's say that your film requires 0.5 seconds correction at tm=1 second. For that film, Af = 0.5. Now you're out shooting lumps of coal at in the deep woods (a common, though not often photographed, sight in West Virginia) and your meter tells you it will take 100 seconds. (Wish I had such a meter.). You whip out your TI30 and do:

100^1.62*.5+100 and the answer is 20424.9 seconds. But suppose you have another film with Af = 0.1. then:

100^1.62*.1+100 = 273.8

 

[gainer]

Lee:

Assuming that your equation gives an accurate corrected value for Acros at 120 seconds metered, I calculate by your power fit numbers that the corrected exposure is 165 seconds. Working backwards, I calculate that the Af coefficient in my equation is 0.0193. Now I find that at 60 metered seconds, Acros need 75 seconds, at 30 seconds it needs 35, and at 1 second, it needs 0.019. I would say that the correction is negligible at 30 metered second or less.

 

[Lee L]

I'm aware of only a few people who have published very good reciprocity data. Bond is the only one I've seen do the really heavy lifting, testing carefully at a wide range of times. Others use a long-standing, sometimes modified, Schwarzschild calculation, based on the necessary adjustments in stops between a 1/8th second (0.125 second) exposure and an exposure with a 3.0 log density (10 stops) filter at 125 seconds (sometimes at 128 seconds) to get equal density. You have to be careful here to get a really neutral filter. The B+W 110 and Wratten ND filters are often recommended, with the caveat that the Wratten gel is too leaky in the red for accuracy with some films. Covington and Reeves in their books on astrophotography outline the procedures for this method. The result is a Schwarzschiild exponent 'p', which can be used to calculate corrected exposures with the formula:

corrected time = (( metered time +1)^(1/p))-1

Covington notes that he has seen significant batch-to-batch differences in the same film. If anyone's interested, I can post Covington's and Reeves' films and Schwarzschild exponents. There is some significant variation between their results using exactly the same procedure and films, but a few years apart.

My point is that none of this is pinpoint accurate. Gainer's method, the power and log formulae, and the Schwarzschild formula can all be fit to any experimental data I've come across within about 1/3 stop or better. If anyone assumes that any experimental data they've seen has a margin of error of less than 1/3 stop at these extended times, they're probably fooling themselves. Most of the data floating around the web is not documented at all. Some people have posted fourth order polynomial expressions in an effort to model experimental error and hit each observation exactly.

The difference between the 120 and 165 second example in Gainer's last post for Acros is about 1/3 stop.

So my point is that you can just pick your mathematical poison among several reasonable models for reciprocity failure, and get within a fraction of a stop, assuming the models are built on decent data such as Bond's. I'm not interested in worrying about thirds of stops or less.

I like working with charts in the field, which is why I produced the one I did. It's only as good as the data, none of which I've confirmed by exhaustive work like Bond's, but I've tried to choose the data reasonably carefully.

Lee

 

[gainer]

When it comes to pinpoint accuracy, reciprocal behavior of film exposure-density is not a good place to look for it, not is it needed when one considers the logarithmic transformations used to get a nearly straight line of exposure vs density. Considering the personal equations involved in the processes that take one from exposure to print, it is a wonder any of our charts and tables come close to agreeing.

When it comes to simplicity of use, I'll put my equations and a TI-30XIIS against any set of charts for anyone who uses more than one film. I haven't seen any comment on my posting of the reverse engineering of Acros. If anyone has data for metered exposures between 1 and 120 seconds, please post it. I do not use the stuff yet.