I have been doing some long exposures on FP4 and am experiencing reciprocity failure . Using the adjustment curve supplied by Ilford on their FP4 data sheet a 20 second metered exposure should be adjusted to approximately78 seconds This is much more than the 1.45 factor mentioned elsewhere in the APUG threads on reciprocity. What am I missing ?
Reciprocity failure is controlled by many variables and there's no guidelines that suit evrey possibility. Ilfords chart is just a starter you raelly need to do your own tests.
Reciprocity will differ in daylight and artifical light, actual light intensity and choice of format, lens and exposure. Something you shoot at f2 and 1/30th with 35mm needs 8 seconds at f64 with an LF lens same amount of light though.
It's just a case of bracketing and remebering what works best for you.
The failure rate is not a straight line curve, but instead increases (curve gets steeper) as the metered exposure gets longer -- thus a singe multiplication factor will not be accurate
The "factor" would be the exponent in the equation, not a straight multiplication factor, hence the curve.
Ilford pretty much uses one graph for all their films so I wouldn't rely too much on them. Kodak's publications are similarly murky and/or incorrect.
Some of the more useful data I have seen comes from tests performed by Howard Bond on several films (unfortunately FP4+, one of my favourite films was not one of them):
The chart is at the end if you don't want to read through the entire thing, but it's worth the read.
Based on my experiences with T-Max 100 and Delta 100, Bond's results work quite well. I did my own tests and came up with similar (although not identical) results.
Thanks -- I was assuming something simpler.
I think my last exposure on FP4 was on News Years Day. The meter said 8 minutes so I was going to go for 30 mintues. Noticed a slight drop in the light partway thru the exposure, so I went for 60 minutes -- nice negative!
Reciprocity correction is not linear. You can't just multiply a meter reading by a factor and come up with a valid result.
1.45 is an exponential factor not a scalar...
20^1.45 ~ 77 secs
at which time it becomes appropriate to say: acros is very nice when it comes to reciprocity... :)
The Ilford formula is a generic one for B&W film used by both Kodak and Ilford since at least the late 60's, and has not changed over time with changes in the film. Most critical users report significant differences between that formula and the behavior of modern films.
Attached are two .pdf files. One is a table of adjustment times with adjustment advice calculated from three sources:
Leon Taylor's tests at:
Robert Reeves' Schwarzschild exponent testing:
Robert Reeves, Wide-Field Astrophotography, Willmann-Bell, ISBN-13: 978-0943396644
Taylor's and Reeves' empirical tests differ significantly from the generic Ilford curve suggestions, but only about a stop or so from each other. You'll see that deviation in the attached graph. I have friends using the Reeves numbers with FP4+, including Shawn Dougherty here on APUG who report excellent results.
The metered times in the first column are all in seconds. The adjusted times for Ilford, Taylor, and Reeves are shown in HH:MM:SS (hours:minutes:seconds) format for ease of timing in the field. The formula used to regress the Taylor and Reeves test data is at the top and the coefficients a and exponents b for that formula are also listed. This is the same formula as Patrick Gainer used, but for calculating a chart to be used in the field I prefer to use a variable exponent rather than the fixed exponent of 1.62 that Gainer found to work well for all the films in the test data he used. I find that varying the exponent more accurately models a wider variety of films, especially those with low reciprocity like Acros.
You can find a lot of discussion of Gainer's formula here on APUG by searching for reciprocity misbehavior.
The last time I posted on reciprocity failure someone said that the books on astrophotography and Gainers' derivations that I referenced were not based on empirical data. Reeves and Covington both tested reciprocity failure with documented standard procedures, and Gainer used empirical data from a carefully done study of five films by Howard Bond. So you may get advice that this post is a complete waste. I'm not interested in arguing online, so you can use this info as you see fit,or ignore it.