Some films have extremely severe reciprocity departure. Most manufacturers have that info on their websites. T-Max 400 has very low reciprocity failure. Still you should make tests and then come up with exposures for different scenes, depending on their inclusion of light sources. I have one exposure and developing time for city streets. I haven't needed to alter that in years. If you're out there doing star trails in the wilderness, when you get the right exposure, it will work most of the time. But trial and error are necessary.
I calculate my reciprocity adjustments, such as in the studio, by find a taking reading for a very small aperture, smaller than my lens even has. For my Commercial Ektar, I find an exposure for f-90, or f-128, and then open the number of stops the manufacturer recommends. This forgoes having to change the time yet again. Eventually experience will guide you.
It depends on the approach you're taking. If I tell you that film A needs three stops of compensation at 30 seconds exposure you can do one of two things. You can open up the aperture three f-stops and you're fine. But if you try to double the exposure three times, you're going to get compound reciprocity failure because at the adjusted exposure of 240 seconds, you've got additional reciprocity to deal with, not just the adjustment for 30 seconds. You can calculate that adjustment with one of several formulae for adjusting for reciprocity with exposure time (as opposed to adjusting with lens aperture).
Originally Posted by holmburgers
There's a marathon thread on a lot of this here: http://www.apug.org/forums/forum37/1...sbehavior.html and there have been subsequent threads as well.
The Howard Bond article discussed there is currently available at: http://www.willwilson.com/articles/0...ciprocity2.pdf
Most manufacturers reciprocity data for B&W film (especially Kodak and Ilford) has not changed in the last 30-40 years, while the films have changed a lot. See the Bond article for comments on that.
In selecting a meter for low light you should pick one on how low light is capable of not how long the exposure time can be set. Among the low light meter, like someone has said the Gossen Luna Pro can go down to EV-4@ISO100. The Minolta booster unit can get my Flashmeter III to read down to EV-6@ISO100 but with an angle of 60 degrees. Of course, with long exposure time there is the reciprocity failure problem which there is no meter that would handle this automatically.
Once you have the meter reading the calculation can be done manually with the same accuracy.
That is not correct. Reciprocity failure is due to the film responding poorly to low luminous flux, i.e. the loss of sensitivity is a function of flux not exposure time therefore the correction factors listed by manufacturers are a time-extension factor.
Originally Posted by Lee L
Say the table that comes with the film says you need a 1-stop correction, you double the exposure time. If you were to open the aperture by one stop instead, you would double the flux hitting the film and thereby improve the reciprocity performance, which would result in a more-than-one-stop increase in exposure, and an over-exposure (fine for negatives, bad if you're shooting chromes at night).
There's a further wrinkle in that reciprocity failure increases contrast, therefore increasing the exposure isn't a complete fix. I believe the manufacturers tables aim to maintain the mid-tones, but it's possible that that's not universally true.
Anyway, the easiest solution is to buy a film with excellent reciprocity performance like Acros, TMY2 or (for colour) Provia; they require almost no correction at all and therefore have very little contrast expansion, even for exposures of half an hour.
And for metering, the best solution is to use a DSLR set to a very high ISO setting e.g. 3200 and wide-open (about f/2). Even if the DSLR's meter is not accurate in the dark, the sensor is itself a very powerful and sensitive light-meter, particularly if you learn to read the histogram.
Say for example you get a well-exposed image (according to histogram) at ISO1600, f/2.8, 2.5s and you want to shoot your film at ISO100 (Acros) and f/16. That means you need +4 stops for the ISO change and +5 stops for the aperture change, so +9 stops from 2.5s is 1280s. At that point you'd add about 1/2-2/3 stop of reciprocity correction for Acros, so about 2000s (30-35 minutes) total exposure depending on how bored/fidgety/cold you get.
If you use a film with poor reciprocity performance, you can expose it all night and still get less image than the Acros or TMY2 collected in half an hour.
If you're only looking at Kodak T-Max film technical data you might get that impression. However if you look at Kodak Tech Pubs F4017 and F-4018 for Tri-X and Plus-X films, you'll see a significant difference between the recommended adjustment in aperture and the recommended adjustment in exposure time. For each of these films you get +2 stops of aperture adjustment at 10 seconds and +3 stops in aperture adjustment at 100 seconds. That would equal 40 seconds of adjusted exposure at 10 seconds measured and 800 seconds adjusted exposure for 100 seconds measured if there were no compounding effect with reciprocity adjustments using time rather than aperture. However, the tech pubs recommend time adjusted exposures of 50 seconds from 10 seconds and 1200 seconds from 100 seconds.
So with Tri-X and Plus-X that's 80% of correct exposure if you apply the aperture correction to time at 10 seconds metered and 66% of correct exposure if you apply the aperture correction to time at 100 seconds metered, the compounding effect of extending by time for reciprocity correction failure.
In the case of the T-Max films, which have less reciprocity failure, Kodak is just rounding down because the reciprocity failure of those films only amounts to about 5% underexposure at the adjusted time for 10 seconds and about 15% underexposure at the adjusted time for 100 seconds. Few people meter that accurately in the first place.
You might want to look at the Schwarzschild formula or Covington's modification of it, or at Gainer's reciprocity formula and variations on that to understand the compounding effect of adjusting for reciprocity with extended times.
I assume you mean the table on Page 2 of F-4017; that's demonstrating what I mean in that adjusting by aperture is not the same as adjusting by time because adjusting by aperture improves the film's sensitivity (reduces the reciprocity failure). Taking the metered-100s example, there is approximately is 3.6 stops (12x) of reciprocity failure but if you applied that much correction to the aperture, it would be an over-exposure. This is because the film responds non-linearly to light at this point and opening the aperture means increasing the light flux and therefore increasing the film's sensitivity. Changing the aperture means changing the flux, which means you are now at a different point on the film's reciprocity curve.
Say you meter at f/16 100s. At that level of luminous flux on the film, the film has 1/12 of the normal sensitivity so you would need to extend the exposure by a factor of 12. You don't do any further compounding to that adjustment, you just apply it to the time. Single step, no compounding. Look up the table, apply the multiplier to the time.
If you instead open the aperture, you need to open it by less than 3.6 stops or you'll get additional exposure, hence the table tells you to use only 3 stops. Opening by 3 stops means you're now shooting f/5.6 for 100s (meter says 12.5s). That's actually pretty close to the metered situation on the previous line of the table (10s requires +2 stops aperture or +2.3 stops time), for which the table would recommend about f/5.6 70s. The 70s/100s difference is mostly down to the change in recommended development time. Arguably you would be better off just using the previous line of the table if you can reasonably open the aperture that far - in other words, treat your meter as saying f/5.6 12.5s, go to the second-last line in the table, adjust it to 70s and shoot.
Using the aperture adjustments (instead of time adjustments) in that table will result in lower-contrast negatives too. From the f/16 100s situation: if you start at the last line and aperture-adjust you will shoot f/5.6 100s at -30% development but if you start at the second-last line (from the same meter reading!) and time-adjust you will shoot f/5.6 70s at -20% development. Those negatives will both be fine but they will be different. The -30% contrast adjustment is intended for use shooting at f/16 (in this light level) so if you instead use it to shoot at f/5.6 (in the same light level), you will produce a contraction of about 1 stop. The higher flux at f/5.6 means that there is less contrast expansion (we're on a flatter part of the exponential reciprocity-failure curve), hence the contraction.
When shooting at night, I don't think it's ever really feasible to go around opening the aperture by 3+ stops because you either won't have enough DOF or your aperture just isn't that big. So it's much simpler to just get the time-adjustment table (which, we agree, is not the same as the aperture-adjustment table) and apply that to the time... once, not compounded in any way.
It's also worth noting the development adjustments in that table. 20-30% is a significant adjustment in contrast and it's required because severe reciprocity failure increases contrast. In effect, those large time increases in conjunction with the development reduction amount to a pull of the film in order to retain shadow detail and tame the contrast. In other words, the time-adjustment table provided by Kodak is a preserve-the-shadows adjustment, not a preserve-the-midtones adjustment.
(I'm pretty sure we are in complete agreement wrt mechanism and what ones does when shooting, but your use of the term "compounding effect" wrt time-adjustments muddies the situation. Yes, we agree that sensitivity loss and therefore exposure time is exponential).
The BEST light meter for night photography used to be made by Quantum Instruments, and it is called Calculight XP. I bought it in 1990s but it was discontinued around the turn of the century. Even to date I haven't seen a light meter as sensitive as this one.
None of those people understood the chemical mechanism of reciprocity failure, nor did they try to. Those equations are empirical fit, and they don’t fit REAL data (not the stuff the manufacturers provide as a rule of thumb for practitioners).
Originally Posted by Lee L
The reason for reciprocity failure is instability of latent subimage. A silver halide crystal needs to be hit by several photons before a stable latent image is formed. If the illuminance is very low, photons hit a crystal very infrequently and thus latent subimage can be destroyed and go back to square one before additional photons strike the same crystal.
The lifetime of the latent subimage, as well as the “threshold” of the latent image depend a lot on the emulsion crystal technology as well as the chemical sensitization techniques used. Generally speaking, small tabular grain crystals are the best.
I have been using a JTL LM8 Light Meter for last three years and am very satisfied with it's performance.
It has the range you are looking for.
Shutter speed range 30 min. to 1/8000 sec. (1/3 step)
Aperture range f1.4 to f90 in 1/10 stop increments
EV scale -6 to +25.9 in 0.1 increments
ISO range 3 to 8000
Your meter doesn't have to have the same settings as your camera, that is the main thing to understand. So what you do is get a meter that allows the lowest possible f number (e.g. f/1) and the highest possible speed (E.g. 6400) and you make yourself an equivalent exposure chart or spreadsheet. it is quite easy. With your meter set at f/1 and ISO 6400, you should be able to get a decent reading at least somewhere in the scene. Then, once you have that exposure, you have to do a reciprocity correction. I used to have equiv exposure and reciprocity charts in my smartphone for handy reference.
A common approach is to pick the very brightest part of the scene and let that go to the highest zone. So e.g. if you're photographing a moonless sky, you let the stars be the brightest objects in the scene, and there are exposures you can look up that are based on star magnitudes etc. If there is moonlight, then there are charts for that too, which depend on phase and inclination etc. Yeah... most people prefer just to wing it!
In the very worst case, if all else fails, you just shoot whatever exposure you can and develop by inspection (or shoot two and develop one to your best guess, and get it right on #2).
With the latest full-frame dslrs, you can also get very precise readings in near total darkness. With a d700, I shot things that I couldn't actually see. That camera is spooky :)