Reciprocity failure

Shooting film · ~7 min read

The reciprocity law says exposure = intensity × time, and that any equal product gives equal density: f/2 at 1/1000 s = f/8 at 1/60 s. It holds beautifully across the normal range – and breaks down at long exposures (roughly past 1 second) and at extremely short ones (past ~1/10000 s, flash and high-speed work).

In short

  • Past ~1 s, film becomes progressively less sensitive than its rating – a metered “8 s” exposure might really need 20 s or 60 s.

  • Cause: at low light, latent-image specks form so slowly that some decay before they reach the stable ~4-atom threshold.

  • Every film has its own reciprocity behaviour. Fuji Acros 100 II and Provia 100F are exceptional (little/no correction to ~2 min). Traditional films (Fomapan, Pan F) are poor.

  • Colour films also shift colour, and each layer shifts differently → a cast that needs a CC filter to fully fix.

  • Practical rule: meter, then apply the film’s published correction (a table or formula), and for colour add the suggested filter.

Why it happens

Forming a latent image needs ~4 silver atoms to cluster on a crystal before the cluster becomes stable (see How a photograph forms). Photons must arrive fast enough that the second, third and fourth atoms form before the first ones regress back to silver ions.

  • Normal light: photons arrive in a steady stream; specks build efficiently.

  • Very dim light (long exposure): photons trickle in. A lone silver atom sitting on a crystal may revert before the next photon arrives, so a lot of absorbed light is “wasted” and the film acts slower than rated. This is low-intensity reciprocity failure (LIRF) – the one that matters for night and astro work.

  • Very intense, ultra-brief light: too many electrons arrive at once for the silver ions to keep up; also inefficient. High-intensity reciprocity failure (HIRF) – matters for electronic flash duration studies and scientific work, rarely for pictures.

Compensating – black and white

Published data comes as a table or a formula (t_corrected = t_metered ^ p, with p a film-specific exponent around 1.2–1.6, or a Schwarzschild constant). Rough guide for common films (always check the current datasheet):

Film

metered 1 s → give

metered 10 s → give

metered 30 s → give

notes

Fuji Acros 100 II

~1 s

~10 s

~30 s

best-in-class: no correction at all up to 120 s, then only +½ stop out to 1000 s. Same behaviour as the original Acros.

Ilford Delta 100 / FP4+ / HP5+

~1.3 s

~26 s

~2 min

Ilford publishes a simple curve; moderate correction

Kodak T-Max 100 / 400

~1.2 s

~15 s

~50 s

good; T-Max films are relatively reciprocity-tolerant

Kodak Tri-X 400

~2 s

~50 s

~5 min

significant – old-school emulsion

Fomapan 100 / 200 / 400

~2 s

~1–2 min

~5–10 min

poor – Foma’s tables are steep; test before relying on it

Ilford Pan F Plus

~2 s

~1 min

~4 min

poor; also fades latent image – not a night film

For contrast: long exposures also raise contrast (highlights keep building while shadows stall), so B&W night work often pairs reciprocity compensation with −10 to −20% development.

Compensating – colour

Colour negative and slide films add a colour shift because the three emulsion layers fail reciprocity at different rates. Datasheets give both a time correction and a CC (colour-compensating) filter.

Film

Long-exposure behaviour

Typical fix at ~1 min

Fuji Provia 100F

excellent – minimal shift

no correction to ~128 s; the astro/night slide film

Kodak Portra 400 / 800

good to ~1 s, then modest; slight warm/blue shift

+⅓–1 stop, sometimes CC05–10 correction

Kodak Ektar 100

shifts blue/cyan and loses shadow colour

+1 stop, CC filtration; not ideal for long exposures

Fuji Velvia 50

poor past ~1 s – strong green/blue shift, contrast spike

+⅔–1⅓ stop and a CC magenta filter (e.g. CC10–15M); avoid for very long work

CineStill 800T

designed with night in mind; modest correction, some cyan shift

+⅓–⅔ stop past ~1 s

Fun fact

Astrophotographers in the film era obsessed over reciprocity because a deep-sky exposure could run 90 minutes. Kodak sold hypersensitised film (Tech Pan baked in forming gas – a hydrogen/nitrogen mix – beforehand) that dramatically reduced LIRF for exactly this audience. Digital sensors don’t suffer LIRF at all, which is a big reason astrophotography went digital early and completely.

Practical workflow for a long night exposure

  1. Meter normally. Suppose it says f/8, 4 s.

  2. Look up the film’s correction for 4 s. Say it’s ×3 → give 12 s.

  3. If colour, add the suggested CC filter (and its ⅓-stop factor → ~15 s).

  4. If B&W and the scene is contrasty, plan N−1 development.

  5. Bracket +1 stop as insurance – reciprocity data is approximate and temperature- dependent (it’s worse in the cold).

  6. Use a locking cable release / B or T mode, and a timer.

Where to go next