What is reciprocity failure?
In normal exposure ranges, the relationship between shutter speed and aperture is reciprocal: halving the time and doubling the light (or vice versa) produces the same density on film. However, at very long exposure times (typically beyond 1 second), film becomes progressively less sensitive to light. This phenomenon is called reciprocity failure, or the Schwarzschild effect
after its 1899 discoverer.[1] A metered exposure of 10 seconds might actually require 30 to 50 seconds to achieve correct density; at 60 seconds metered, real exposure might need 5 to 10 minutes.
Why it happens — the physics in brief
Photography's latent image forms when enough photons strike a single silver halide grain in a short-enough time to stabilize a silver nucleus before thermal energy undoes the intermediate states. At long exposures, photons arrive slowly enough that thermal "unfixing" of these intermediate states becomes significant — some would-be latent-image centers dissolve before they can reach the critical stable size. The film has to collect more photons per grain to produce the same density.
At the short end, a related failure occurs for different reasons: very short high-speed flash durations outrun the latent-image-forming chemistry. The short end is rarely practically relevant for conventional photography, but the long end reshapes everyday photographic practice anywhere shutter speeds extend beyond a second or two.
Reciprocity failure bends the toe of the film's characteristic curve (the low-exposure region where shadows live) much more than it bends the shoulder. Long exposures therefore lose shadow detail first — by the time the meter's "correct" time has elapsed, shadows have recorded less density than the meter predicted, while highlights have recorded about what was expected.
When it applies
Reciprocity failure is relevant for:
- Night photography — city lights, neon, moonlit landscapes
- Astrophotography — stars, constellations, Milky Way
(where exposures routinely run 30 seconds or longer)
- Pinhole photography — tiny apertures drive very long exposures even at daylight brightness
- Long-exposure daylight with ND filters — waterfalls, moving crowds, smoothed ocean surfaces
- Dim indoor scenes without flash — churches, caves, interiors by ambient light only
- Any exposure beyond ~1 second — the threshold above which published film data shows measurable correction requirements
Different films have different reciprocity characteristics. Fuji Neopan Acros II 100 is the reciprocity champion among current B&W films — essentially no compensation needed up to 2 minutes, and only modest compensation at 10 minutes. T-grain films (Kodak T-Max, Ilford Delta) generally do much better than traditional cubic-grain stocks. Traditional cubic-grain films (Tri-X, HP5 Plus, FP4 Plus) show the classical reciprocity curve that Adams and Archer described in the 1930s.[1]
Compensation tables
Every film manufacturer publishes reciprocity correction data on the film's technical data sheet. Representative values for common modern films:
| Film | 1s metered | 10s metered | 60s metered |
|---|---|---|---|
| Kodak Tri-X 400 | ~1.5s actual | ~30-50s actual | ~5-10 min actual |
| Ilford HP5 Plus | ~1.3s | ~30s | ~4-6 min |
| Ilford FP4 Plus | ~1.3s | ~25s | ~3-5 min |
| Kodak T-Max 100 | ~1.3s | ~15s | ~2-3 min |
| Kodak T-Max 400 | ~1.3s | ~12s | ~1.5-2 min |
| Ilford Delta 100 | ~1.2s | ~12s | ~2 min |
| Ilford Delta 400 | ~1.2s | ~10s | ~1.5 min |
| Fuji Neopan Acros II 100 | ~1s | ~10s | ~1 min |
| Kodak Portra 400 | ~1.2s | ~15s | ~2 min (with color shift) |
Always consult the specific data sheet for your film; these are representative rather than definitive. Ilford publishes a convenient formula for their films (t_corrected = t_metered^1.31 for most FP4/HP5/Delta-era stocks), which is useful for estimating intermediate exposures not in published tables.
The math pattern
Traditional cubic-grain films approximately follow the Schwarzschild exponential form:
t_corrected ≈ t_metered^p
where p is typically 1.3 for older cubic-grain films, 1.1 to 1.2 for T-grain films, and essentially 1.0 for Acros II (meaning no meaningful reciprocity failure for practical exposures). For a metered 10-second exposure:
- At p = 1.0: t_corrected = 10s (no compensation)
- At p = 1.1: t_corrected ≈ 12.6s (+¼ stop)
- At p = 1.2: t_corrected ≈ 15.8s (+⅔ stop)
- At p = 1.3: t_corrected ≈ 20s (+1 stop)
- At p = 1.4: t_corrected ≈ 25.1s (+1¼ stops)
This pattern lets you estimate compensations for intermediate exposures between published data points by computing t^p directly or reading off a log-log chart.
Color shift considerations
For color films, reciprocity failure does not affect all emulsion layers equally, which causes color shifts. Color negative films may develop a color cast (typically toward cyan or green) that can be corrected in printing or scanning. Color reversal (slide) films are more problematic because their correction flexibility is tighter.
For critical color work at long exposures, use filtration recommended by the manufacturer (a weak CC magenta filter is a common Portra long-exposure recommendation) or choose films specifically designed for long exposures. If you're doing serious astrophotography on color negative film, bracketed testing is the only reliable way to calibrate for your specific stock and subject.
Development adjustments
Long exposures with reciprocity compensation can increase contrast because the shadow areas (which drove the compensation) build up density while the highlights gain proportionally less. More precisely: reciprocity bends the curve's toe more than its shoulder, so extending exposure preferentially strengthens upper zones while shadows remain compressed. Consider reducing development by 10-20% (N-1 development) for exposures beyond 10 seconds to keep highlight density manageable. See [[characteristic-curve]] for the sensitometric basis of this adjustment.
Interaction with the Zone System
Reciprocity failure has a specific Zone System consequence: metered-placed shadows drop one or more zones below their placement target. A shadow placed on Zone III at a metered 10-second exposure may render at Zone II after reciprocity pulls it down the curve's bending toe.
Practitioners working the Zone System at long exposures compensate with deliberate over-placement: meter the darkest important shadow, place it on Zone IV (one zone higher than the usual Zone III), and after reciprocity drop it lands back at the intended Zone III. The workflow otherwise is unchanged.
Stacking with ND filters
Long daytime exposures with ND filters compound two different adjustments: the filter factor (reduction in light reaching the film) and reciprocity failure. Apply them in order:
- Meter without the filter (or compute with it) to get base exposure — say, f/16 at 1/60 s
- Apply the ND filter factor — a 3-stop ND3 filter extends the exposure to 8 times the metered value: 1/60 × 8 = 1/8 s. Still comfortably in the linear range; no reciprocity needed.
- A 10-stop "big stopper" ND extends the same exposure
to 1/60 × 1024 = 17 s. This is firmly in reciprocity territory — compensate the 17-second filtered time per your film's reciprocity table, not the 1/60 original metered time.
Order matters: the filter factor scales exposure duration, and reciprocity operates on the final scaled duration. Getting this backwards (applying reciprocity first, then filter factor) produces an incorrectly-calculated exposure that's usually too short.
Self-testing your own film
For committed long-exposure photographers, bracketed testing against a stopwatch is the most reliable calibration method. Shoot a gray card or standard test scene at the metered exposure plus +¼, +½, +1, +1½, +2 stops of extra time. Develop normally. Examine the negatives on a light table and pick the exposure that renders the scene at desired density. Your personal reciprocity correction for that film, that meter, that developer will be slightly different from the manufacturer's published table — and more useful.
Worked example — starlit landscape
A concrete walkthrough: you're shooting a starlit landscape.
The meter (a handheld or spot meter with long-exposure capability) reads f/4 at 45 seconds for the darkest shadow you want detail in (Zone III placement).
- Apply reciprocity correction for your film. With Kodak Tri-X 400, p ≈ 1.3. At 45s metered: t_corrected = 45^1.3 ≈ 165 seconds ≈ 2 min 45 sec. With Fuji Acros II: ~50 seconds (almost no correction).
- Check Zone System implications. The 45-second placement landed at Zone III via the meter. But 2:45 actual exposure + Tri-X's curve-toe bending means the shadow records at roughly Zone II — about a stop darker than you placed. If Zone III detail matters, bump initial placement to Zone IV (add one stop at the meter step: f/4 at 22 seconds metered → 45 seconds corrected on Tri-X). This over-placement compensates for the reciprocity drop.
- Develop N-1. Highlights that built density proportionally more during the long exposure will overdevelop in normal development. A 10–20% reduction in development time keeps them in range.
- Accept color shift if using color film. Factor in a CC magenta filter on Portra, or accept the correction-in-scanning workflow.
- Shoot and evaluate — and adjust next time based on what the negative actually shows.
This workflow shows the full interaction of the three long-exposure variables — reciprocity, zone placement, and development — that a casual "add some time to the exposure" approach misses.
Pinhole and zone-plate shooting routinely produces shutter times of 5 seconds to several minutes; see Pinhole Photography and Zone Plate Photography.