Reciprocity Failure in Practice: Worked Exposure Examples
Our reciprocity failure primer covers why the effect happens and where the correction formula comes from. This companion piece skips straight to numbers: a run of metered exposures, worked through the same power-law correction our Reciprocity & Exposure Calculator uses, so you can see exactly how the corrected time, the extra time, and the stops added all move together as the metered exposure grows, from a fraction that needs no correction at all to a multi-minute exposure that needs a great deal of it.
The short end: exposures under 10 seconds
At a metered exposure of exactly 1 second, the correction does essentially nothing: 1 raised to any exponent is still 1, so the corrected time, the extra time, and the stops added all come back as zero regardless of which exponent you use. This is worth internalizing on its own — reciprocity failure isn't a tax on every exposure, it's a tax that only starts to bite once the metered time climbs meaningfully past a second or so. At 8 metered seconds and an exponent of 1.3 — a reasonable working figure for many historical emulsions — the corrected time comes out to 14.93 seconds, an extra 6.93 seconds, or 0.9 of a stop. Push the exponent to 1.5 for the same 8-second metered reading, representing a more reciprocity-prone material, and the corrected time jumps to 22.63 seconds, an extra 14.63 seconds, a full 1.5 stops. The same input time, run through two plausible exponents for two different real materials, produces meaningfully different real exposures — which is exactly why the calculator asks for an exponent rather than assuming one.
The middle range: 10 to 45 seconds
A metered 10-second exposure at an exponent of 1.3 corrects to 19.95 seconds — an extra 9.95 seconds, essentially a full extra stop (1.0) on top of the metered reading, almost exactly doubling the nominal time. Stretch the metered reading to 30 seconds at the same 1.3 exponent, and the correction grows disproportionately: 83.23 seconds corrected, 53.23 seconds of extra time, 1.47 stops added — the corrected exposure is now nearly triple the metered figure, not just double. Raise the exponent to 1.6 for that same 30-second metered reading, modeling an older or more reciprocity-prone plate, and the corrected time balloons to 230.88 seconds — just over three minutes and fifty seconds — a full 2.94 stops of correction, meaning the real exposure is roughly seven and three-quarter times longer than the metered reading suggested. At 45 metered seconds and an exponent of 1.4, the corrected time comes to 206.3 seconds (about three minutes and twenty-six seconds), 161.3 seconds of extra time, and 2.2 stops added.
Long interior and low-light exposures: 60 to 120 seconds
Once metered times reach a minute or more — a realistic figure for a dim interior on a slow historical process — the corrections stop being a minor adjustment and start dominating the actual working exposure. A 60-second metered reading at an exponent of 1.4 corrects to 308.61 seconds, or a little over 5 minutes 8 seconds: 248.61 seconds of extra time and 2.36 stops added, meaning the real exposure runs more than five times longer than the meter alone would suggest. A 90-second metered reading at 1.3 corrects to 347.15 seconds (about 5 minutes 47 seconds), an extra 257.15 seconds and 1.95 stops. And a full 120-second (2-minute) metered reading, still at 1.3, corrects to 504.59 seconds — about 8 minutes 25 seconds — an extra 384.59 seconds and 2.07 stops added, meaning the plate actually needs roughly four and a quarter times the metered figure.
Comparing exponents side by side at one fixed time
It's also useful to hold the metered time constant and vary only the exponent, to see how much a single, seemingly small difference in material behavior changes the outcome. At a metered 8 seconds: an exponent of 1.3 corrects to 14.93 seconds (6.93 extra, 0.9 stops); an exponent of 1.5 corrects to 22.63 seconds (14.63 extra, a full 1.5 stops). That's roughly a 51% difference in the corrected exposure time between two exponents only 0.2 apart on paper. At a metered 30 seconds, the same 0.3 exponent gap (1.3 versus 1.6) separates an 83.23-second correction from a 230.88-second correction — nearly a threefold difference in the final exposure from what looks, at a glance, like a fairly modest difference in the underlying material constant. This is the clearest illustration of why "just use 1.3, it's usually about right" is risky advice: the same modest-looking uncertainty in the exponent turns into a large, easily plate-ruining uncertainty in the actual exposure once the metered time gets long.
Translating stops added into f-stops, if that's more intuitive
Photographers who think more naturally in aperture stops than in raw seconds can use the "stops added" figure directly, since it's already expressed in the same doubling-based unit an aperture ring uses. A correction of 1 stop (the 10-second example above) is equivalent to opening the aperture one full stop wider while keeping the metered time unchanged, or equivalently, doubling the exposure time at a fixed aperture — both routes deliver the same extra light. A correction of 2.36 stops (the 60-second, 1.4-exponent example) doesn't map onto a single closer full aperture stop, since aperture stops are also spaced in doubling increments and 2.36 falls between the 2-stop and 3-stop marks, but it does tell you that opening two full stops still leaves the shot meaningfully under the corrected requirement, while three full stops slightly overshoots it. In practice, most historical process photographers find it easier to work directly in the corrected seconds this calculator provides rather than translating back into fractional aperture stops, but the equivalence is there if your workflow is built around thinking in stops.
What the pattern tells you
Lay these numbers side by side and a clear pattern emerges: the ratio between corrected time and metered time keeps growing as the metered time grows, for any exponent above 1. That's the defining signature of a power-law relationship rather than a flat percentage add-on — a "just add 20%" rule of thumb might work passably at short exposures and fail badly at long ones, or vice versa, because a genuine power law doesn't scale that way at all. It's also why treating any single stops-added figure as universal is a mistake: 0.9 stops at 8 seconds and 2.36 stops at 60 seconds, from exponents that only differ by 0.1, show how sensitive the real-world correction is to both the exposure length and the specific material's exponent. Neither number transfers cleanly to a different exposure time on the same material, let alone to a different material entirely.
Where the metered time itself comes from
Every example above starts from an already-metered time, but it's worth remembering that the metered time for a historical process is itself the output of a separate correction — converting a modern meter's ISO-100-or-400 reading down to the process's true, much slower ISO, as covered in our ISO↔EV Converter and its companion guide. That conversion is what typically produces the long metered times — 30 seconds, a minute, two minutes — that make reciprocity failure relevant in the first place. A process running close to a modern film speed rarely needs a reciprocity correction at all, simply because it rarely needs an exposure long enough to trigger one; it's specifically the combination of a very slow historical ISO and a not-especially-bright scene that produces the multi-second and multi-minute metered readings this page's worked examples are built around. Treat the two corrections as a pipeline: the ISO conversion tells you roughly how long the metered exposure needs to be in the first place; only once you have that figure does it make sense to ask whether reciprocity failure applies to it, and if so, by how much.
Using this table as a planning tool
In practice, these worked figures are most useful as planning anchors rather than exact predictions for your own plate. If a meter reading for a dim interior portrait comes back around 30 seconds and you're working with a material you'd estimate somewhere in the 1.3–1.4 exponent range based on prior results, you now have a concrete expectation — somewhere between roughly a minute and a half and three and a half minutes of actual exposure — before you even open the shutter, rather than discovering the true figure only after developing a badly underexposed plate. That expectation is exactly what lets a photographer plan around a long exposure practically: briefing a sitter on how long they'll need to hold still, choosing a headrest or support, and picking a moment when nothing in the frame is likely to move unexpectedly.
Run your own metered time and exponent through the Reciprocity & Exposure Calculator directly, see Understanding Reciprocity Failure in Historical Photographic Processes for the underlying theory, and read Why Victorian Portraits Look the Way They Do for what exposures in this range actually meant for the person sitting in front of the camera. For these same figures alongside ISO/EV and print-size tables in one place, see our Exposure & Print Planning Reference.