Long exposures, counted properly
Stack ND filters in whatever units they are labelled in, get the corrected shutter time, and run a timer that keeps counting when you switch apps.
ND filter calculator
Filters on the lens
Filters are labelled three different ways and the numbers look nothing alike. All three mean the same thing here.
Counts from the clock, not from a tick loop — switching apps or locking the phone will not lose time.
Nothing leaves this device
How the three labels convert
corrected time = metered time × 2^(total stops)
Stops is the direct figure. Optical density divides by 0.301 — each 0.3 of density is one stop, because that is log₁₀2. ND factor (the number in "ND1000") is how much the light is divided, so its stops are the base-2 logarithm.
Stacked filters add their stops. Two six-stop filters make twelve stops, not twelve times the exposure.
A detail worth knowing: an "ND1000" is sold as ten stops but density 3.0 is really 9.97 stops. The difference is a fifth of a percent of exposure time and nobody will ever see it, but this page reports the honest figure rather than the marketing one.
What the arithmetic cannot tell you
Reciprocity failure. Digital sensors are close to linear, so the calculated time is usually right. Film is not — past a second or so most emulsions need considerably more exposure than the maths says, and each stock behaves differently. The extra compensation field is there for that; your film's datasheet has the numbers.
Colour cast. Strong ND filters, especially stacked ones, rarely stay neutral. Ten stops of cheap resin often lands somewhere magenta or brown. Shooting raw and correcting later is the usual answer; a custom white balance through the filter is the better one.
Light leaking past the filter. On very long exposures, light entering through the viewfinder or around a square filter holder can fog the frame — the exposure is long enough for a small leak to matter. Cover the viewfinder. This is the most common cause of a mysteriously washed-out long exposure.
Sensor heat. Past a few minutes, sensors warm and noise rises, unevenly and more in warm weather. Bracket rather than trusting one frame.
What each range of stops is actually for
3 stops takes a 1/125 to a 1/15 — enough to blur a waterfall in shade, or to open a lens up in daylight for shallow depth of field.
6 stops reaches roughly half a second in daylight: moving water goes smooth, people walking become ghosts, and clouds start to streak.
10 stops is the landscape standard. Bright midday becomes 8 to 30 seconds — water turns to mist, clouds draw lines, and anything moving disappears.
15 stops and beyond pushes daylight into minutes. This is where filter quality, viewfinder leaks and sensor heat all start to matter more than the calculation.
Related
The exposure comparator solves the same relationship without filters. For blur from a moving subject rather than a long exposure, see depth of field for the sharpness side.
Questions people actually ask
Do stacked ND filters multiply or add?
Their stops add. A six-stop and a three-stop together give nine stops, which is 512× the exposure — not 18 stops and not 6×3. Their ND factors multiply, which is the same statement: ND64 × ND8 = ND512.
Why does my ten-stop filter need more exposure than calculated?
Three usual causes. On film, reciprocity failure means long exposures need more light than the maths predicts, and every stock differs. On digital, the filter may simply be denser than labelled — cheap ten-stops are often 10.5 or 11. And light leaking through the viewfinder or around a square holder fogs the frame rather than exposing it, which looks like underexposure of the subject.
Should I meter before or after fitting the filter?
Before, then apply the correction. Through-the-lens metering behind ten stops of glass is working in near darkness and most cameras hunt or give up. Autofocus too — focus and meter first, then fit the filter without touching anything.
What is the difference between ND8, 0.9 and 3 stops?
Nothing. They are three labels for the same filter: ND8 divides light by 8, density 0.9 is 0.9 ÷ 0.301, and both come to three stops. Manufacturers pick whichever convention they like, which is why this page accepts all three.
Does the timer keep running if I switch apps?
Yes. It reads the system clock rather than counting ticks, so it stays accurate whether the tab is backgrounded, the phone is locked, or the browser throttles the page. Come back and the remaining time is correct.
Why does it say 9.97 stops for a filter sold as 10?
Because density 3.0 really is 9.97 stops — 3.0 ÷ log₁₀2. The round number is marketing. The difference works out to about a fifth of a percent of the exposure time, so it will never show in a photograph, but the page reports what the filter does rather than what the box says.