Which ND keeps your stop in this light?

What the tool answers

A shot designed at T2 with a 180° shutter has a fixed exposure. Outside, the sun decides how much light arrives, and the ND is the only part of the chain that does not change the image. The tool works out how many stops have to go, which filters from your kit get closest, and what is left over for the iris.

The stops ladder shows the gap to close and the glass that closes it, internal ND in blue, filters in green. Below, the same exposure in other combinations: a stop deeper, a narrower shutter, a different ISO — each with the ND it would need, so you can decide whether the look or the filter swap is cheaper on the day.

How we calculate

Light is expressed as an exposure value at ISO 100 (EV100). A camera setting needs EV100 = log2(N² ÷ t) − log2(ISO ÷ 100), where N is the stop and t the exposure time. The scene is too bright by the difference between the two, in stops. Each 0.3 of optical density removes one stop, because 10^−0.3 ≈ ½.

Exposure time comes from frame rate and shutter angle: t = angle ÷ 360 ÷ fps, so 180° at 25 fps is 1/50 s and at 24 fps 1/48 s. Light situations use the standard EV100 table for front light between about two hours after sunrise and two hours before sunset; side light costs roughly one stop, backlight two. A meter reading is converted the same way.

For the filter kit the tool tries every combination of up to two filters plus each internal ND position and picks the one closest to what you need, preferring less glass. The remainder is turned into an iris correction: N × 2^(remainder ÷ 2).

EV100 needed = log2(N² ÷ t) − log2(ISO ÷ 100) stops over = EV100 scene − EV100 needed ND density = stops × 0.3 transmission = 2^−stops t = shutter angle ÷ 360 ÷ fps

Where the tool is wrong

• The EV table is a starting point. Light on a real location varies by a stop or more with season, altitude and reflections. Meter the subject and enter the reading — that replaces the table entirely.
• T-stop is not f-stop. The maths treats the stop you enter as the effective aperture. Cine lenses are marked in T-stops, which already include transmission loss; with photo lenses expect the image to be up to a third of a stop darker than the f-number suggests.
• Filters are not exactly their number. A stated ND 0.9 can sit a few hundredths off, and stacked filters add reflections and a slight colour shift. Check a grey card on the monitor after building the stack.
• Infrared. Dense plain ND passes a larger share of far-red and infrared, which many sensors still record: blacks go magenta-brown, foliage muddy. The tool cannot see your filter type — with heavy stacks, use IRND glass.
• Flicker depends on the lamp. The check covers the classic case: light pulsing at twice the mains frequency. Many LED fixtures run their own PWM frequency, which no shutter angle fixes; test at the shooting frame rate.

From the set

A summer exterior in full sun, Alexa 35 at EI 800, T2.8, 25 fps and 180°: the scene is 9.4 stops too bright, which is ND 2.8. The internal 1.8 plus an external 0.9 gives nine stops; the remaining 0.4 stops go into the iris, T3.2 instead of T2.8. Many older ND tables quote six stops for this situation — on the day that would leave the image more than three stops over.

The same camera on a beach, EV 16, T2 and 172.8° at 24 fps to stay flicker-safe under practicals: 11.4 stops. Internal 1.8 and an external 1.5 make eleven, the iris goes to about T2.3. At a total density of 3.3 this is exactly where plain ND shows infrared contamination — the 1.5 should be an IRND, and the colourist should get a grey card at the head of the roll.