What can you do with the result?
Use the pixel count to compare recording or scan sizes for the same detail frequency. For example, 30 LP/mm across 36 × 24 mm gives a theoretical reference of 2,160 × 1,440 pixels at margin 1×. Margin 4× gives 8,640 × 5,760 pixels: four times as many along each axis, so sixteen times as many in total. This is a useful planning comparison; the lens and recording chain still determine the captured detail.
Understand the calculation
A line pair is a cycle, not a pixel
LP/mm expresses spatial frequency: one light and one dark line form a pair. At frequency f, a theoretical two-sample-per-cycle boundary corresponds to 2 × f pixels per millimetre. Multiply by the active recording width or height to get the corresponding pixel count. This boundary does not promise reliable contrast at that frequency.
Nₓ = 2 × f × W × S
Nᵧ = 2 × f × H × S
Pixel pitch [µm] = 1000 / (2 × f × S)
W and H are millimetres; S is your sampling-margin multiplier. The historical setting S = 4 corresponds to 8 pixels per line pair, or 4× the theoretical two-pixel reference. It is an engineering assumption you can vary. It is not a theorem that every camera needs this amount of oversampling.
MTF and the complete imaging chain
An MTF curve describes contrast transfer at spatial frequencies. Lens, aperture, focus, optical filtering, sensor layout, demosaicing and processing all affect the captured image. A pixel-count conversion cannot predict an MTF curve or assign one universal resolution to film. Use the actual sensor crop or film aperture for your project. The presets are editable examples; S35 and large-format labels do not define one fixed recording size.
Original MTF illustration
The Kodak VISION3 illustration from the previous page remains available for comparison. MTF shows how strongly fine detail is reproduced at a given spatial frequency.
View the MTF illustration
