As far as I've understood, optimal diameter for pinhole size is calculated by formula
d - the optimal diameter for pinhole
c - constant
f - focal length (distance between pinhole and film/sensor)
λ - wavelength of light the pinhole should be optimized for
f and λ should be in same units as desired for d
Different sources seem to agree that about 550nm (green-yellow) is a good value for λ, and the focal length part is also quite clear.
However, each source seems to provide different value for the magic constant c —
- Wikipedia article cites (~1.414)
- Stanford Pinhole Math suggests lookup values based on c value of 1.542…1.543
- Stanford Complex Pinhole Calculator uses 1.562
- mrpinhole.com Pinhole Size Calculator results work out at ~1.8
- David Balihar provides 1.9 as "lord Rayleigh value"
The difference of 34% between smallest and largest suggested value seems quite significant.
Why are there so many different values for the constant? Are different constant values optimizing different properties of resulting image? Or perhaps different constants apply to different pinhole material thicknesses (if that's the case, do bigger constants go for thicker materials)?