Response to A Super film from Germany Gibabit film

I wasn't contesting Gigabit's claims for its film, just questioning its sensationalist publicity material.
Let's put some facts on the table:
1) The resolution of a film cannot exceed the limitations set by laws of physics. Primarily, light cannot be focused to a spot size smaller than its wavelength. For blue light this is 450nm, a resolution limit of ~1100 lppm, for red light its 700nm, or 700 lppm. To actually test a film to these limits of resolution isn't a trivial task.
2) The resolution of a lens is diffraction limited to a figure of 1392/N using light of wavelength 698nm (N = f stop number). Most lenses approaching this level of performance work at an aperture of f/5.6 or so, and cover a very small field. This means that the practical limit of resolution is 248 lppm.
3)The OTF of the combined film and lens reduces this figure even further. Optimistically, a figure of 200 lppm might just be obtainable.
4)All the above assumes perfect focal plane alignment and flatness. An error of one wavelength of light results in the limit of resolution of the lens being reduced to 20% of its maximum. i.e. 248 lppm drops to 50 lppm. The surface roughness of a film based gelatine emulsion is probably on the order of + or - 100nm.

I used to work in micro-photolithography, producing masters for semiconductor devices, so I've seen these theoretical limits in action.
Achieving 5 micron features using visible light is not an easy task, and that's using a Lippmann emulsion on glass plate, exposed with monochromatic light, through a very expensive specialist lens, set on an accurately aligned optical bed, in a temperature and vibration controlled environment, and developed to maximum gamma.
Perhaps now you can see why a continuous tone snapshot of two girls on a museum roof, showing the same kind of resolution stretches my credulity a little.

[ Previous | Next ]