4 ms·

It’s more than a bit difficult to talk about films in terms of pixel resolution, but for example I used to shoot a lot of FP4, a fairly but not exceptionally fi

It’s more than a bit difficult to talk about films in terms of pixel resolution, but for example I used to shoot a lot of FP4, a fairly but not exceptionally fine grain black and white film, and in 35mm (actual film gate 24mm x 36mm) I’ve seen that stock estimated at around 25 Megapixels equivalent, which works out to 0.0289 Mp/mm^2, or about 1,492 Megapixels for an 8” tall x 10” wide negative… assuming square pixels that’d get you around 43K horizontal resolution. That’s if my very tired brain is getting the math right. The perceived resolution would basically melt the mind… if you’ve ever looked at a 35mm frame blown up to 8x10 side by side with an 8x10 contact print under a loupe, it’s just incomparable.

rbanffy 3y ago
I was thinking more in terms of dpi than total bitmap size (but, then, specifying the frame physical size was silly). One possible way to measure it would be to do a contact print with a reference etched with patterns of a higher resolution than the one we want to measure. The film’s resolution would then be the densest pattern that’s still resolvable. Either that, or directly exposing the film to that pattern (which would require some very fancy semiconductor-manufacturing-grade optics).
- yawpitch 3y ago
  It’s really hard to come up with any kind of a solid digital-resolution-relatable number for any given film stock, if only because multiple layers of silver halide grains randomly distributed in an emulsion across three dimensions has no direct relationship to a two dimensional matrix of regularly distributed pixel elements, but the MTF curve of the stock describes its theoretical limit of resolution in terms of line pairs per mm. There’s published MTF curves for every stock, but the limiting factor is almost always the optical system, not the stock. An 8x10” negative shot with a very slow, black and white stock (say PanF for telescopic astrophotography) would exceed the practical DPI capacity of any extant commercial drum scanner… at a somewhat informed guess FP4 would probably top out somewhere north of 6400 DPI, as I’ve seen those scans not pick up details visible through a really good high power loupe.

rbanffy 3y ago

I was thinking more in terms of dpi than total bitmap size (but, then, specifying the frame physical size was silly). One possible way to measure it would be to do a contact print with a reference etched with patterns of a higher resolution than the one we want to measure. The film’s resolution would then be the densest pattern that’s still resolvable. Either that, or directly exposing the film to that pattern (which would require some very fancy semiconductor-manufacturing-grade optics).

yawpitch 3y ago
It’s really hard to come up with any kind of a solid digital-resolution-relatable number for any given film stock, if only because multiple layers of silver halide grains randomly distributed in an emulsion across three dimensions has no direct relationship to a two dimensional matrix of regularly distributed pixel elements, but the MTF curve of the stock describes its theoretical limit of resolution in terms of line pairs per mm. There’s published MTF curves for every stock, but the limiting factor is almost always the optical system, not the stock. An 8x10” negative shot with a very slow, black and white stock (say PanF for telescopic astrophotography) would exceed the practical DPI capacity of any extant commercial drum scanner… at a somewhat informed guess FP4 would probably top out somewhere north of 6400 DPI, as I’ve seen those scans not pick up details visible through a really good high power loupe.