Let’s assume film’s response to light is binary.
I suggest we should think about differences between the binary photo sites’ spatial distribution in film and digital imaging.
For film the gain particles are distributed in a pseudo-random fashion. In addition the grain particles are suspended in a three-dimensional matrix.
The distribution of conventional digital sensors are highly organized (discrete) in a two-dimensional plane.
Does it matter if film truly is binary at atomic and, or molecular levels? Differences in the chemical and electronic photo sites’ spatial distributions are more relevant.
___________
“…the mathematical rules of probability theory are not merely rules for calculating frequencies of random variables; they are also the unique consistent rules for conducting inference (i.e., plausible reasoning)”
E.T Jaynes, Probability Theory: The Logic of Science
I would say the specific spatial arrangement is somewhat trivial with the same number of binary sensing elements, as 3D emulsion grains project in the Z axis into a 2D image. The biggest difference is that the sensor would likely have simple rows and columns, which are more prone to aliasing with the same number of elements forming the same basic image. Of course, that is only tradition, and it could be possible to have quasi-random photo-site placement in an exotic moiré-proof sensor, if the pixel density would otherwise be prone to aliasing.
You could actually simulate a 3D grain-like effect with the "illusion" (what we mistake for reality) of a range of tones by using charged photo-sites ("ones") as the nuclei for a larger transparent shape of semi random size and shape added on top of each other, overlapping, if the image is going to be highly magnified and you don't want distracting gridded disks, or, you can just blur it at a wide radius of several pixels if it will be displayed at low magnification or is to be downsampled.
