I agree with tomfrh. If one examines the actual strip of film to determine the amount of particles that responded to the light, then this is the most accurate measurement. It might be tedious, but if/when one determines the number of particles per millimeter of film, this should be an accurate determination of film resolution. Particles per millimeter can be converted to line pairs per millimeter (a common term for determining and comparing resolution), which can then, through a simple calculation, determine a given piece of film's MP content.
For example, if a very fine grained film is examined -- one that resolves 100 lppmm, for example -- this corresponds to (24mm x 100lppmm x 2) x (36mm x 100lppmm x 2) = 34.56 mp.
The x2 is necessary to convert from line pairs per mm to lines per mm, or pixels (or dots) per mm.
If one foregoes the conversion to lppmm in the above example, then this means that the pixel (or particle) count must be 200 particles per millimeter. Thus (24mm x 200ppm) x (36mm x 200ppm) = 34.56mp
But let's say one counts "only" 100 particles per millimeter on a film strip, then we're looking at (24mm x 100ppm) x (36mm x 100ppm) = 8.64 mp
So, determining a type of film's resolution is independent of any sort of lens or interpretation. It's just observation, measurement, and math. Obviously a limiting factor in terms of an image's results is the resolution of the lens being used to produce the image. 100lppmm is a very high number for any given lens. Typically the best I've ever seen with lens resolution tests is around 70lppmm. So let's use that number instead of the film's theoretical limit:
(70lppmm x 24mm x 2) x (70lppmm x 36mm x 2) = 16.9344 mp. Which is, if you ask me, a real world number, one that accurately reflects a fine-grained film's working resolution.
Fine from your examples. But before some get confused now we should make clear first : to make a determination of a thin black line on white ground you just have to count lp/mm and from your example that means in digital you definitive need the min. of 2 lines of pixel. (one pixel for black/one pixel for white).
So 100lp resolution definitivly is describing 2400 x 3600 x2 = 4800 x 7200 = 34,56MP resolution.
Thats not the fact. It is the contrast of 1:1000 ! So it is just theoretical. Kodak gave with introduction of new Ektachrome films >200lp resolution.
That would give theoreticaly 9600x14400 = 138,24 Mp with 35mm.
And it is objective reachable in Kodak research laborathorys. But you can't reach it with a camera (because a lens can't reach 200lp and you can't have a contrast of 1:1000 in normal pictures.
(in film tests of manufakturers they use a
grid (the lines of resolution tests) with is projected under laboratory condition with the help of special lenses simular to microscop optics of best characteristics)
So it is just a theoretical test standard to
compare different films.
The film in color c41 and E6 and [very special] in bw isn't the limiting factor within your "system of resolution" - it is ALLWAYS the lens wich may limit the theoreticaly power of modern emulsions.
In digital it is (meanwhile) the same - the
digital sensor isn't the first limitation - it is your lens.
BUT IF YOU MIX A FILM WORKFLOW VIA SCANNING THE SCANNER IS THE FIRST LIMITATION. TO COMPARE DIGITAL AND FILM ONTO A SCREEN IS THE NEXT FAILURE WICH LET MANY SAY = 6mp digital looks better than ISO 100 films.
Imagine you would expose data from 60Mp digital onto film and then you look at resolution.
Or imagine you would print datas of different digital cameras for comparison onto normal paper via cheap ink jet machines.
What is the reference you would get.?
It would then nothing say about your camera comparison but it would tell you somethink about the characteristics of you lousy ink jet workflow.
Think about
with regards
PS : Very special is the system resolution between film and lens : IF your lens would reach 120 lp/mm and your film is
(theoretical from 150lp/mm [within lab conditions].. ) your system resolution isn't in the middle of each factor (here this would be 135lp/mm) it is ALLWAYS below the aritmethic mean.

But seriously in 35mm high resolution bw films with special characteristics (700lp/mm WITHIN LABS) in combination with Leica M lenses of modern type gave an extrem high resolution at F 5,6/F8 with tripod.
I've seen such results and I like to compare them with digital cameras of big sensor type.(like Leica S2 system - digital midt format).
But it make no sense to have a scan of this films for comparision.
It make no sense to compare such films with prints of normal optical enlargement (with workflows of normal darkrooms)
You definitive have a need of high tech darkrooms with best "rodenstock"enlarger lenses (the > $3000,- class) at F5,6/F8 AND best experienced darkroom operators.
Then you will get the todays optimum with simple 35mm.
In 120 midt format with such film - film is superior in comparision to 120Mp digital - beliefe me. The same is with 4x5 and higher formats.
But you definitive need a true basis to compare. It make no sense to scan 35mm / 120 films / 4x5 inch with a good epson flat bed scanning maschine.
In such test the result is allways the same : 35mm digital is better than 4x5 bw (delta 100) but imagine : this All says nothing about 4x5 delta100 but everythink about your epson flat bed workflow.
