Thanks Scott.
This has turned into a bit of a dive for me, but I've always been a bit curious about film. Not so much to get into it, but just wondering "how good is/was it" type of thing.
I am still looking into the resolving power of film, as it does not seem simple. But, I can totally see what you describe in the "A". There may be two aspects of film coming into play: 1) Potentially impressive resolving power; and 2) Film manufacturers may be "cheaters." Of course, the third option is I did underexpose the MF relative to the R5.
But, assuming it was a legitimate difference between film and digial, a quick explanation.
Film may have more impressive resolving power than I appreciated. Keeping this on the simple side, Fujifilm reports the resolving power of their film and several people rave about Velvia 50 for its rendering and detail. Looking at Fujifilm's site, they list the resolving power of Velvia 50 at 160 line pairs per mm (lp/mm).
Converting this, a line pair is two lines, that means 320 lines per mm. Assuming a line is a single row of pixels, that gives us the digital equivalent of 320 pixels per mm. Then some quick math: 36 mm x 320 = 11,520; and 24 mm x 320 = 7,680 or, 11,520 x 7,680 = 88.5 MP equivalent to digital. My 645 MF film camera would be ~238 MP equivalent.
So, if we just take Fujifilm at their word, digital still has not caught up to film in terms of resolving power. Fujifilm backs this up with MTF charts. Velvia ISO 50 is on their page 48, PDF page 24.
This is where you can start to see the "cheating." Notice the MTF chart was that it peak significantly above 100% response. In most instances, that isn't possible. This guy has looked at this and has noticed with Velvia 50 (not all film) had whiter whites and blacker blacks than expected right at the boundary of the white and the black. Thus, the differential is greater than 100% of the contrast ratio of the chart.
Did something like this come into play with Kodak Portra 400 used to shoot the "A"? Not sure, but on the last page you can see Portra 400 has a similar MTF curve (last page in link). Also, if you are wondering, I think FujiFilm extrapolates their curve out to 7% response to get the 160 LP/mm value, but their curve does end before you reach 160 LP/mm and 7% line is bolded. In reading about how MTFs work, it seems like 2-5% response is still visible contrast.
I am still working on if converting LP/mm at 1000:1 contrast ratios translates as linearly as I did just above and have seen a few others do online. With film, they do indicate that resolving power drops with lower contrast. Using the Velvia 50 as a reference, they rate it at 160 LP/mm at 1000:1 and 80 LP/mm at 1.6:1 contrast ratio. Also, I came across this test where the 24 MP Sony A900 looks a lot better than various films. So, true resolution is not settled in my mind....
If people are interested, and also for my future reference: resolution units, Understanding MTF curves, and How to Read MTF curves.





Reply With Quote