Real resolution of films?

[Glenn Kroeger]

I have been testing some new LF lenses. Based on the results on Chris Perez's we b site, I decided to use TMX 100 for the tests. At the recommendation of local " experts" I used Tmax developer, at 1:4 dilution, 72 degrees for 7 min, and rated the film at EI 100.

I am using the standard USAF lens charts at a 50:1 distance to emulate infinity.

What I am finding is that 1) (the good news) is my 55mm Apo-Grandagon clearly re solves 70 lp/mm, and 2) (the subject of this question) beyond about 70 lp/mm it appears that grain size really plays havoc with resolution.

On paper, these films are supposed to be able to resolve > 100 lp/mm at high con trast, and although my charts aren't 1000:1 contrast, I cannot see how, with the size grain I observe, this film could ever resolve 100+ lp/mm.

Am I too stringent on my definition of resolution? I expect to see ightlines sep arated along their entire length by a dark line.

Is my developer/time/temperature giving me too much grain?

I am using a 40x Zeiss stereomicroscope to view the negs, so that's not the weak link.

Thanks for any feedback.

[M.]

Try Rodinal 1+50 for finest sharpness with TMax 100.

[Bill C]

Glenn, I don't know if it's current, but there used to be an ANSI standard that described how to interpret results using the 1951 USAF test target. The key thing was that you needed a "reasonable" certainty that 3 bars were resolved. In a nutshell, that level of certainty basically asked whether you were more than 50% sure that three bars were there. Since you said, "I expect to see lightlines separated along their entire length by a dark line.", I'd say yes, you are being too stringent.

I would suggest that when you examine each set of bars, ask yourself if you would be willing to bet a beer as to whether three bars were actually present on the original target (as if that target were unknown to you). If you are willing to bet, then you should consider it to be resolved.

[Pete Andrews]

Hi Glenn. The manufacturers figures for film resolution are the MTF of the film itself, not the OTF of a camera/lens/film combination.I don't know exactly how they're arrived at, maybe by contact printing, but in practise you can never acheive those figures. There's a formula for combining the MTFs of the lens and film to give a final OTF, but I don't know it offhand. Something like the RMS value of all the system MTFs sounds about right.Anyway, the good news from this is that your lens resolution is actually higher than you can measure from the film. If you're getting a combined resolution of 70 lppm, then it's likely that both the film and the lens have separate resolutions closer to 100 lppm.Whatever, for LF work 70 lppm is more than enough.

[Pete Andrews]

PS. On a more practical note, I've found that I have to rate T-max at 80 ISO and cut the development to get the finest grain. I use Ilfosol-s, for its tonal quality with Tmax 100, even though I know the sharpness is compromised.

[Daniel Taylor]

TMX in Xtol yields reduced grain.

[Duane]

Just some thoughts:

As it turns out, what we call grain on the image isn't really grain. It's a grouping of a number silver halide cystals. As a result, the resolution can in fact be better (read smaller) than the grain size. If the cyrstals were actually the size of the observed grain kernels, the whole kernel would be dark.

Rodinal is a high acutance developer but gives grain like almost nothing else. Ilfosol reduces grain but the negatives get a softer feel. I suspect Tmax developer is in between.

Regarding the measurement technique, there's quite an interplay of factors here. Did you use a white, a tungsten or a monchromatic light source? Did you use a diffraction limited lens? What was the developer agitation method? What instrument was used to measure the contrast ratio of the film?