Measuring film resolution

[RobC]

but what use are the scans without the ability to output them onto a target at the correct size and definition

 

[Photo Engineer]

I am using originals Rob from film and metal on glass. However, if I do scan, I can do it and print to transparent material with a hires digital printer! So, I have no problem. And, when I was at EK, I had many many more options to print, photograph, process and measure, so this is not a new subject to me.

PE

 

[RobC]

Yes but what is Rudeofus supposed to do with a scan and without the necessary equipment.

 

[Photo Engineer]

If I could scan and upload at a high enough resolution and in a format with minimum aliasing, then he could probably use it.

PE

 

[Rudeofus]

Rob&Ron, I think you mixed up something here: Ron has suitable targets, contact printed them onto various photographic papers, scanned the results, and is now unable to post scans to APUG because of file size limitations. Rob has resolution targets in nearly arbitrary resolution, but only as PDF, and is unable to print them onto a transparent medium in the resolution he wants. Both problems are only barely related (insofar as they popped up in this thread), and both can be solved easily: Ron could crop his scans, thereby limit the image area to the relevant parts and post them here as high res TIFF or PNG images. Rob could go to a wafer mask making company and have arbitrary patterns with submicron resolution printed onto perfectly flat glass, for insane amounts of money

 

[RobC]

And there was me thinking it was you wanting these things. Clearly this isn't the case. I will leave you to your own devices...

p.s. the usaf 1951 chart is not arbitrary resolution if you have read up on it and print it at the correct scale. But you would need to know what you're doing to achieve that.

 

[Photo Engineer]

USAF retiree in the field of photography!

Kodak Research retiree in the same field!

PE

 

[Rudeofus]

Never been in the field of photography myself, not retired either, but prepared the STL files for my own special purpose wafer masks myself (with correct scale! ), and wrote the software for doing this myself, too. Still unwilling to pay for a wafer mask, though

I guess it's time to focus on how to practically make that repeated resolution pattern as suggested by Bill Burk on Fuji Velvia 50.

 

[Bill Burk]

The thought occurs to me that the pattern could be smaller than 3/16, and repeated several times in a 3/16 square... It could be positive or negative, because it seems to me you get equal black and white bars either way. If the pattern is negative, I will just make two test exposures: one test exposure with the Stouffer scale + pattern: for resolution, and one with Stouffer scale alone: for characteristics.

As for making the pattern... well it's not too difficult a design job on paper (for in-camera reduction).

 

[Photo Engineer]

Well, after all of this, I must say that if you have a lousy camera and/or lens or you have a tremor (and no tripod ) then you really don't know the overall resolution of your "film".

PE

 

[Bill Burk]

PE,

I'm talking of using a good camera with micro- or graphic arts film to make a best-effort home-made master resolution filmstrip...

It might reach 100+ lp/mm if I am lucky

Then using that filmstrip in the bed of the sensitometer for contact test exposures on everyday film.

It might evaluate up to 80 lp/mm if I am lucky - but it will be the resolution of the film, not the optical system of the camera at that point.

It's only for my home laboratory, self-educational purposes... the information might help me understand the shape of the exposure/resolution curve for my everyday film and developer combinations.

 

[Photo Engineer]

In the long run, we end up arguing about who's system is better, sharper, more accurate or whatever.

PE

 

[Alan Johnson]

Erwin Puts seems to have had such a debate with Zeiss, in the end he concluded that 150-175 lppm is the maximum attainable on film but in tests he did not approach this:
http://web.archive.org/web/20051017022108/http://www.imx.nl/photosite/technical/bwstateofart.html

 

[Nodda Duma]

Re: issues uploading large files... Look into using Dropbox for big file transfer.

 

[Athiril]

Erwin Puts seems to have had such a debate with Zeiss, in the end he concluded that 150-175 lppm is the maximum attainable on film but in tests he did not approach this:
http://web.archive.org/web/20051017022108/http://www.imx.nl/photosite/technical/bwstateofart.html

See Henning's tests.

 

[Rudeofus]

100 lp/mm and 100 lp/mm can be very different things. Let's say you start with 1:1000 or 1:50 contrast and look at the pattern on the output medium, where do you declare the limit of resolution? Is it where contrast drops 10/50/90 % below maximum contrast? Is it where your eyes can't reliably detect the pattern any more? Which limit is more relevant, and doesn't it depend on your subject matter? Either way, you will get very different resolution numbers for both cases, both with their own justification.

A real MTF graph would completely eliminate this ambiguity.

 

[Athiril]

It's where you can no longer see the lines generally, the limit of the resolving power in a photograph, since photographs are for looking at.

 

[Alan Johnson]

Yes, the visible lines.
I have an older chart which, photographed from the correct distance, gives a direct reading in lppm.I read the attached as 80-90 lppm.

 

[Athiril]

I photograph from various distances, I just measure the size of the chart on the negative

 

[Bill Burk]

Can someone check my work:

In an 11x14 print from 35mm TMY-2, I can clearly see and count at least a half dozen hairs in a millimeter. (At one point I counted 10 hairs but I think I cheated by counting thicker "hairs" as two because I could tell there were two hairs there. So take 6 as really resolved.) Now the enlargement is 9.3x (35mm frames are 24x36mm and the 36mm dimension of the frame is 335mm on the print) so that equates to about 56 line pairs per millimeter achieved on the negative. (6 times 9.3 = 55.8).

I find this amount of detail on an 11x14 print to be quite pleasing. This is from something my daughter did with her first rolls of film, her first prints.

If this calculation is correct, I think I'm going to start with 56 lp/mm as a baseline for acceptable quality in 35mm work. I will expect more resolution from TMAX 100. But for now I have a baseline of what I like.

 

[Rudeofus]

Yes, the visible lines.

The problem with this measurement is that it's impossible to combine the measurements of two consecutive transfers. Let's assume one evaluates a lens at 200 lp/mm and someone else evaluates some photographic film as 200 lp/mm. If you take an image though that 200 lp/mm lens onto that 200 lp/mm film, you get nowhere near 200 lp/mm if you define lp/mm as "I can barely see the lines". If you define the lp/mm limit as "I get 50% of max contrast", you won't get 200 lp/mm either, but you'll be much closer.

 

[Alan Johnson]

Use the formula to calculate the resolution R in lppm that you will see from the resolution of the lens and film separately:
1/R^2=1/Rfilm^2 +1/Rlens^2
Its on Erwin Puts site I linked to previously,he spent a long time doing tests similar to what you propose.

 

[Rudeofus]

Alan, this formula works for "I get 50% of max contrast", but not for "I can barely see the lines" limit. Given that Erwin's numbers are far below the numbers given by Zeiss, I have reason to believe that he used the "I get 50% of max contrast" limit.

 

[Nodda Duma]

FYI, lower limit for contrast detectable by people is established. This curve is typically laid over the system MTF to predict useable resolution for human viewing. I'll try to dig up a reference.

Most spatial frequencies it's ~5-8% up to 1 arc minute (which you can convert to lp/mm by assuming a viewing distance).

 

[Bill Burk]

I was pleasantly surprised to see Erwin Puts mention 6-10 lp/mm is about the resolution limit for the unaided human eye... That agrees closely with what I found.