Scanners - current status

[Loris Medici]

I'm glad that we're understanding each other finally. OTOH, as a matter of fact, 40 lp/mm also is poor performance (in terms of 35mm); that means you have an airy disk size of 12.5 mikron, which is (roughly) what you'll get (at max.) when the lens is stopped down to F16. Almost every 35mm lens will perform much better (than what they're capable at F16) in the aperture range of F5.6 to F11... (For instance, "theoretical" limit at F5.6 aperture is something like 117 lp/mm. Even after chopping 40% of that you'll still have a respectable 70 lp/mm, which translates to a sharp / full-frame print on 11x14" paper - in other words; 10x enlargement - even when viewed from 25cm...)

Regards,
Loris.

 

[Rudeofus]

OTOH, as a matter of fact, 40 lp/mm also is poor performance (in terms of 35mm); that means you have an airy disk size of 12.5 mikron, which is (roughly) what you'll get (at max.) when the lens is stopped down to F16.

You still focus entirely on theoretical lens performance, and yes, 100 lpmm is entirely doable with a good lens at F/8. That airy disk you mention is tiny compared to COC limits for focus (25-30m, depending on vendor) or blurriness from camera shake. Just to give you the idea: Canon specs its AF sensors to be accurate within DOF, and only the few "super accurate" ones are 1/3 DOF.

So it all depends on how you work with your camera. Landscape and studio shooters will get 100 lpmm, but e.g. street photography? No way.

 

[Worker 11811]

I did some more experiments with my SMPTE test film.

I remembered an important fact about movie film versus photo film. Movie film orients the emulsion the opposite way. The emulsion faces toward the projection lamp and away from the lens/screen. The reason for this is because "release prints," the copy of the film that gets sent to theaters for projection, are contact printed off an interneg. (Actually they are 3rd or 4th generation.)

I am so used to working with film in that orientation that I forgot to flip the film over before I scanned it. :eek:

Consequently, I flipped the film over and rescanned. I was able to get as high as 48 lines/mm. That's a little better. In a theater, I would call that an "okay" reading. (i.e. - The minimum acceptable.)

I'm a little more satisfied with that. If my math is right that comes out to just a hair over 1,200 dpi. I can live with that.

This exercise has got me thinking that movie film probably isn't the most optimal test target for this kind of work. It's probably good as a "basic" test but not optimal for best results.

So, where does one get a proper test target for scanner calibration? How much does one cost?

 

[Loris Medici]

25-30 mikron is coming from a very loose (and pretty old - from the times where lenses didn't have coatings, films had much less resolution and so on...) Zeiss standard for marking DoF scales on lenses. (Roughly = film's diagonal length in mm / 1730) It's nothing near to what you can achieve, even in the context of street shooting. (Please!? I DO street shooting with my RFs!) If that were all we were able to get, no one would be able to have sharp / detailed 8x10" (even!) prints from 35mm.

Do the math; 30 mikron would mean 5 lines per mm resolution (print) in case of a full 35mm film frame enlarged to 8x10" paper. I don't think anyone will agree that 5 lines per mm is sharp (at that print size) by any standard. (Even pt/pd prints on *plain hot press watercolor papers* can do better believe me; been there done that...)

Both practical experience (at least mine) and logic/math doesn't support your assumption...

Regards,
Loris.

You still focus entirely on theoretical lens performance, and yes, 100 lpmm is entirely doable with a good lens at F/8. That airy disk you mention is tiny compared to COC limits for focus (25-30m, depending on vendor) or blurriness from camera shake. Just to give you the idea: Canon specs its AF sensors to be accurate within DOF, and only the few "super accurate" ones are 1/3 DOF.

So it all depends on how you work with your camera. Landscape and studio shooters will get 100 lpmm, but e.g. street photography? No way.

 

[Loris Medici]

BTW, reading your answer again, 100lp/mm @ F8 isn't doable even in theory. See the formula below:

Airy disk size (mm) = 1.22 x Wavelength (mm) x F stop

1.22 x 0.000625 (625nm; full spectrum minus red) x 8 = 0.0061

1 / 0.0061 ~= 164 l/mm = 82 lp/mm... (You need 22% more resolution for achieving 100 lp/mm.)

100 lp/mm isn't theoretically possible at F stops above F5.6.

Not 100% relevant but see this also:
Diffaction Limited Aperture Calculation (Google Docs Spreadsheet by Loris Medici)

Regards,
Loris.

 

[Rudeofus]

It's nothing near to what you can achieve, even in the context of street shooting. (Please!? I DO street shooting with my RFs!) If that were all we were able to get, no one would be able to have sharp / detailed 8x10" (even!) prints from 35mm.

I did the math but not in terms of "but it looks sharp at 8x10", instead I consulted a popular DOF calculator.

Here are the results based on the following assumptions:

• assume perfect lens, absolutely no field curvature, spherical abberation or other crap
• You are like Chuck Norris and have no camera shake. None.
• focal length 24mm
• lens stopped down to F/5.6
• distance to subject 150cm, so you cover a width of 225cm on 135 film

If you want 40 line pairs per mm, which equals a max airy disk of 12.5m, you end up with a DOF of about 55cm. For a COC below 6m (-> 80 lp/mm) your whole subject must lie within a distance range of 26cm! Achieving this in a candid situation is challenging, euphemistically speaking.

Next thing is motion blur. People walk with about 1m/s, which is one whole image frame per 2.25 s. If you shoot with an exposure time of 1/125s, that person moves across about 3/1000 of your image frame. Given a film width of 36 mm, you have motion blur of 128m! If you move your camera with that moving person, assuming you do this perfectly while accurately maintaining perfect focus, at least other people in the picture may not move in the same direction so they will be inevitably blurred.

Note that none of this is relevant for carefully composed shots with near static subjects.

 

[Loris Medici]

Thanks for the trouble. 56cm DoF is perfectly manageable; it's a depth that may enclose two people... OTOH, *your example doesn't rule out antyhing else on the frame that may be in the plane of focus*, no? (BTW, I have tack sharp images from my Bessa + 21/4 combo, shot handheld at 1/8 - or as a practical example, I can go to bed and sleep like a baby even after drinking two double cups of espresso, therefore, I don't believe I have any agitation / manual dexterity problems...)

Anyway, I still insist on the fact that 40 lp/mm is "technically" poor performance for 35mm format, and that's all I say since from the start; 35mm photography doesn't solely consist of chasing children or (almost "lomography style") street shooting (and street photography doesn't necessarily imply moving photog and/or moving subject), please refrain from making up conditions/situations favouring your claim...

Regards,
Loris.

 

[sanking]

One thing that one needs to consider is that a print does not have to be tack sharp all over to give the impression of sharpness. Often if the important subject is one that has a lot of detail and we focus on it at a relatively wide aperture and use a high shutter speed the print will have the appearance of great sharpness, even though much of it is slightly out of exact focus.

I also agree with Loris that 40 lp/mm is fairly mediocre performance for 35mm. I have done a lot of work hand held with the Mamiya 7 handheld (landscape shots that allowed optimum depth of field and a relatively high shutter speed) that about twice that much effective resolution. With a high resolution ASA film of 100 like Tmax-100 or Fuji Acros this is not hard to achieve.

Sandy King

Thanks for the trouble. 56cm DoF is perfectly manageable; it's a depth that may enclose two people... OTOH, *your example doesn't rule out antyhing else on the frame that may be in the plane of focus*, no? (BTW, I have tack sharp images from my Bessa + 21/4 combo, shot handheld at 1/8 - or as a practical example, I can go to bed and sleep like a baby even after drinking two double cups of espresso, therefore, I don't believe I have any agitation / manual dexterity problems...)

Anyway, I still insist on the fact that 40 lp/mm is "technically" poor performance for 35mm format, and that's all I say since from the start; 35mm photography doesn't solely consist of chasing children or (almost "lomography style") street shooting (and street photography doesn't necessarily imply moving photog and/or moving subject), please refrain to make up conditions/situations favouring your claim...

Regards,
Loris.

 

[Rudeofus]

56cm DoF is perfectly manageable; it's a depth that may enclose two people... OTOH, *your example doesn't rule out antyhing else on the frame that may be in the plane of focus*, no?

But these 56cm DOF apply only if you accept airy disks of 12.5m, giving you about 40lp/mm, something you continue to call mediocre or poor. And I don't think it's acceptable to have some random part of an image in perfect focus, and I'm sure you, too, try to control which exact part is in focus.

35mm photography doesn't solely consist of chasing children or (almost "lomography style") street shooting (and street photography doesn't necessarily imply moving photog and/or moving subject), please refrain from making up conditions/situations favouring your claim...

Well, if you call urban landscape and architecture shots "street shots", then of course you need no people in the frame. You may call it "street shooting" if you take pictures of standing people with your camera on a tripod, and obviously you can get more than 40 lp/mm with this shooting style. I would, however, not call this typical for street shooting as I know it.

My take at this whole thing is this: if I care about resolution, I use an MF camera, which can give me not only lp/mm, but also lots of mm to begin with. The main strength of 35mm format cameras are their unobstrusiveness, versatility and speed, which I try to take full advantage of, even at the cost of resolution.

 

[Loris Medici]

You're constantly twisting my words Rudeofus and taking the discussion away from the main subject. I'm sure you'll be able to write at least 50 street shooting scenarios not including tripods and/or slow shutter speeds and/or high angular speeds, therefore I won't bother... Why you insist on pushing the examples into the extremes? Such as: shooting a 2 y/o child (one extreme) or standing people/architecture on a tripod (the other extreme)...

I think I've made my point, you take it or not, that's your problem; I have other concerns than educating people. Good luck to you; continue to think 40 lp/mm is good for your 35mm shots and live in happiness... (Ignorance is bliss? Never mind...)

BTW, see one of my street photos that would easily give you resolution over 40 lp/mm, if there was a resolution target present in the scene:

Dead Link Removed
(Bessa T + 21/4)

I have plenty of these of you like to see more...

Bye bye,
Loris.

 

[Rudeofus]

Thank you, Loris, for that example! Just look at the depth of the subject matter and use some basic DOF calculator! You want 80 lp/mm with your 20/4? Yeah, stop it down to F/5.6 FWIW. What's the distance of the guy holding the scarf? 2m? Less? See, what you get?

You get your 80 lp/mm in a range from 1.7m to 2.4m. If you accept 40 lp/mm, you get 1.5m - 3m, which still doesn't cover most of the scene. The folks in the back ground? Blurred! The mosc? Drowned in a COC of 35m. Sorry, dude, if you really thought that even most of this image was tack sharp (according to your definition), then you have very little reason to throw insults at me.

And if you still think, ones 2 year old girl is not a valid reason for photography, then well, keep taking tack sharp pictures of bazaars, then

 

[Loris Medici]

Whatever...