is this Moiré pattern on some of my scans?

[Ed_Ingold]

Your definition of "optical resolution" does not conform to the accepted industry terminology. The fact that you disagree doesn't mean everyone else is "lying."

 

It has been my understanding from the get-go that Epson and other scanner manufacturers define "optical resolution" as the number of pixels captured per inch of film. That is true for Nikon, Hasselblad, Scitex, and all others I am aware of. If I copy a slide or take a photo using a Sony A7Rii, the "optical resolution" is that of the sensor, 5304 ppi.

 

In the last case, the end-to-end resolution will depend on the lens, and is expressed as "pixels per image height," or something to that effect, and refers to the lens, rather than the camera. What, then, is the resolution of the camera? Resolution of the lens can be measured independently with greater resolution than the lens itself can provide. The net result of lens and sensor is always less than either taken alone.

 

In the case of dual-purpose flatbed scanners, for reflected and transmitted light, the focal point is usually at the surface of the bed. Film holders keep the film a millimeter or two above that surface, which significantly degrades the effective resolution. Scanners which shift the focus depending on the mode will tell you so, and tend to be very expensive. I have the best results placing the film directly on the glass, emulsion side down (less problem with Newton's Rings), inside a thin mask made of black matting cardboard, held down by a piece of anti-Newton glass. The focal point is easily measured using an inclined rule, preferable an engraved engineering rule, at a known angle (q.v., "sine bar").

 

The greatest disappointment of a flatbed scanner is that its effective resolution is only about half of the stated optical resolution, whereas a more sophisticated device like a Nikon Coolscan is about 3600 vs the stated 4000. A really great lens on a Sony A7Rii might achieve a score of 4800 pixels/image height, far short of the theoretical 5304. Should Sony ask which lens you intend to use before printing the specifications ;)

 

It is incumbent on the consumer to read specifications with skepticism, an hopefully to understand how they were derived. Discussions such as this help clarify misunderstandings, as do many technical reviews. As always, Caveat Emptor.

[glen_h]

Your definition of "optical resolution" does not conform to the accepted industry terminology. The fact that you disagree doesn't mean everyone else is "lying."

 

 

(snip)

 

The greatest disappointment of a flatbed scanner is that its effective resolution is only about half of the stated optical resolution, whereas a more sophisticated device like a Nikon Coolscan is about 3600 vs the stated 4000. A really great lens on a Sony A7Rii might achieve a score of 4800 pixels/image height, far short of the theoretical 5304. Should Sony ask which lens you intend to use before printing the specifications ;)

 

It is incumbent on the consumer to read specifications with skepticism, an hopefully to understand how they were derived. Discussions such as this help clarify misunderstandings, as do many technical reviews. As always, Caveat Emptor.

 

For an interchangeable lens camera, there is no assumption as to which lens you might use,

or which lens someone might build in the future.

 

For a scanner (or camera) with a built-in lens, though, you know the lens. You can give a diffraction limited resolution value.

Focus is a different question, and usually not considered.

 

But yes, right or wrong, it is usual to measure by the number of pixels in the sensor.

 

Someone should not sell a camera or scanner with pixel resolution many times higher than the diffraction limit of its optics, but I suspect that they do.

 

In some cases, one might be able to do deconvolution on the output, though probably only with quality raw data.

[rodeo_joe1]

Whatever. The long-and-short of it is that Epson use pretty crap lenses in their flatbed scanners, such that the sensor resolution cannot be realised practically.

 

And mere sensor pixel numbers do not constitute "true optical resolution", since no optical system is involved. I don't know which mendacious organisation might have invented such an erroneous definition, but their understanding of the English language is obviously very defective.

Edited November 8, 2018 by rodeo_joe|1

[Ed_Ingold]

Welcome to the world of marketing, Rodeo. If you want that world to change, sue and get someone with a robe and gavel to agree. Otherwise content yourself to spread the truth to others (as you see it), It may not matter since film scanners of any merit have gone dinosaur. We are left with inexpensive (and largely ineffective) flatbed scanners, re-packaged digicams, or whatever we can find on the used market. Consider the real need in a world where the introduction of a new brand of film is headlined at PNet or DPReview.

[glen_h]

Anyone know what the diffraction limited resolution is for iPhone cameras?

 

It seems that phone cameras have more and more pixels, but are they really giving more resolution?

[rodeo_joe1]

'Diffraction limited' implies a perfect lens whose only resolution limit is its numerical aperture size.*

 

I doubt that even the sacred, can-do-no-wrong-in-the-eye-of-its-believers Apple can achieve that miracle.

 

A phone camera is, what, about 8mm square in total? With a sensor size of 5x7.5mm or thereabouts? With 12 (or 7) megapixels spread across it.

Do the maths.

 

And at a guess, I suspect the f-number of a phone camera lens is no greater than f/2 or f/2.8.

 

Anything else is all software enhancement, which if applied to a decent-sized camera could obviously do even better.

 

Why Canon, Nikon, et al don't get those 'trillion calculations per second' working for them is anyone's guess.

 

* Diffractive Airy disk diameter in microns = 2.44*lambda*f-number

Where lambda is the wavelength of light, also expressed in microns.

 

See this link for a fuller explanation.