Resolution Limits of 35mm Photography

[Snapshot]

Hi All,

I have a general question/comment regarding the practical resolution limit of 35mm pictures. It is my impression that it is usually the lens that is the limiting factor for maximum sharpness in 35mm image capture. A good lens can resolve about 80 lp/mm or more while many other standard lenses struggle to attain 50 lp/mm. Other individuals have commented that the film is typically the limiting factor and I would agree that some films are not capable of resolving much above 40 lp/mm but many can resolve over 100 lp/mm. Furthermore, other argue that the aperture ratio limits what can be resolved, with apertures of f/22 limiting resolution to about 70 lp/mm due to diffraction.

I must admit I'm quite skeptical of individuals claiming their DSLR superior to film because their 10MP camera is supposedly capable of resolving 200 lp/mm. Doubtful, especially when their lens can barely resolve more than 50 lp/mm. So, I guess my question is what in your opinion is the prime limiter of image resolution for 35mm? My perspective is that the lens limits 35mm photographic resolution. You can adjust your aperture and you can use film with good resolving capabilities but you are basically stuck with what your lens can resolve.

 

[reub2000]

How many times do you shoot at f/22? I sure don't do it often.

Also, f/22 means that focal length divided by 22. It would be more useful to measure the actual size of diaphragm opening in this case.

 

[Pinholemaster]

I'm resolved to whatever resolution I can get regardless of the medium.

But one thing I can say, I rarely even shoot at f/22 with 35mm. A lens is far sharper in the middle of the f/stop scale.

 

[nicolai]

If you're that concerned about it, I'd worry about a solid tripod and a remote release with MLU (if appropriate) before I worried about the lens. Obviously a soft lens is a soft lens, but hand-holding a squillion-line lens at 1/30 on an SLR will still be soft.

If you really, really cared about resolution, you wouldn't be shooting 135 in the first place, you'd be shooting wet plates on a monster tripod with crazily expensive lenses. If sharpness and resolution were all that mattered in photography, we could all go out and buy the sharpest lens available, put it on a vibration-dampened tripod, shoot one wet plate of a resolution chart, and call it a day, secure in the knowledge that we'd made the best photograph ever. Since they're not, why not just go with whatever gives you the results you're happy with?

If you're not happy with the resolution you're getting with 135, go to medium format, change the way you shoot, or don't enlarge as much.

There is no universal "better". Technical mumbo-jumbo aside, maybe DSLRs *are* better for some people to accomplish what they want to accomplish. I'd let them do their thing and worry about what *you* want to accomplish.

 

[Michel Hardy-Vallée]

So, I guess my question is what in your opinion is the prime limiter of image resolution for 35mm?

The sharpness of the photographer's mind.

 

[kunihiko]

>>what in your opinion is the prime limiter of image resolution for 35mm?

Camera shake.

 

[Nick Zentena]

Good 35mm lenses can out resolve anything you might want. But you won't be shooting at F/22. F/4,F/5.6 or maybe F/8.

Can you get this all to the film? No.

Biggest problem with 35mm is the enlargement factor. You aren't going to record more lp/mm with a bigger camera. But you'll enlarge less. Inches matter just like in racing.

You could try and turn a 35mm camera into a tripod camera. That will help you get all you can out of the lenses etc. But why? If you're tied to a tripod it's not much harder to mount a slightly bigger 645 or a much bigger 6x7 camera to the tripod. Then inches matter again.

Use 35mm for what it's good at. Speed.

 

[Snapshot]

The sharpness of the photographer's mind.

Then I guess its fuzzy pictures for me.

 

[Ray Heath]

why even be concerned, does any of this matter if you are producing work that you value

 

[copake_ham]

why even be concerned, does any of this matter if you are producing work that you value

I concur.

Seems to me you're seeking a justification for using film that isn't required.

Whether film can out res digi at 35mm (m/l since not all DSLRs are true 35mm) is less important than what you like to do.

In the pro fields - the "workflow" of a DSLR is superior enough to overcome any disadvantages against a similar film image. Look at your daily newspaper - none of the photos "res" high enough to matter. If I were a pro PJ - I'd have to shoot digi - it gets the image into the paper faster - and that's all that matters.

Shoot what you enjoy, whatever it is....

 

[DrPablo]

I must admit I'm quite skeptical of individuals claiming their DSLR superior to film because their 10MP camera is supposedly capable of resolving 200 lp/mm. Doubtful, especially when their lens can barely resolve more than 50 lp/mm.

To be fair many 35mm lenses can resolve well over 50 lpm.

Just to answer the issue re: DSLRs, it's easy to calculate the theoretical maximum resolution of a digital sensor (in terms of lpm). I mean a grid-shaped sensor can resolve nothing more than one detail per photosite, so a line pair would require two photosites. So the Nikon D2x, with the highest of all pixel densities (12.8 MP on APS-C) can theoretically resolve no more than 89 lpm, because it only has 89 pixel-pairs per mm. This, of course, will be limited by the antialiasing filter, low contrast subjects, Bayer algorhithms, etc.

The thing is, people (including former film users) who proclaim superiority over film often mention non-optical considerations, such as convenience, versatility, and cost-benefit factors. These can't be argued, I mean they are what they are. They also mention, fairly commonly, application-specific circumstances where digital may not equal film but it's good enough. That's pretty fair, I mean it doesn't take many megapixels to fill a wedding album with small prints. But a lot of this debate is idealism as well, and selective ignorance of the factors that make a "film versus digital" debate fairly trite and irrelevant.

Anyway, if you take a 35mm film camera and use a fantastic lens, you may or may not have greater resolving power depending on the film you use. Keep in mind that with digital sensors that resolution stays constant at all ISO speeds, whereas with film a higher ISO film is basically a different sensor. If you use a phenomenal lens, like a Leica or Contax-Zeiss lens (which claim 200 lpm resolutions sometimes) and a film like Tech-Pan, and avoid diffraction limits and motion or mirror-blur, then you can have extraordinary resolving power in 35mm format.

Now, small format shooting is quickly limited by a few factors, including its diffraction limit (which often happens at around f/11 or f/16) and its circle of confusion (in other words, the requirement for extremely fine resolution in order to enlarge a smaller film image). This is true whether film or digital, but these factors certainly favor 24x36 over 15x24. And the diffraction limit becomes a nearly insignificant issue by the time you get to large format, because your enlargement factor is so small that subtle losses in resolution are trivial.

There was a great series on print sharpness in the Sept/Oct and Nov/Dec issues of View Camera. What is clear to me from these articles is that the degree to which diffraction limits sharpness depends highly on your depth of field requirements. If you need a huge depth of field, because of critical foreground and background detail, then this can limit sharpness much more than diffraction. On the other hand, if your only important subject is at infinity and your DOF requirement is trivial, then excessively small apertures will make diffraction your limiting factor.

Thus, you want to operate below the diffraction limit when possible, which would basically make film/sensor characteristics and your intrinsic lens resolution (and its aberrations) the ultimate delimiters of system resolution. But sometimes it's not possible to operate below the diffraction limit, because your DOF will be too small for a sharp print if you do.



Now, there are many people on this forum who use digital cameras -- it's just not discussed here. But what I can offer, for what it's worth, is that debates over 'better vs worse' are just not worth wasting your precious brain cells. I mean on another forum I just witnessed a whole bunch of Nikon users claim they were switching to Canon when Canon's new DSLR was introduced -- then a rumor circulated about a new Nikon pro DSLR that prompted Canon users to do the same -- and then when the Nikon rumor proved false, the pendulum swung to the other extreme again. I swear, sometimes watching the digital world is like being a Red Sox fan (and trust me, I am a Red Sox fan) -- just too many highs and lows, too many emotions, too much posturing.

 

[los]

paul may I?


THE END

 

[reub2000]

There isn't a "diffraction limit", image quality simply gets worse the smaller the aperture is. Other things affect the image quality at various aperture, things that I don't really understand.

With large format cameras the loss of quality becomes less significant because the focal lengths are longer. With a 50mm lens at f/22, the diaphragm opening is 2.3mm, while at 150mm, f/22, the lens opening is 6.8mm which results in lot less loss of resolution due to diffraction.

 

[Woolliscroft]

It probably depends on the film. With fast films the film is probably the limiting factor whereas with higher resolution, slow films the lens' limitations might be the more significant factor.

David.

 

[Ed Sukach]

There isn't a "diffraction limit", image quality simply gets worse the smaller the aperture is. Other things affect the image quality at various aperture, things that I don't really understand.

Yes, there IS a "diffraction limit". It is part of the design criteria that determines the smallest (numerically highest) f/stop the manufacturer will allow. Beyond that, the image quality will be degraded (by diffraction) to less than the manufacturer will allow. It is determined mathematically - and not influenced by lens design. I have the formula around here - somewhere - but it has not drawn my interest for long, long time - ever since I got out of the lens manufacturing business.

In short, it is not something I'd EVER be concerned with - the manufacturer has already done all of that that is necessary - or even, reasonable - unless one really WANTS to be anal to a fault.

.

 

[Snapshot]

I want to thank everyone that contributed to this discussion thus far. It has been informative (as usual) for me. Hopefully, I can digest what has been discussed here and put it to good use.

 

[Ole]

"Diffraction limited at f:8" means that at f:8, the resolution of the lens is so good that only the increasing diffaction limits the total resolution on film.

So a lens that is diffraction limited at f:8 is a lot better than one that is diffraction limited at f:16, and this is likely to be visible all the way from wide open to f:22!

 

[Ed Sukach]

"Diffraction limited at f:8" means that at f:8, the resolution of the lens is so good that only the increasing diffaction limits the total resolution on film.

So a lens that is diffraction limited at f:8 is a lot better than one that is diffraction limited at f:16, and this is likely to be visible all the way from wide open to f:22!

Whooo .... You LOST me there, Ole!!


The aperture where diffraction does not degrade the image more than allowed by the manufacturer is ... where?

 

[Ole]

Whooo .... You LOST me there, Ole!!


The aperture where diffraction does not degrade the image more than allowed by the manufacturer is ... where?

That's where you got lost - by assuming it's "allowed by the manufacturer"-

Simply put, an ideal lens would have the best resolution wide open, and only diffraction would reduce the resolution as the aperture is stopped down.

With real lenses, the aberrations and "stuff" makes the wide-open resolution poorer. And those faults decrease as the aperture is decreased.

At some point, the diffraction will become the main limitation on resolution. This can be described as "diffraction limited at..." whatever fstop.

Think of it as two crossing lines in a diagram: One, "lens fuzz" starts high and decreases with aperture. The other, "diffraction fuzz", starts low and increases with aperture. The diffraction-limited value is where the two cross each other.

It's not quite as simple, and the lines aren't straight, and so on. But it's a simple way to see it...

A lens which performs to the limit on resolution determined by diffraction at a large aperture (say, f:4) is very good. another, diffraction limited at f:16, won't perform anywhere near as well at any larger aperture than f:16!

 

[Ed Sukach]

That's where you got lost - by assuming it's "allowed by the manufacturer"-

Uh ... I did not "assume". That is the standard - and gerearlly accepted description of "Diffraction Limit".

Simply put, an ideal lens would have the best resolution wide open, and only diffraction would reduce the resolution as the aperture is stopped down.
With real lenses, the aberrations and "stuff" makes the wide-open resolution poorer. And those faults decrease as the aperture is decreased.
At some point, the diffraction will become the main limitation on resolution. This can be described as "diffraction limited at..." whatever fstop.

I understand what you are writing here. This is the first time I've ever heard of "diffraction limits" described in that way. The idea that "Nothing else really matters except diffraction beyond a certain aperture", is to me, strange. Do you have any references to this use?
I have never seen the characteristics of a lens system described in this manner ... "Diffraction limited at f/ - whatever". Can you direct me to an example?

BTW - All manufacturers ... except those who convert the bottom of Coke bottles into lenses (not necessarily "bad" - see "Holga") DO pay scrupulous attention to design characteristics - resolution is only one of many.

For REALLY "fine" (standard definition of resolution is the minimum distance where two points can be seen as two points - not "lines per millimeter" - although L/mm is a useful derivation), look to Aerial Mapping and Military Recon.... OOPS! Forget that last one.

 

[DrPablo]

I have never seen the characteristics of a lens system described in this manner ... "Diffraction limited at f/ - whatever". Can you direct me to an example?

All manufacturers publish it in their MTF charts. You see lens resolution peak at a certain aperture (often f/8 or f/11 for 35mm lenses), beyond which resolution progressively decreases with narrower apertures.

 

[Ed Sukach]

On further thought...

Isn't that "Diffraction Limited" definition saying,

"Diffraction has no effect on the resolution of the lens until it is stopped down to an aperture of f/... whatever"?

I could have an abysmally poor lens - will not resolve 1 (one) line per millimeter. It does not get worse until it is stopped down to f/4 (the smallest amount of diffraction imaginable has a negative effect on its performance).

Or - given the opposite, "This astoundingly **wonderful** lens is not affected by diffraction until it is stopped down to f/32" ...?

Do I have it "right"?

 

[DrPablo]

Ed, the "limit" part of "diffraction limit" simply describes where lens resolution becomes limited due to diffraction, as opposed to being limited by aberrations or other things intrinsic to the lens.

So diffraction does affect the lens, including when it's wide open, because you're still passing light through a little hole.

But the salubrious effects of stopping down, like decreasing the effect of aberrations, make the increase in resolution outweigh the decrease caused by diffraction. At some point, however, the effects of diffraction become more important than any beneficial effect of stopping down, and that is where it becomes diffraction limited.

As I said in my first post, at that point you have to decide whether your resolution will be limited by diffraction or by depth of field (because, obviously, you can have critical subjects out of focus if you don't stop down enough).

As to your example, diffraction per se is not intrinsic to the lens elements, so it is not a feature of good or bad lenses. It's that the theoretically perfect lens will never benefit from stopping down -- so if you stop down from f/2.8 to f/4, diffraction will limit resolution rather than a smaller aperture improving performance.

 

[Nick Zentena]

Not to my way of thinking.

At F/4 the highest perfect lens is going to be close to 400lpmm. So most lenses won't be made worse by diffraction at this point.

OTOH at some point the best possible resolution will be lower then what the lens can get more open.

To me the point a lens becomes diffraction limited is the point stopping it down reduces resolution. Of course some times the smaller F/stop is more important then the resolution.

 

[Nick Zentena]

I need to type faster