Pages: [1]   Go Down

Author Topic: depth of field MF vs. 35 mm  (Read 1713 times)

craigrudlin

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 68
depth of field MF vs. 35 mm
« on: August 15, 2011, 06:13:21 pm »

I believe I read somewhere that the depth of field for MF is different than for 35 mm.
That is, if I feel that I need f/16 on my full frame nikon 35mm camera to obtain sufficient
depth of field, that I would use a different f-stop in MF.

What I am pondering is "diffraction."  I have read several times that on 35mm full frame
sensors, that going above f/16 (and for some lenses, less than this, for it may vary with lens),
diffraction will actually begin to erode the resolution and image quality.

Yet, I notice that a number of MF lenses have f/32 as an aperture option (which is certainly
beyond any of my pro-level nikon lenses).  So, when, typically does diffraction become an
issue for MF?

Am I correct to consider that diffraction is a "function" of the size of the lens' aperture opening
and the spacing of the pixels on the sensor?

Thank you.

craig

craig rudlin
Logged

ondebanks

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 858
Re: depth of field MF vs. 35 mm
« Reply #1 on: August 16, 2011, 04:35:22 am »

Ah, Mr. Rudlin. You have uncovered medium format's dirty little secret. I congratulate you. But now, you must die, to protect the secret! Guards, take him to the shark tank!

Maybe I should get out of my Bond villain mode to discuss this further...just give me a second to set this ridiculous cat down :D

What you've hit upon is that in a scenario where a wide depth of field is required, whatever IQ gain you make from the larger sensor and more resolute longer focal length lenses in MF, is instantly eroded by the greater diffraction arising from the requirement of using a smaller aperture (larger f-stop value) to achieve that equal DOF.

Yes and no.

You are correct in thinking that diffraction is an absolute: all systems operating at a given f-stop, like f/11, have the same linear resolution, and for a given pixel size, all systems operating at the same f-stop are equally affected by diffraction. But there are other factors which come into play too.

This is an old debate, going back to the film days. Once you factor in DOF and diffraction, why use 4x5 inch sheet film or 6x7 cm rollfilm over 35mm? Then, as now, the winning argument was that even if a 50mm lens at f/11 projects a smaller but identically resolved image to a 100mm lens at f/22, spreading the same FOV over much more film real estate always gave you finer grain, better tonality, and actually greater sharpness and detail, because your optical system is rolling off at spatial frequencies where the film itself has much higher MTF.

Then, as now, using a system where the lens or film back could be tilted to avail of the Scheimpflug principle also neutered the problem - it decouples the taking aperture from determining the depth of field in the plane of interest.

And if you can't use tilts, then with Bayer pixel arrays, it is better to drop a diffraction-fattened f/16 PSF onto the pixels than to drop a narrower f/11 PSF onto the same size pixels in a smaller format sensor. If neither camera has an anti-aliasing filter, there will be more colour moire with the f/11 PSF. Or if the smaller format sensor has an AA filter to avoid moire (they nearly all do), then it is basically smearing the PSF out to a similar extent as the medium format system with more diffraction. Thus the image quality per pixel is about the same, but the MF system still has more pixels.

And finally, there will be many times when you will either be shooting something at "infinity" and have no DOF concerns about the foreground, or when you want to exploit the thin DOF of a fast lens wide open - and then you will get the full benefits of MF lenses and sensors, with no concerns over stopping down and diffraction.

Ray
Logged

TH_Alpa

  • Full Member
  • ***
  • Offline Offline
  • Posts: 214
Re: depth of field MF vs. 35 mm
« Reply #2 on: August 16, 2011, 05:13:50 am »

Craig,

Diffraction is also a function of the reproduction scale: the larger, the more diffraction. In other words, one has the less diffraction when the lens is set to infinity.

Best regards
Thierry


Am I correct to consider that diffraction is a "function" of the size of the lens' aperture opening
and the spacing of the pixels on the sensor?

craig

Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: depth of field MF vs. 35 mm
« Reply #3 on: August 16, 2011, 12:40:21 pm »

Thierry,

Effective aperture is changing when going close up. At 1:1 its normally two stops down, so f/22 on the aperture ring is about f/45 at life size.

The above relation is pretty exact for symmetrical lenses. For asymmetrical designs it's still a decent approximation.

Best regards
Erik


Craig,

Diffraction is also a function of the reproduction scale: the larger, the more diffraction. In other words, one has the less diffraction when the lens is set to infinity.

Best regards
Thierry

Logged
Erik Kaffehr
 

TH_Alpa

  • Full Member
  • ***
  • Offline Offline
  • Posts: 214
Re: depth of field MF vs. 35 mm
« Reply #4 on: August 16, 2011, 01:01:16 pm »

Yes, Erik, that's true of course, am aware of this relation. Yes, at 1:1 one needs 2 f-stops more light, because the effective aperture is not the same.
Effective aperture is always for infinity.
This being said, the diffraction "rings" still become bigger, when the distance to the image plane increases, making it more critical when stopping down.
As an example, when using 4x5", at a scale of 1:1, the critical aperture (= K = when the diameter of the diffraction "rings" equals the one of the CoC) is f45, when it is f128 when at infinity.
But you are absolutely right, it is all in relation, the distance, the reproduction scale and relative/effective aperture.

Thierry

Thierry,

Effective aperture is changing when going close up. At 1:1 its normally two stops down, so f/22 on the aperture ring is about f/45 at life size.

The above relation is pretty exact for symmetrical lenses. For asymmetrical designs it's still a decent approximation.

Best regards
Erik


Logged
Pages: [1]   Go Up