Pages: [1]   Go Down

Author Topic: Nikon D800/D800E and stopping down  (Read 2284 times)

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Nikon D800/D800E and stopping down
« on: March 16, 2012, 01:25:05 am »

Hi,

A year ago I made an apertures series on my Sony Alpha 55 checking out the effects of defocusing and diffraction. To me it was quite obvious that diffraction set in early, with the lens I used (100/2.8) macro I had optimum sharpness at f/5.6.

http://echophoto.dnsalias.net/ekr/index.php/photoarticles/49-dof-in-digital-pictures?start=1

On the other hand, diffraction is relatively benign to deconvolution sharpening. I would suggest that using a "gaussian blur" in Photoshop's smart sharpen can be used to reduce the effect of diffraction. Sharpening matters a lot. To begin with, some detail can be regained, if the PSF (Point Spread Function) is good enough as long as MTF is reasonably high, around 20%. AA-filtered systems generally end up around 20% at the Nyquist limit.

The enclosed sample shows the effects of diffraction and defocus and extensive sharpening.

The Sony Alpha 55 used for these tests has the same pixel pitch as the Nikon D800/D800E but in an APS-C package, so the results would be perfectly applicable for Nikon D800.

Best regards
Erik
Logged
Erik Kaffehr
 

marcmccalmont

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1780
Re: Nikon D800/D800E and stopping down
« Reply #1 on: March 16, 2012, 03:47:42 am »

I've been shooting my IQ180 at f8.0 (pixel pitch 5.2um) with very sharp results (Rodenstock HR lenses)
I figured the D800 at a pixel pitch of 4.8um should be the same, perhaps f7.1 would be better?
I kind of stumbled into being lens independent of camera. All the discussions of the Leica M lenses on the NEX7 and Leica R lenses being both equally superb and reasonably priced I purchased two zooms for landscape work on my 5DII replacement which I thought was going to be a 5DIII. The Leica 28-90 2.8-4.5 R and the 80-200 4.0 R both being as good as their f2.8 bigger brothers stopped down. Since for most landscape work these will be used at f8.0 (maybe f7.1) lighter weight, more reasonable cost was a consideration. At any rate now that I've decided on the D800E all I need are the Nikon Leitax adapters :)
Marc
Logged
Marc McCalmont

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Nikon D800/D800E and stopping down
« Reply #2 on: March 16, 2012, 04:35:57 am »

I've been shooting my IQ180 at f8.0 (pixel pitch 5.2um) with very sharp results (Rodenstock HR lenses)
I figured the D800 at a pixel pitch of 4.8um should be the same, perhaps f7.1 would be better?

Hi Marc,

My rule of thumb is that when the diameter of the diffraction blur pattern exceeds 1.5x the sensel pitch, the effects of diffraction blur will start to be come visible at the pixel level. The microcontrast will be visibly reduced. That should not be a surprise, because starting at that diameter there will be 8 surrounding pixels that add a part of their infuence as well.

Using that criterion, it becomes easy to calculate the aperture where diffraction will start to take it's toll. For green light, which contributes most of the luminance signal, just multiply the sensel pitch in microns by 1.108 and the f-number is the result. So for a 5.2 micron sensel pitch that means f/5.7, and 4.8 micron means f/5.3 will be the onset of visible diffraction limitation. A lens that is free of most resolution limiting aberrations at that aperture is assumed. Many lenses will benefit from stopping down further to improve corner resolution, at the expense of center resolution.

Mind you, that doesn't mean that all resolution is lost at narrower apertures, but it does mean that low contrast microdetail becomes harder to reconstruct. It also means that aliasing artifacts will start being suppressed a bit. Restoration with an accurate Point Spread Function (PSF, model of the blur pattern for the lens at hand) can be quite effective if deconvolution sharpening is used. The issue then becomes finding the optimal PSF to use.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==
Pages: [1]   Go Up