11
\$\begingroup\$

According to http://mansurovs.com/nikon-d800-vs-d800e the D800E differs from the D800 in that one of the layers of the filter reverses the effect of the initial low-pass filter. How does this work?

Is the second filter the same as the first, but mounted in a different orientation? Are they identical birefringent crystals? On how many axes must the second filter be aligned with the first? What will I see in the resulting images if there is a misalignment?

If the use of a second filter to reverse the effect of the first arises from the problem that entirely removing the AA filter would change the plane of focus, why not just do that, and calibrate the AF system in the camera accordingly?

Improve this question
\$\endgroup\$
2
  • 6
    \$\begingroup\$ In terms of removing it.... I bet its a thousand times cheaper to just load up a cartridge of a different filter in the manufacturing process than to change the actual build process. Changing automated process is expensive and time consuming, even in the slightest. \$\endgroup\$
    – rfusca
    Apr 23, 2012 at 2:00
  • \$\begingroup\$ I'm still looking for an answer which addresses the second and third paragraphs of my question. \$\endgroup\$
    – James Youngman
    Apr 27, 2012 at 0:24

3 Answers 3

Reset to default
8
\$\begingroup\$

A traditional AA filter consists of two parts, one splits the image horizontally with a ~1 pixel offset (effectively giving one image superimposed on itself with a very small horizontal offset). Behind this is a second filter that does the same vertically. The effect of this is to split each light ray four ways so some of it lands on each of the four RGGB pixels in the Bayer array.

The D800e has the first horizontal split filter but immediately behind it a second horizontal filter which combines the two images to undo the effect of the first filter. The filter material has two refractive indices for different polarizations so each incoming ray (which will contain photons with different polarization) is split into two. If the second filter has equal and opposite refractive indices, the diverged light rays are bent the other way so the they land on top of each other again, cancelling the effect of the first filter. The key point here is that the each of the superimposed images have different polarizations, allowing them to be recombined. If this weren't the case adding a second filter would produce three images not one!

As for why they do this instead of just not installing a filter, rfusca hit the nail on the head, it's much easier (and therefore cheaper) to occasionally swap a batch of vertical filters for horizontal filters when you want to build d800es than it is to disable the filter mounting part of the production line and pass the cameras through to a different AF calibration stage.

In addition to this, modern lenses are designed to correct for the refractive properties of the filter stack. Omitting the AA filter would introduce a subtle aberration into all images shot with digital era glass.

Improve this answer
\$\endgroup\$
1
  • \$\begingroup\$ conslayer's argument about manual focus lenses is also convincing on part 3 of the question. However all round I think this is the best answer to the question asked. \$\endgroup\$
    – James Youngman
    Mar 30, 2013 at 13:21
3
\$\begingroup\$

The answer is in the quoted article.

"We know that the full low-pass filter cannot be completely removed, because it would cause the focal plane to move; plus, the camera still needs to be able to reflect infrared light rays."

Manufacturing concerns, as pointed out by rfusca probably have some contribution, but you still need to have the 3 filters in place for the reasons in the article.

I've run a camera with the infrared filter removed (for astrophotography). I was told that the auto focus would no longer work, but for me it did.

Improve this answer
\$\endgroup\$
4
  • \$\begingroup\$ Removing the IR filter does not affect the AF functionality, aside from introducing a focus offset. The AF filter has a separate IR filter, so if you have a visible-light-cut filter on your lens, it will indeed no longer focus normally. You can remove the AF IR filter, but it's a giant pain in the ass, and I would recommend against it. \$\endgroup\$
    – Fake Name
    May 5, 2012 at 23:04
  • \$\begingroup\$ Actually, recalibrating the AF sensor is pretty easy to do, it a bit monotonous. I did it myself on my Nikon D80 IR conversion. Basically, there are three screws that adjust the position and planarity of the AF sensor relative to the sensor. You take lots of pictures, and adjust the screws until the AF sensor plane is identical to the sensor plane. \$\endgroup\$
    – Fake Name
    May 5, 2012 at 23:07
  • \$\begingroup\$ The only really valid reason I can think of is the one about reflecting IR. Basically, most DSLRs have the hot-mirror as part of the AA filter. As such, they would have to produce a set of hot-mirror only sensor covers instead of the two AA filter thing. It still seems like a ridiculous solution to me. \$\endgroup\$
    – Fake Name
    May 5, 2012 at 23:09
  • \$\begingroup\$ @fakename a lot has been posted about "sensor stack thickness" since this thread. Changing the thickness of the glass in front of the sensor will be bad for dSLR lenses with a close exit pupil. The picture above should be modified to show that the flat glass has refractive effects for rays that come in at an angle (as pictured). \$\endgroup\$
    – JDługosz
    May 22, 2015 at 20:06
3
\$\begingroup\$

If the use of a second filter to reverse the effect of the first arises from the problem that entirely removing the AA filter would change the plane of focus, why not just do that, and calibrate the AF system in the camera accordingly?

  1. Adjusting the AF system calibration does not fix the focus offset for manual focus. You would have to adjust the physical position of the focusing screen as well, so you're back to a manufacturing difference between the camera models.
  2. By removing the filters (replacing them with an air gap, which has a lower index of refraction), you effectively move the sensor farther away from the lens mount (in an optical sense). A lens whose infinity focus is "perfect" at the physical limit of its focusing range will no longer be able to achieve infinity focus when the filters are removed.

enter image description here

Improve this answer
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.