5dIII Removing AA filter

midluk said:
Mt Spokane Photography said:
I could use a better AA filter on mine. Imagine this photo with no AA filter. That photo was the one with the least Moire, I finally gave up.
I can't see moire at 100%. The moire is a problem of the downsampling process for reduced size preview. Nothing an AA filter on the sensor will change.

Absolutely right. Write "moire" into google, pick images, and scroll trough. You´ll see awesome optical effects on the whole display :)

But the question still holds. Is it possible to remove AA filter if you are friend with electronics and careful, without any harm or malfunction?
As it has been presented, many cameras use two filters stacked "against" each other. Well, if you remove these, what with empty space...
 
Upvote 0
crashpc said:
midluk said:
Mt Spokane Photography said:
I could use a better AA filter on mine. Imagine this photo with no AA filter. That photo was the one with the least Moire, I finally gave up.
I can't see moire at 100%. The moire is a problem of the downsampling process for reduced size preview. Nothing an AA filter on the sensor will change.

Absolutely right. Write "moire" into google, pick images, and scroll trough. You´ll see awesome optical effects on the whole display :)

But the question still holds. Is it possible to remove AA filter if you are friend with electronics and careful, without any harm or malfunction?
As it has been presented, many cameras use two filters stacked "against" each other. Well, if you remove these, what with empty space...

I read something about how the AA filter is set up in the 5DS / R and that the R version actually still has one but they added an extra filter that reverses the effect as this was easier to implement than to change the design and subsequent manufacture of the camera. This leads me to believe you wouldn't be able to simply remove it from the 5DIII without mucking up the optical properties in a negative way. Maybe a solution would be to add something rather than remove it.

I'll try and find that source.

Edit - found it.

"Chuck Westfall (Canon USA) explains the difference between the 5Ds and 5Ds R: "The EOS 5DS uses a conventional low pass filter design in which a single point of image data entering the filter is first separated into two points and ultimately into four points by the time the data reaches the image sensor. By comparison, the EOS 5DS R uses a different low pass filter design in which the single point entering the filter is first separated into two points and then the two points are merged back into a single point by the time the data reaches the image sensor."

Using the optical cancellation technique (vs. complete filter removal) greatly reduces development and implementation costs. The "R" delivers sharper images, but moiré and aliasing are potential side effects, notably in scenes that include patterns repeating at a specific frequency. Fortunately, Chuck expected this phenomenon to rarely occur and indicated that "If it does, it can usually be mitigated during post processing."
 
Upvote 0
Jul 21, 2010
31,186
13,043
Zv said:
I read something about how the AA filter is set up in the 5DS / R and that the R version actually still has one but they added an extra filter that reverses the effect as this was easier to implement than to change the design and subsequent manufacture of the camera. This leads me to believe you wouldn't be able to simply remove it from the 5DIII without mucking up the optical properties in a negative way. Maybe a solution would be to add something rather than remove it.

I'll try and find that source.


Correct about the 5DsR, but as you say the reason for the self-canceling AA filter is to save development/manufacturing costs (having effectively the same filter stack means no adjustments for the self-cleaning sensor motor, etc.).

Full removal of the AA filter from a 5DIII is technically not possible. The portions that are in the self-cleaning assembly (one sheet of lithium niobate crystal and a quarter wave plate) can be removed and replaced with clear glass. The second sheet of lithium niobate is bonded to the sensor stuck and cannot be removed.
 
Upvote 0
neuroanatomist said:
Full removal of the AA filter from a 5DIII is technically not possible. The portions that are in the self-cleaning assembly (one sheet of lithium niobate crystal and a quarter wave plate) can be removed and replaced with clear glass. The second sheet of lithium niobate is bonded to the sensor stuck and cannot be removed.

Maybe not "full" but technically it *is* possible and has some effect (but not recommended as it quite difficult):

http://www.eoshd.com/2012/03/james-miller-removes-optical-low-pass-filter-from-5d-mark-iii-for-resolution-increase/

And it has it share of problems:

"James says he needs to replace the filter glass with a clear piece of filter to prevent back-focus issues from rearing their head."

I suspect you'd need to use an IR filter as well.
 
Upvote 0
Mar 25, 2011
16,847
1,835
midluk said:
I can't see moire at 100%. The moire is a problem of the downsampling process for reduced size preview. Nothing an AA filter on the sensor will change.

Moire is explained by the Nyquist Theorem. Removing the AA filter just means you need to limit high frequencies in a image in a different way.

Digital cameras use a combination of a physical AA filter as well as a digital AA filter. As you get more pixels, the Nyquist frequency rises, so you can use less filtering to get needed detail.
 
Upvote 0
Jan 29, 2011
10,673
6,120
Mt Spokane Photography said:
midluk said:
I can't see moire at 100%. The moire is a problem of the downsampling process for reduced size preview. Nothing an AA filter on the sensor will change.

Moire is explained by the Nyquist Theorem. Removing the AA filter just means you need to limit high frequencies in a image in a different way.

Digital cameras use a combination of a physical AA filter as well as a digital AA filter. As you get more pixels, the Nyquist frequency rises, so you can use less filtering to get needed detail.

Yes, but that doesn't alter the fact that there is no moire in this image at 100%, therefore it is not sensor/AA filter induced moire. The moire visible in the smaller versions is, as midluk said, produced by the resampling process, not the initial capture.
 
Upvote 0
privatebydesign said:
Mt Spokane Photography said:
midluk said:
I can't see moire at 100%. The moire is a problem of the downsampling process for reduced size preview. Nothing an AA filter on the sensor will change.

Moire is explained by the Nyquist Theorem. Removing the AA filter just means you need to limit high frequencies in a image in a different way.

Digital cameras use a combination of a physical AA filter as well as a digital AA filter. As you get more pixels, the Nyquist frequency rises, so you can use less filtering to get needed detail.

Yes, but that doesn't alter the fact that there is no moire in this image at 100%, therefore it is not sensor/AA filter induced moire. The moire visible in the smaller versions is, as midluk said, produced by the resampling process, not the initial capture.

There IS false coloring on the full size version, though. I suspect that is a side effect of the same problem that causes moire.

Thing is, if you do have issues with moire, you can always defocus ever so slightly. It only happens when the frequency of the object being imaged is similar to the frequency of the sensor pixel matrix as focused at the image plane.
 
Upvote 0