[c-meier]

As you know, the reason most digital camera sensors, save for high resolution backs, usually have so-called AA filters is to prevent the phenomenon known as moiré: that pesky low-frequency ghost image that appears superimposed on images of fine, typically man-made, materials.
The imminent introduction of the AA-free not-that-high-res K5IIs got me thinking: will photographers have to learn to deliberately introduce diffraction when confronted with moire-inducing situations? Or will a market for optional AA filters spring up to cater for the K-5IIs owner that wants to do shoot fashion or a wedding?

Then I had an idea.
You know, the sort of idea you think is simple and clever, but are not sure if makes sense to anyone else. So here it goes.
Pentax*: add an option, let's call it "meier moiré suppression", to your AAless cameras, which uses sensor shift to introduce artificial, deliberate 1px motion blur in all directions. Voilá! Instant low pass filter.

What do you think?

* If you read this, I'll gladly give up the rights to this idea for the low-low price of 1 of each camera that uses the technology. Oh, and a DA*50-135, please.

---

Last edited by c-meier; 09-12-2012 at 01:48 AM.

[Na Horuk]

Oh yeah, I think this idea was had before, and in theory it should work. Not sure Pentax would use it, though. I mean, if they had, they would have announced it.

[Timd]

With the advance in digital processing, I do not know why they do not introduce the filtering in software rather than a mechanical solution.

[Rondec]

The question I have is how much will moire be an issue with this camera? Everything I have read about the Nikon D800e says that it has minimal problems with moire and the K5IIs will have the same pixel density as the D800. Once you get to a high enough pixel density (probably a little over 30 megapixels for APS-C), you don't need an AA filter any more.

[c-meier]

With the advance in digital processing, I do not know why they do not introduce the filtering in software rather than a mechanical solution.

By the time the data is in software the problem is much worse. ("In this house, we obey the laws of information technology!") It would be like stopping making sharp lenses because sharpening can be applied in post.

Once you get to a high enough pixel density (probably a little over 30 megapixels for APS-C), you don't need an AA filter any more.

The higher the res, the finer the detail has to be to induce moiré, and the more negative effects the conventional AA filter has.

[kh1234567890]

Once you get to a high enough pixel density (probably a little over 30 megapixels for APS-C), you don't need an AA filter any more.

Only because your pixels get so small that you hit diffraction limitations.

The problem is inherent in the Bayer filter design, where there are gaps between neighbouring pixels of the same colour sensitivity (effectively giving spatial undersampling). The AA filter does not just 'smudge' the image, it ensures that any given image point is spread both horizontally and vertically (but not diagonally). This, combined with the working of the demosaicing matrixing algorithm, usually reduces moire effects to acceptable levels.

There are also other problems caused by the Bayer filter design and the necessary spatial interpolation, for example false colour banding and jaggies at diagonal edges etc.

The answer is the Foveon sensor approach, with no spatial gaps between pixels. But the Foveon design introduces other potential problems and is more difficult to fabricate

[Timd]

With the advance in digital processing, I do not know why they do not introduce the filtering in software rather than a mechanical solution.
Originally posted by c-meier
By the time the data is in software the problem is much worse. ("In this house, we obey the laws of information technology!") It would be like stopping making sharp lenses because sharpening can be applied in post.By the time the data is in software the problem is much worse. ("In this house, we obey the laws of information technology!") It would be like stopping making sharp lenses because sharpening can be applied in post.

What the low pass filter does is take information out of the image (blur it); what sharpening does is put information in the image. You can only do so much post processing sharpening - a sharp lens is the answer. However to de-moire an image is to filter (take out / reduce) the information. Thus it should be done PP. Just a thought!

[c-meier]

. However to de-moire an image is to filter (take out / reduce) the information.

But... how do you know what information to take out? Before sampling it is easy: all signals that have a period of less than 2 pixels. Hence the artificial motion blur suggestion. But after that, it becomes hard (or impossible, really) to discriminate the alias from the desired image.

Of course, if there aren't any high spatial frequency patterns to begin with, which is most of the time, this is all an non-issue. But when you are shooting the fine texture of that bride's veil or the delicate strands in the feathers of a bird, it would come in handy.

[c-meier]

I told you guys...

PENTAX K-3The new flagship of the PENTAX K digital SLR camera series, featuring an array of new technologies including a new anti-aliasing simulator?RICOH IMAGING

Now, Pentax, ehem:

* If you read this, I'll gladly give up the rights to this idea for the low-low price of 1 of each camera that uses the technology. Oh, and a DA*50-135, please.

[JohnnyXD]

Holy crap, we got a visionary.

[Mystic]

Oh man you are the scientific prophet! Very i.pressive indeed, concerning that you wrote it a year ago. Congrats. I believe you should freely ask pentax for the camera +your desired lens if i would be them, i'd give it to you. I even would vote that pentax would give the camera to you :-) any other revolutionary ideas for the next pentax flagship? :-)

I told you guys...

PENTAX K-3The new flagship of the PENTAX K digital SLR camera series, featuring an array of new technologies including a new anti-aliasing simulator?RICOH IMAGING

Now, Pentax, ehem:

[Ex Finn.]

You did file a patent...I hope.

[Giklab]

You did file a patent...I hope.

I don't think he needs one to claim originality. He posted it here, under a name registered to an e-mail address which can be traced back to a specific person (him).

[c-meier]

Ok. I don't have much hope, but you have convinced me: I am going to write them an e-mail.
I'll let you know how it goes

Edit: Now I am wondering whom I should contact... marketing? public relations? in Japan or the US? Jim Malcolm? Their e-mail is harder to find than I thought...

[c-meier]

Oh man you are the scientific prophet! Very i.pressive indeed, concerning that you wrote it a year ago. Congrats. I believe you should freely ask pentax for the camera +your desired lens if i would be them, i'd give it to you. I even would vote that pentax would give the camera to you :-) any other revolutionary ideas for the next pentax flagship? :-)

Thanks for the praise and support!
As for other "revolutionary" ideas: no, not at the moment. But the sensor-shift anti-aliasing was somewhat of a sudden inspiration, so who knows...

I don't think he needs one to claim originality. He posted it here, under a name regidtered to an e-mail address which can be traced back to a specific person (him).

I am not a lawyer, so, legally, I have no idea where I stand. Maybe my post constitutes "prior art" or something similar, maybe not, depending probably on a lot of things, including when the relevant patent was filed. Obviously, the moral question is more relevant than the legal one... since only a camera and lens are in question.

I'll let Pentax decide if I earned it or not If I manage to contact them, that is.

EDIT: Sent the email to Ms. Amber Renard of public relations, only person whose e-mail I found. pentaxinfo@pentax.com doesn't seem to be working, either.