Will anti-aliasing filters doom even forthcoming Canons to early obsolescence?

[vbi]

I have the usual Canon suspects (5D2 and 7D) as well as the X100. The IQ from the X100 is simply amazing, easily equal to that of the 5D2, definately superior to the 7D. So far I have not picked up any problems with moire, but then I don't use it for fashion where it might be more prevalent.

However, I would never think of changing my 5D for an X-Pro for model portfolios or wildlife images. The deciding factor for me is the extensive range of lenses and the SLR feel.

The X100 is an excellent travel camera, making one step back and actually be part of the experience rather than being slightly remote. And it is light enough to carry around all day with ease.

So, to get back to the OP's original question...the 5D2's IQ is good enough that one can continue to use it for many, many years. I am sure the same applies to the 1D4 and the new 1DX (my next aquisition). The 7D will not fare so well in terms of IQ, but the superior AF system and speed will guarantee it's usefulness for many years as well.

I think a camera's virtue is about it's usefulness (appeal?) as a whole as opposed to 1 or 2 innovations.

[RAKAMRAK]

@kapanak_ I am not an expert but as far as I have read a 15MP Foveon sensor is actually a 5MP sensor, since the RGB channel captures the three primary colours separately it becomes 5*3=15MP.... Or did I read it wrong. Please those who know more about the science behind the thing, could you confirm?

[necator]

@necator -- from your experience, and if you have the time, a list of types of "moiree-critical things" would be very interesting.  We hear and see often this effect from repetitive fine patterning on fabric.  What else?

For me it's fabric ... since I do mostly sashion and science. Within science I experienced no problem to date, with fashion sometimes it can be a issue, esp. when doing lookbooks, where you can't change the distance to the model, since each picture has to look alike. (Still the 5Dc has a AA-filter, so without it it would happen more often).

Which clothes: Hard to say. Of course those with fine patterns, made out of the threads. But not with all. Sometimes hard to predict. However it happens seldom enough to not really beeing worried about it.
Most times I also saw it together with the Canon 100mm f2.8 @~f5.6. (So, lens must be pretty good at resolving fine structures also.)

Oh, and forgot: Never had any Moiree on my leisure-pictures (family, holiday, parties, ...)

[KacperP]

@kapanak_ I am not an expert but as far as I have read a 15MP Foveon sensor is actually a 5MP sensor, since the RGB channel captures the three primary colours separately it becomes 5*3=15MP.... Or did I read it wrong. Please those who know more about the science behind the thing, could you confirm?

You read wrong. Sensor from Sigma is 4800x3200x3 layers and while it's 15 MP in practise, it is being advertised as 46 MP.
http://www.sigmaphoto.com/shop/sd1-dslr-sigma

[torger]

AA filters are good. They are quite weak in current cameras so not much resolution is lost, but moire is avoided and a pleasing softness is added which allows for huge magnification without pixelation.

If you want higher resolution, increase pixel count but don't remove the AA filter.

The worst kind is low pixel count cameras without AA, (ie sigma), producing aliased pictures which really can't be printed large without visible pixel artifacts.

[JerryBruck]

http://nikonrumors.com/2012/02/04/why-remove-the-anti-aliasing-aa-filter-in-the-nikon-d800e.aspx/

[wockawocka]

As I shoot portraits and weddings I'd much rather not have an AA filter on my sensor.

Much sharper images, so much so I'm testing out medium format for my portraiture.

Once you get to around 40mp the need for an AA filter is greatly reduced.

[Positron]

As a Computer Science major with an Electrical Engineering concentration, I have a fairly good grasp on the general concepts of aliasing and filtering. What I don't understand is why having an anti-aliasing filter should ever be problematic for images. As I understand they are implemented as low-pass filters in front of the sensor, which prevents high-frequency data (that data in which there is extremely rapid change from pixel to pixel) from getting through to the sensor, while letting everything else in. High-frequency data could be noise or moire, but the threshold for "bad" data is so high that there's no reason that the data you actually want should ever be stopped by it. That is to say, the local frequency of moire is so much higher than the frequency of, say, a transition from a person's clothing to the background behind them, or even the transition between edges on a repeating pattern, that I don't understand why having an anti-aliasing filter could ever be detrimental to any but the most extreme images.

Please enlighten me. 

[JerryBruck]

@Positron: Does this do it?
http://nikonrumors.com/2012/02/04/why-remove-the-anti-aliasing-aa-filter-in-the-nikon-d800e.aspx/#comment-230160

I don't pretend to know how significant the advantage will be, but it's interesting that several posters who seem to know, declare that moire effects are not triggered very often.

[jrista]

As a Computer Science major with an Electrical Engineering concentration, I have a fairly good grasp on the general concepts of aliasing and filtering. What I don't understand is why having an anti-aliasing filter should ever be problematic for images. As I understand they are implemented as low-pass filters in front of the sensor, which prevents high-frequency data (that data in which there is extremely rapid change from pixel to pixel) from getting through to the sensor, while letting everything else in. High-frequency data could be noise or moire, but the threshold for "bad" data is so high that there's no reason that the data you actually want should ever be stopped by it. That is to say, the local frequency of moire is so much higher than the frequency of, say, a transition from a person's clothing to the background behind them, or even the transition between edges on a repeating pattern, that I don't understand why having an anti-aliasing filter could ever be detrimental to any but the most extreme images.

Please enlighten me. 

An AA (low-pass) filter can  be problematic when the frequencies it blurs are larger than the frequencies the sensor can still capture properly, and to a lesser degree, when it does not blur frequencies that are too small to be usefully capture properly. Generally speaking, the latter is not really a problem, and I would opt for that over no AA filter at all. However, if the AA filter is blurring frequencies that are still resolvable by the sensor, then that is eating away at GOOD detail, and thats not a good thing. Some of Canon's camera's have had low pass filters that are just a bit too aggressive, resulting in softer images than the competition at similar pixel sizes. That means your losing a small amount of detail (permanently, although you can use post-process sharpening to "recover" it...or rather, lessens softness), and losing it without cause.

[RedEye]

As I shoot portraits and weddings I'd much rather not have an AA filter on my sensor.

Much sharper images, so much so I'm testing out medium format for my portraiture.

Once you get to around 40mp the need for an AA filter is greatly reduced.

+1 for medium format, me too.  If the next pro DSLR is not a close contender to at least an older 28MP Phase One I'll basically be out of the DSLR market.