I am aware that most DSLRs use a so called anti alias (AA) filter in front of the sensor to avoid formation of Moire patterns in the original image. The AA filter is made from clear crystalline lithium niobate with different thickness depending on the use in many kind of digital cameras.
Now my question: Why do manufacturers (e.g. Canon) really use this AA filter? It lowers the resolution of the sensor up to 30% - is it just to make sure that nobody sees a bit of a moire pattern when the cameras are tested professionally later? The moire pattern can be removed easily in post processing software, too. Is anybody aware of another reason why AA filters are used? It does not make a lot of sense to me that such kind of filters are added as standard in most digital cameras and which is not providing the best of digital resolution outcome especially when consumers try the best and sharp lenses! Medium format cameras and also Leica's digital M8/M9 cameras do not use AA filters.
Nikon might have realized that this is an issue because they offer now the D800e without anti aliasing. This will increase the resolution and makes most out of the 36 MPixel sensor.
I am also aware that the AA filters are coated to reflect infrared. But the infrared reflection can also be achieved without using an AA filter in the first place. If anybody has some insight especially from the camera manufacturers, please post!
First ... for most people ... doing a good job of removing moire is NOT easily done. To that end the majority of the cameras being sold are to people who either don't have the expertise to do a good job or it ... or don't want the hassle of it.
Second, while it ranges from about 10%-30% with most being in the 15%-20% from what I've seen.
+1 @ Leica & MF (option) not using AA filter. Very different breed (i.e. not the masses) of photographer. Personally, I shoot with the Kodak SLR/c due to it's absence of an AA filter also.
Even though I've chosen to shoot non-AA, moire' still kicks my butt from time to time, but not that often as I don't use it for many moire inducing subjects. I'm a fan of non-AA, but the Nikon D800E is not a true non-AA filter system. It employs a blur/unblur (for 1/2 the direction) attempt. There is a much more detailed thread in the Alt Gear Forum that addresses the D800E and the non-AA filter issues in great depth.
Thanks, Kent! I am aware that the D800e's AA filter is not removed but has an unblur layer on top of the common AA filter to neutralize the blur effect. The way I understand you is that the AA really is just used to suppress moire patterns. I wasn't even aware that this filter blurs so much until I converted one of my DSLRs for infrared where the AA filter is removed. I was very surprised to see that my photos gained in RAW file size compared to my regular color photos in the same camera model and lens setup - and this even knowing that infrared at 720 nm cutoff provides less sharp photos compared to regular color ones! I traced this back to the loss of the AA/blur filter. So I was wondering if the AA filter serves another less well known purpose why it is so common in digital cameras.
There are also a few P&S cameras that do not use an AA filter as well. I recall when Bifurcator posted some P&S stuff that was sharper than DLSR with great glass, only to learn that the P&S had no AA filter.
As to some other reason for an AA filter ... I got nuttin'
I have my SLR/c ... but I've also got my 1D II N / 5D. Interesting to note however, I can still get moire' with those cameras (weak AA filter) when used with stellar glass. Talking with MaxMax, he share his opinion @ moire' that was this: If you aren't getting moire' ... your glass isn't sharp enough.
So, there is a relationship between the glass being used and the need for / not AA filter ... be that P&S or MF.
I'm not sure, but I wonder to shat degree the optics design / transitions come into play. I've not seen a lot of Leica people complain @ moire ... but Leica (and MF) designs their lenses differently than the likes of Canon & Nikon. So, while I don't have a clue @ the optical design relationship ... I wonder if it is part of the equation.
I don't think moire is easy to remove in post-processing. I think that moire (especially color moire) is often impossible to remove using automated tools.
It also seems to me that the mythical "average" shooter would care more about occasional moire on fabric and such (remember, people usually wear clothing and that clothing is usually made out of fabric) which would make the photo ugly/unusable than about ultimate sharpness.
It can also be argued that a lot of what's perceived as detail on AA-less images is actually aliasing artifacts that just happen to provide pleasing "texture" but are not actual real detail.
Thanks, KaaX! But I really scratch my head about this - how often it is discussed here how sharp or soft a lens is, how good or bad a sensor with resolution is - and then there is this AA filter in front of the sensor which blurs out the best resolution you would theoretically get from a very sharp lens! I personally would rather live with a bit of moire effect than being forced to cope with a crippled resolution!
Note -- "discussed here" :-) This is a gearhead pixel-peeper forum :-D Most people, including some very good photographers, don't care that much about pixel-level sharpness.
But yes, there is a group of people who prefer AA-less cameras. They shoot the old Kodaks, or they perform surgery on their cameras, or they line up for the D800e. The number of these people is not large and most of them are landscapers.
Consider, e.g. the preferences of portraitists: would you accept moire in the eyelashes to get a better definition of skin pores? :-)
Non-AA is not for the faint of heart for general shooting. If you are specific at avoiding moire' inducing subjects (although you still get bit by it no matter how careful you are), you can do fine without an AA for a lot of things ... just be aware that it isn't without it's issues.
I'm a fan of Deming, and understand the aspect @ systematic process, with the AA filter being contrary to the holy grail of ultimate sharpness and viewed as the "weak link" in the process that doesn't need to be there, but it is basically a "safety device" (imo) that is necessary to play with a bayer sensor that makes up its decision (i.e. guess) as to what the color should be, and it helps it not get fooled / guess wrong with such scenarios.
I get drawn to the Fovean sensor as the possibility of being non-AA and non-bayer sensor. I haven't ever tried one, so I can't speak to it, but it does capture my theoretical attention. If the mount were EF vs. SA, I'd probably try out the Merrill, but I'm pretty heavily into my alt glass for use on the EF mount.
For most people ... I'd steer them away form non-AA ... even though I do like shooting mine. It does require more thought, and incurs more risk, and warrants more effort. I've seen the crops at the D800 vs. the D800E ... even as a fan of the non-AA ... I'd pass on the D800E, as it isn't void of moire' issues, and it isn't even a true non-AA ... it's kinda half-way is, half-way isn't ... so why bother
Rather, I'd recommed people step up their glass (old school) to the likes of shooting great primes rather than consumer zooms if they really want excellent sharpness. I find there is more to be gained in great glass, than their is to be lost in a (mild) AA filter. For that reason, I have little worry @ switching from my SLR/c to my 1D MK II N / 5D. I use a cadre' of primes including Oly, Mamiya, Leica, Nikon, Zeiss ... and even Canon of my EF bodies. My few zooms are nice, but when I want to 'obsess' over detail ... I know the specific nuances of my various primes. I mean, we are talking about obsessing over detail, right.
The AA filter is simply more comvenient for most, with only a modest amount of penalty that can also be compensated for in PP ... more easily than trying to correct for moire'. If you want the best of the best, then mating the best glass with the best non-AA filter makes sense (Deming 101), but from a pragmatic perspective for the masses ... non-AA systems are not ever going to a high priority for mainstream (Canon/Nikon/etc.) DSLR mfr's.
RustyBug wrote:
I get drawn to the Fovean sensor as the possibility of being non-AA and non-bayer sensor. I haven't ever tried one, so I can't speak to it, but it does capture my theoretical attention.
Conveniently, dpreview did a review (or re-review? :-) ) of the Sigma's DSLR. Basically their conclusion about the sensor is that at low ISO it provides excellent detail and colors, but at high ISOs it just horribly falls apart.
There's also a bunch of issues with the camera itself (15 seconds to write one RAW image to card?!!) and, of course, there is (or, rather, isn't) the glass.
Well, the "falls apart" at high ISO is kind of "moot" if you ask me. If you are shooting for max detail ... you are never shooting at high iso. That's what tripods are for.
15 seconds ... WOW ... and I thought my SLR/c was slow.
And that "glass thing" ... that's a killer problem. Why do I want to use Sigma glass on a Fovean, when I could be using Zeiss or Leica or Mamiya (obsessing @ detail again) or some of the killer Canon/Nikon stuff. It is just a bit counter-productive and now we've made the glass the "weak link" to maximum detail ... i.e. robbed Paul to not pay Peter ... ooops.
I'm not bashing Sigma glass (I like my 100-300/4), but getting real ... the Canon 24L TS-E II smokes anything on earth for that FL ... but I can't use it on Fovean. Same could be said for Zeiss 100/2 or Contax N realtive to their drawing style/sharpness ... and it is a long list of glass that this could apply to, with Sigma glass not being on THAT list.
Yeah, they have just missed the boat a bit for me when it comes to the glass.
retrofocus wrote:
Thanks, Kent! I am aware that the D800e's AA filter is not removed but has an unblur layer on top of the common AA filter to neutralize the blur effect. The way I understand you is that the AA really is just used to suppress moire patterns. I wasn't even aware that this filter blurs so much until I converted one of my DSLRs for infrared where the AA filter is removed. I was very surprised to see that my photos gained in RAW file size compared to my regular color photos in the same camera model and lens setup - and this even knowing that infrared at 720 nm cutoff provides less sharp photos compared to regular color ones! I traced this back to the loss of the AA/blur filter. So I was wondering if the AA filter serves another less well known purpose why it is so common in digital cameras....Show more →
In an aliased image detail can't be distinguished from artifacting without a person guessing where detail is and where artifacting is, and the process to remove it necessarily results in an image with less fine detail than would have been there if an appropriately AA filtered camera had been used.
That said, people seem to like artifacting (calling it "crisp") so...
There are companies that will remove the aa filter, as it improves things a lot for video. People are also diy'ing (below). I guess with the lower resolution of video, moire is not a problem with the AA filter removed. Rumor is that the next EOS Cinema camera will not have the AA filter.
KaaX wrote:
Note -- "discussed here" :-) This is a gearhead pixel-peeper forum :-D Most people, including some very good photographers, don't care that much about pixel-level sharpness.
But yes, there is a group of people who prefer AA-less cameras. They shoot the old Kodaks, or they perform surgery on their cameras, or they line up for the D800e. The number of these people is not large and most of them are landscapers.
Consider, e.g. the preferences of portraitists: would you accept moire in the eyelashes to get a better definition of skin pores? :-)
Yep, moire might really be an issue in some kind of shooting. Best option would likely be if the photographer had the option to move in or out such AA filter depending on the need.
You will also see the effects of diffraction much more readily sans AA filter. Technically the higher the resolution the less chance of seeing moire. You push the interference effect to smaller scale. You shouldn't see it in say fabric i the pixels are now much smaller, but may see it in macro for example where the insect has very minute repeating pattern.
1. Has anyone used the LR4 Moire brush? Does it help when you have an image from a non AA cam that has Moire?
2. I don't know to much about less moire as pixels get smaller. I'll trust Pixel on that as he knows far more than I ever will on many subjects . But I saw a thread on the Nikon forum that said a guy used a D800 (standard NOT the E) and was shocked at the Moire it gave from people clothing. The assumption was/is if the standard one gives moire then how bad will the d800E be?
retrofocus wrote:
I am aware that most DSLRs use a so called anti alias (AA) filter in front of the sensor to avoid formation of Moire patterns in the original image. The AA filter is made from clear crystalline lithium niobate with different thickness depending on the use in many kind of digital cameras.
Now my question: Why do manufacturers (e.g. Canon) really use this AA filter? It lowers the resolution of the sensor up to 30% - is it just to make sure that nobody sees a bit of a moire pattern when the cameras are tested professionally later? The moire pattern can be removed easily in post processing software, too. Is anybody aware of another reason why AA filters are used? It does not make a lot of sense to me that such kind of filters are added as standard in most digital cameras and which is not providing the best of digital resolution outcome especially when consumers try the best and sharp lenses! Medium format cameras and also Leica's digital M8/M9 cameras do not use AA filters.
Nikon might have realized that this is an issue because they offer now the D800e without anti aliasing. This will increase the resolution and makes most out of the 36 MPixel sensor.
I am also aware that the AA filters are coated to reflect infrared. But the infrared reflection can also be achieved without using an AA filter in the first place. If anybody has some insight especially from the camera manufacturers, please post!...Show more →
it's not that easy to remove in post and when you do you get blurring away anyway, it's worse than real AA filter, some tricky stuff wouldn't be fixed without making it look a lot blurrier than AA to begin with I believe
the 5D3 has only 20D pixel pitch I don't think it's nearly close enough to start doing away with AA filter, even the D800e seems borderline, it's not even 7D pixel pitch
