So, like an earlier poster, I am a signal processing person. Here is my perspective on AA in audio. Perhaps some of you have some thoughts on how it applies here.

Aliasing must be avoided at all costs. Any sampling technology must have something ahead of it that limits the frequency of what it receives to half of the sampling rate. Anything else is an ERROR. If you can be absolutely sure that you will never encounter such frequencies, then fine. However, I think in photography that would be equivalent to saying that the resolution of your lens system is low enough that fine details will not appear. That seems a poor bet.

In audio we have sampling rates such as 44khz. Which means that we must never see frequencies above 22khz. The problem is that real world low pass filters have rolloff. So to be sure that 22khz is 96db down (the limit of 16bit samples) requires that frequencies such as 19khz will be somewhat pulled down as well. No one likes this. The solution to this in audio was to over sample. So you sample at 88khz. So now you have to avoid things above 44khz. A real crappy AA filter pulls things down but by the time you get to 22khz there is no rolloff at all. So you over sample, anti-alias, and then down sample to 44khz. Presto, you have great audio frequency response out to the edge of the system.

So applied to photography, it looks like this: at 12mp you use an AA filter and you lose some fine detail because AA filters have roll off. Instead, run things at 36mp (oversample) with an AA and down sample to 12mp and you have essentially perfect 12mp images. Running the 36mp system without an AA would only make sense (to me) if you KNEW that the lenses were low enough resolution that they were doing the AA for you.

So...That leads to my question(s). Are the Nikon lenses high enough resolution to exceed the resolution of the 36mp sensor? I think the answer is "yes". This I suspect because the pixels at 36mp are very similar to the grain size of film. Has Nikon or anyone else done studies or white papers on this?

Thoughts?

TaoTeJared said:
Moire is one of those things that is really hard to remove in post,

Just noticed, LR4 beta, has a moire slider on the adjustment brush , tested it on Niko's sample and it works

PS

The new Clarity slider is awesome
as is Shadows adjustment

@7X - I already have a prime camera, which in my case is d3s, I also have d300s, which I'll be most likely selling this year in order to get d800. as I'm not that much concerned with moire effect I can allow myself to have one pure camera (non AA) in order to get extremely sharp snaps, the way I want.

@mike - it's very expensive hobby, especially in the field of animals and sport photography :(, although I think that d800 will handle the job of d7000, as the dx gonna be around 16mpx, more than I really need.

The idea is that you'd buy a camera without an AA filter as a second, third or fourth body. I see it as similar to owning an IR converted camera. You aren't going to use it all the time and certain photos wont lend themselves terribly well to using it, but you've got it for the things that you know that it will excel at. The ideal future scenario for someone like me is having a D4 or D3s for most of my work, the D800 for premium quality studio and landscape, and the D700 as a backup and lightweight alternative to both. And the D7000 for wildlife. :]

This is going to get expensive! lol

So what I can gather from all of the previous posts is:

An AA filter gets rid of the Moire (funny patterns)that you get on certain items with close spaced lines on them when photographed. But at the expense of a small ammount of sharpness and resoloution.

Without an AA filter, you can use the full potential of the sensor, but with a chance that you will get moire on certain items when photographed.

Now we all have to decide wether we are concerned about the moire.

My thoughts are that this will only be critical when taking portraits or buildings. So I may possibly go for the no AA filter option. But I may be wrong.

Since I havent had the chance to buy my new camera just yet, Ill now wait and see what is in the forthcoming Nikon announcement and decide then if it will be a D700 or D800.

I`d just like to thank everyone for explaining this confusing subject for me.

If I were to buy the D800, it would def be a non AA-filter D800.

Of course, we'll have to see how much I like the D800. I've got to get through the D4 first! ;)

sevencrossing said:
adamz can explaine your choice
I am still trying to get to grips to reason for two versions
am i right in thinking it is one or the other

Think of taking a picture of satin ties or seeing newscasters who's tie is "trippy" looking - more lines, fun-house wavy looking. Basically the cloth's lines exceede the lens and sensor frequency and it converges. An AA filter helps remove that. The way it does that is by having a zebra pattern filter (clear-dark-clear...) that removes every other line - in a way. (I'm keeping it layman simple.) That helps remove lines that have met the resolution frequency of the lens but you loose a small amount of sharpness and resolution. Moire is one of those things that is really hard to remove in post, kind of like adding a polarizer effect in post, it just never quite works well.

If moire is not an issue, remove the filter and you are able to utilize the full potential resolution and sharpness of the sensor. Leica has done this, Sigma doesn't use one, and you can tell a difference if you study the images. The draw back is, if someone is wearing a suit that moires really bad, you are toast.

mirtos said:
Can someone explain why it will be more expensive to have the filter removed? Wouldnt it be cheaper to not include something? Or am i missing something.

It is an exception of normal production thus adds additional time. Just like paying more for a manual transmission on a vehicle. My guess is they are planning on making the changes by hand. That adds time and cost. The actual cost of an AA filter is peanuts next to the cost of hand work on a camera.

I'm willing to bet so many will want one without the AA filter, they will end up running two productions so maybe the cost will drop.

Can someone explain why it will be more expensive to have the filter removed? Wouldnt it be cheaper to not include something? Or am i missing something.

adamz can explaine your choice
I am still trying to get to grips to reason for two versions
am i right in thinking it is one or the other

I will get the non-AA version.

if I will ever get d800 I want the non AA version

So the D800 may be available soon, with and without an A-A filter

Has anyone worked out which one they want

if the main blog is correct $900 is a lot to pay NOT to have something

This was actually included in some posts that got deleted from this thread, Dormant. We got a little far afield, but unfortunately the AA info got lost along with the other stuff!

aesnakes said:
Also Ive gotten some success applying AA using a matrix operation

010
242
010

I'm not disputing the success of your approach, but you cannot apply do anti-aliasing to a digital image unless you are down-sampling afterwards. Aliasing happens when the digital image is created at the sensor, and the AA filter has to happen before this.

Here's an explanation from audio signal processing. If you create a digital signal with a sampling rate of 10 kHz, the highest signal that can be recorded is 5 kHz, or the Nyquist frequency. If you take your 10 kHz recorder and try and record a signal with a frequency of 6 kHz, it will be aliased and appear with a frequency of 4 kHz. There is no way you can turn that 4 kHz signal back to 6 kHz because you cannot distinguish it from a real 4 kHz signal.

Images are the same, but two-dimensional. Images with regularly repeating patterns with a spacing less than the Nyquist frequency (2 pixels) are worst affected. Aliasing will also affect sharp edges, as an edge contains a wide range of frequencies.

Apologies if I've confused anyone.

Drab said:
Could you post a photo somewhere? I don't have a P6000, can't speak to the P6000, but I'd like to see the moire.

Thanks

Hi Drab,

I sold my P6000 a year and half ago to keh dot com, but yes the pattern I got was very similar to what NikoDoby has posted above except instead of the rainbowish colors I got only a single color pattern. The one I recall was on my brother's gray suit and the pattern was silver metallic, intertwined - sometimes in a curvy-wavy pattern, and some times in symmetric zigzag pattern.

I will try to dig up the picture from archive, but this week is busy with homecoming.

True Spy Black, there are differing degrees of AA filter strength but moire will be the strongest when there is no filter.

This is what moire looks like for those who don't know what we are talking about.

It's really not an issue as long as you are aware of how and where it happens. I mostly see it on fabrics with fine detail and on newer cars with carbon fiber body work or certain plastic headlight covers.

You can still get Moire patterns even with the AA filters found in camera sensors. Sooner or later the pattern of a surface, combined with the particular magnification of that pattern on the sensor plane, will cause interferometry with the sensor array and thus Moire.

I've seen this countless times in product shots for catalog work. If the studio photographer is smart, he'll ignore the client/designer request for that "perfect arrangement" and shoot an alt where the sensor's array and the pattern don't clash. Otherwise the only solution in post is typically to blur the Moire pattern and then substitute it with noise. Not always a pretty site (although it does sometimes work well).

Hi all, I need some insight on this AA thing as well. (SO you know where im coming from) Im a VFX artist doing visual fx, do a lot of video and photos every day. Quality is a big issue for me, the image (noise, contrast, black point, sharpness etc) must look exactly the same as it did on import after adding effect CG , 2d elements. I shoot sill elements and video so what would be a better preference..

Ive been able to remove Moire patterns using the YCBCR techniques (convert colorspace to YCBCR and soften the Y channel and it goes away without damaging RGB

Also Ive gotten some success applying AA using a matrix operation

010
242
010

What Im wondering with this is if I got the D800 with AA removed would it be better to applying these method when I need them or is the camera just going to do a different and better job at this in your opinion.

Sorry for the long long post

Nikoner said:
I often got a lot of moire pattern in my P6000, could it be Nikon experimented with a inferior AA filter on it for their upcoming DSLR?

Could you post a photo somewhere? I don't have a P6000, can't speak to the P6000, but I'd like to see the moire.

Thanks

So it will show wrinkles and sagging skin in much more fine details then! ladies are not gonna like this at all.

I often got a lot of moire pattern in my P6000, could it be Nikon experimented with a inferior AA filter on it for their upcoming DSLR?

As P6000 had quite a few innovative features which are now being incorporated with newer camera releases (p&s, mirrorless, dslr):

U1 & U2 settings.
continues shooting in a semi rugged p&s.
filter thread.
GPS.
Network enabled Ethernet connector.

You wouldn't want to shoot portraits without an AA filter. Fabrics are prone to cause moire. You'd probably get moire with fine grids and windows in architecture too. The biggest advantage would be for landscapes.

Whatever you want more detail in and are able to predict, if not control, the risk of aliasing.

All that jargon talk is find & dandy and please carry on, but in simple terms what type of photography (not video) could benefit from a D800 with AA filter removed?

Sports, landscape, cityscape, portrait, macro??? (assume no post processing is done)

Thanks

Absolutely right, Drab. I saw "moire" (time-domain aliasing) today by firing a 160Hz strobe light at a speaker that was being driven by a 162Hz sine wave generator. (I love my job.) This gives a 2Hz beat frequency that you can see: the cone moves in and out as if driven by a 2Hz generator. There are no rectangles there—only discrete pulses of light to sample the position of the speaker.

This is time-domain aliasing, but it's exactly the same principle in space. Eventually the AA filter may be the imperfect lenses we use, but I don't think we're there yet. The fact is, like Drab says, there has to be some filtering done because frequencies just above the Nyquist frequency will be problematic.

This is me thinking out loud: I think it also becomes less of a problem as the sampling becomes less discrete, because that has an effect similar to low pass filtering the signal (i.e. as the photosite becomes better at collecting light from the entire photosite area, very high frequencies are naturally filtered out). This would be analogous to increasing the pulse time of the strobe experiment I mentioned until the pulse time almost equals the time between pulses. In fact, lighting the speaker with an incandescent light is the extreme case of the light on constantly (100% duty cycle), and the aliasing disappears.

TaoTeJared said:
Moire happens when frequencies start to converge which is a basic end effect of Regular Rectangular Patterns on sensors. Computers work in 1 & 0s and has a regular pattern as well so that is why software solutions are haphazard as well. All of it is better solved with physical design then software.

I agree with everything you say except this paragraph.

1 - Regular rectangular patterns are not the cause of Moire. Regular patterns (rectangular or not) at frequencies greater than the Nyquest limit of the sensor are the cause. Moire is the visual artifact of aliasing.

2 - It is irrelevant that computers work in binary. Nyquest and Shannon proved that a digital sampling system can perfectly reproduce a bandlimited analog input so long as there is the said input side bandpass filter and proper reconstruction filtering on the output side.

3 - "Bandlimited" is the key. Once sampling has taken place aliasing can not be filtered out without making assumptions.