Are you ok with AA filter bodies for shooting wildlife?

[sid_19911991]

Hi folks,

I would love your take on the subject.

I have found DSLRs/mirrorless cameras with AA filters (eg. A9 i, Z6 ii, etc.), particularly heavy ones, giving soft output for feather details as opposed to cameras with no AA filters (eg. d500, d850, A7 iii etc) with other factors being the same for all practical purposes.

The only time I was able to get sharp results from heavy AA filter bodies was when the light was exceptional & the ISO was low.
However, in the same situation the non-AA filter bodies would perform better.

I find it a little annoying that often pro photographers don't talk about it as much as I suspect that they often shoot with exceptionally sharp telephoto lenses which somewhat negate the AA filter effect.

Your thoughts?

 

[wild-pics]

I prefer a camera with no AA filter, ie: D500, D850, Z7, XT3 etc: mainly for general/landscape type shots as I find the image crisper. But this is entirely personal. For birds and wildlife I have not noticed a difference as most images are cropped, in fact the D750 or Z6ii with bird feathers etc: appears to control moire better than the Z7. For wildlife which is my area of interest I don't think it matters that much.

 

[RichF]

Once I get above 30MP, I don't care (much). The tiny possible reduction in resolution does not matter to me, as I seldom much an image that large or do a heavy crop and enlarge it to 16x24.

 

[NorthernFocus]

I prefer bodies without AA filters. Ironically the only one I shoot with one is the D4. I've not seen it published anywhere but I believe the AA filter is not effective (IOW not needed) above a certain pixel pitch. This based on the fact that Nikon still used an AA filter on the D800 but dropped it on the D810 and D850 but still uses it on the D750 and flagship bodies(all 24MP or less).

That said I don't see that it bothers me when editing D4 images. Standard sharpening techniques are adequate. I suspect the filter doesn't bother pros because they tend to do a lot less heavy cropping than amateurs therefore have more inherent detail to work with.

 

[sid_19911991]

I prefer a camera with no AA filter, ie: D500, D850, Z7, XT3 etc: mainly for general/landscape type shots as I find the image crisper. But this is entirely personal. For birds and wildlife I have not noticed a difference as most images are cropped, in fact the D750 or Z6ii with bird feathers etc: appears to control moire better than the Z7. For wildlife which is my area of interest I don't think it matters that much.

Strange, I feel it is the other way around! I guess preferences wary...

 

[sid_19911991]

I prefer bodies without AA filters. Ironically the only one I shoot with one is the D4. I've not seen it published anywhere but I believe the AA filter is not effective (IOW not needed) above a certain pixel pitch. This based on the fact that Nikon still used an AA filter on the D800 but dropped it on the D810 and D850 but still uses it on the D750 and flagship bodies(all 24MP or less).

That said I don't see that it bothers me when editing D4 images. Standard sharpening techniques are adequate. I suspect the filter doesn't bother pros because they tend to do a lot less heavy cropping than amateurs therefore have more inherent detail to work with.

You are right, I guess pros typically don't crop much or shoot in bad light.

 

[Tiago C]

Hi folks,

I would love your take on the subject.

I have found DSLRs/mirrorless cameras with AA filters (eg. A9 i, Z6 ii, etc.), particularly heavy ones, giving soft output for feather details as opposed to cameras with no AA filters (eg. d500, d850, A7 iii etc) with other factors being the same for all practical purposes.

The only time I was able to get sharp results from heavy AA filter bodies was when the light was exceptional & the ISO was low.
However, in the same situation the non-AA filter bodies would perform better.

I find it a little annoying that often pro photographers don't talk about it as much as I suspect that they often shoot with exceptionally sharp telephoto lenses which somewhat negate the AA filter effect.

Your thoughts?

I owned a Canon 5D MK IV (32MP with AA filter). I'm quite certain I can get (Edit: way) more detail from my current Fujifilm X-T4 despite the lower MP sensor (26MP no AA).
Edit: Perhaps I should attribute the additional detail I perceive to the higher pixel density of the X-T4 sensor?
I prefer a camera not to have AA filter. However, If I can fill the frame, MP and AA filter become secondary concerns unless I want to print.

BTW, the A7 III sensor has AA filter.

https://twitter.com/i/web/status/971819557085048832

 

[Steve]

No issues with the AA filter at all. I use the 600mm F/4, but also "lesser" lenses and have had no issues that I felt were because of the AA filter. Below are just a handful of shots with cameras like the D5/6. For the most part, any softness from an AA filter is easily sharpened away during processing. I find that often when people think an AA filter is the problem, it's usually something else. Remember, higher res cameras aren't as likely to have an AA filter BUT also capture a lot more detail. Lower res cameras can't capture that came level of detail, regardless of their optical pass filter status.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

You can only see EXIF info for this image if you are logged in.

 

[DRwyoming]

Your thoughts?

As others have said I don't really have a problem with cameras that have or do not have an AA filter but I think what you may be experiencing is cameras with high resolution vs cameras with lower resolution especially if you crop deeply. IOW, manufacturers only started removing the AA filter when sensor resolution jumped into the 20+ or 30+ MP range and lower resolution cameras traditionally had them.

If you also crop a lot you're using even fewer of those pixels on your subject. So what you might be thinking of as AA vs no AA filter might be closer to shooting lower res cameras (possibly with some deep cropping) vs shooting higher resolution cameras (with similar cropping) and forgetting the filter for a moment that alone will tend to yield less detail if you're putting fewer pixels on your subjects.

Assuming I can get close enough to my subjects to adequately fill the frame I don't find the small amount of sharpening I have to add in post to compensate for the AA filter is a big deal as that was just the deal at the outset of digital photography with pro cameras in the sub 5 MP range like my 2.74MP Nikon D1H. But if I can't get close and have to rely on cropping then yeah, high resolution cameras like the D850 or Z7 II will get me more pixels on subject and typically cameras like that will not have an AA filter either.

 

[Tiago C]

Just found this video comparing image sharpness from the Panasonic S1 vs S1H, both recent 24MP cameras, one without and one with aa filter

Thom Hogan suggested a while ago the difference in sharpness between sensors with AA filter VS no AA filter is about 15%. I also heard/read somewhere that video-centric cameras tend to have stronger AA filters. It's one of those things that is not very easy to compare (I haven't found many comparisons), especially between cameras of different types and generations.

 

[sid_19911991]

Just found this video comparing image sharpness from the Panasonic S1 vs S1H, both recent 24MP cameras, one without and one with aa filter

Thom Hogan suggested a while ago the difference in sharpness between sensors with AA filter VS no AA filter is about 15%. I also heard/read somewhere that video-centric cameras tend to have stronger AA filters. It's one of those things that is not very easy to compare (I haven't found many comparisons), especially between cameras of different types and generations.

Oh, my test/experience is totally different. I shoot birds & crop them by a significant margin. I have found AA filter bodies giving a softer output probably by 50% or so.

In the same manner, Nikon Z6 ii despite the AA filter is sharp for landscape type photos when shot with the sharp 24-70 F4 Z lens.

Wildlife especially on higher iso is a whole other ballgame.

 

[Steve]

Oh, my test/experience is totally different. I shoot birds & crop them by a significant margin. I have found AA filter bodies giving a softer output probably by 50% or so.

In the same manner, Nikon Z6 ii despite the AA filter is sharp for landscape type photos when shot with the sharp 24-70 F4 Z lens.

Wildlife especially on higher iso is a whole other ballgame.

Like anything, the heavier you crop, the more exaggerated any detrimental effect will be.

 

[Steve]

Oh, my test/experience is totally different. I shoot birds & crop them by a significant margin. I have found AA filter bodies giving a softer output probably by 50% or so.

In the same manner, Nikon Z6 ii despite the AA filter is sharp for landscape type photos when shot with the sharp 24-70 F4 Z lens.

Wildlife especially on higher iso is a whole other ballgame.

Another thought is to post a 100% crop section of a problem photo. If we can see it, we might be able to give some more advice

(BTW - I like your instagram feed)

 

[bleirer]

I know that aa filters are less necessary as the megapixels rise, since moire is a lot less likely when the resolution is high. I think you can have them removed for a couple hundred bucks, but I dont think I would risk it. With Canon dpp4 software they have the digital lens optimizer that is supposed to correct for the aa filter. I bet other makers have something similar. Or just sharpen more in Lightroom.

 

[Woodpecker]

Hi folks,

I would love your take on the subject.

I have found DSLRs/mirrorless cameras with AA filters (eg. A9 i, Z6 ii, etc.), particularly heavy ones, giving soft output for feather details as opposed to cameras with no AA filters (eg. d500, d850, A7 iii etc) with other factors being the same for all practical purposes.

The only time I was able to get sharp results from heavy AA filter bodies was when the light was exceptional & the ISO was low.
However, in the same situation the non-AA filter bodies would perform better.

I find it a little annoying that often pro photographers don't talk about it as much as I suspect that they often shoot with exceptionally sharp telephoto lenses which somewhat negate the AA filter effect.

Your thoughts?

Although I normally argue with getting IQ in the image in the first place, in this case I would look at it differently.
Remembering that these filters were originally introduced because digital cameras regularly had - and to a certain extent still have - problems with Moiree, I would say that - apart from very specific applications - it is better to use a camera without this kind of filter. This especially applies as similar to the denoising there are quite good functions implemented in post software to cope with this kind of issue. Especially when shooting in nature you rarely have scenarios that are prone to Moiree.
I was shooting bodies with and without filters side by side for a while and the loss in fine detail - despite of precise AF fine tuning etc. - was clearly visible and I ended up preferring the body without filter whenever I had the choice.

 

[sid_19911991]

Another thought is to post a 100% crop section of a problem photo. If we can see it, we might be able to give some more advice

(BTW - I like your instagram feed)

Thank you Steve. I am happy you liked some of my photos.

I will post a bird pic taken on Z6 ii & D500 soon.

 

[thelordofthelight]

I am not sure if an AA filter can cause a 50% difference in sharpness even on a cropped image. There is probably something else going on there. I have used the D850 and Z6II side by side several times and find it very hard to find such a huge difference. There certainly is some extra punch with the D850 files and i'd say the difference is just about let's say pushing that sharpness slider in LR by 5 or 10 more points with the Z6 files. But have never noticed such a huge variation as you seem to have observed in your case.

 

[arbitrage]

I’ve never been bothered with it. Really could care less and never noticed any critical difference between my cameras with or without. Of course those without have always been the pixel dense ones so hard to really compare. I have noticed some shots ruined with odd colours/moire with AA-less ones. But still no big deal.

 

[NorthernFocus]

I'm interested to know how one measures a difference in sharpness in order to assign a percent difference value. I mean I realize that Thom Hogan is way smarter than the rest of us but even the brightest minds have trouble measuring things that are subjective by their very nature.

 

[Steve]

I'm interested to know how one measures a difference in sharpness in order to assign a percent difference value. I mean I realize that Thom Hogan is way smarter than the rest of us but even the brightest minds have trouble measuring things that are subjective by their very nature.

I think he has an engineering background so maybe he came up with some consistent way to test. I agree, it would be tricky for sure.

 

[NorthernFocus]

I think he has an engineering background so maybe he came up with some consistent way to test. I agree, it would be tricky for sure.

Yeah I share the same handicap which is why I ask the question. There is software that detects/measures contrast which I suppose could be used. But I've not seen such a methodology referenced.

 

[eaj101]

Aliasing is simply a necessary artifact of any digital capture of anything, from light to sound waves (it's why your CDs use 44Khz). Knowing the frequency of the input and the sampling frequency you can calculate precisely when aliasing will occur. Without getting too technical it's what happens when in sampling an analog value falls too close to a digital boundary and could be placed to either side of it. Moire is a visual effect that results from that.

Anti-aliasing filters essentially blur/block those wavelengths most likely to fall into that frequency range. Higher resolution sensors essentially raise the sampling frequency and change the range of frequencies likely to cause aliasing. When the D800 was released with an anti-aliasing filter they also released the D800e without the filter (I bought one and loved it.)

But AA filters have to apply to a general case. So AA filters do lower the risk of moire and diminish sharpness. Only you can decide if that matters to you. (I don't take a lot of images of screen doors ). There is, however, some risk of moire in bird feather detail. Luckily most natural scenes have little detail likely to cause moire. I will note that many PC video games have settings for anti-aliasing (You trade visual image quality for higher computational overhead. If you know the resolution of the output device you can absolutely eliminate visual artifacts of aliasing, aka 'jaggies').

As usual, your mileage will vary.

 

[EricBowles]

I'm interested to know how one measures a difference in sharpness in order to assign a percent difference value. I mean I realize that Thom Hogan is way smarter than the rest of us but even the brightest minds have trouble measuring things that are subjective by their very nature.

Thom's comment was in 2009 - and in the context of discussing removing the AA filter from an 8-12 megapixel camera of that time period. I don't think 15% is a realistic assessment.

Probably a better comparison is to compare the DxoMark sharpness scores for the D800 and D800E using a given lens. For example, the Sigma 85mm f/1.4 lens is one of the sharpest lenses made for the F-mount. The Profile Measurements at f/4-5.6 gave it an Acutance % of around 84% for the D800E and 80% for the D800. So that would be a 5% difference in sharpness or resolving power solely attributed to the AA filter or lack of that filter on those two cameras.

Bottom line, it matters but does not matter very much. You probably can't visually observe a difference without a very specific test case, and the difference would mean a little more aggressive sharpening setting.

 

[NorthernFocus]

...When the D800 was released with an anti-aliasing filter they also released the D800e without the filter (I bought one and loved it.)...

Probably a better comparison is to compare the DxoMark sharpness scores for the D800 and D800E using a given lens. For example, the Sigma 85mm f/1.4 lens is one of the sharpest lenses made for the F-mount. The Profile Measurements at f/4-5.6 gave it an Acutance % of around 84% for the D800E and 80% for the D800. So that would be a 5% difference in sharpness or resolving power solely attributed to the AA filter or lack of that filter on those two cameras...

I also had a D800E. As I recall they did not remove the OLPF/AA filter but rather added a second filter layer that "cancelled the effect"(Nikon's wording). They actually added an additional layer in the optical path rather than removing one. So the DxO test might be the best example we have but is not truly a comparison of the same sensor with/without a filter. When the D7100 came out it had no filter. To date my favorite Nikon bodies for pure IQ(given adequate light) were the 71/7200. They were also the highest pixel density sensors other than some of the "consumer level" cameras.

As explained a couple of posts above there are other ways to generate aliasing(aka moire) effects in the digital world. Once or twice I've managed to resize images just right/wrong such that moire effect showed up in bird feathers displayed on my monitor.

 

[eaj101]

I also had a D800E. As I recall they did not remove the OLPF/AA filter but rather added a second filter layer that "cancelled the effect"(Nikon's wording). They actually added an additional layer in the optical path rather than removing one. So the DxO test might be the best example we have but is not truly a comparison of the same sensor with/without a filter. When the D7100 came out it had no filter. To date my favorite Nikon bodies for pure IQ(given adequate light) were the 71/7200. They were also the highest pixel density sensors other than some of the "consumer level" cameras.

As explained a couple of posts above there are other ways to generate aliasing(aka moire) effects in the digital world. Once or twice I've managed to resize images just right/wrong such that moire effect showed up in bird feathers displayed on my monitor.

Right. I forgot about that. There was a difference you could see on 16x20 prints I made (I borrowed a standard D800) but it was not big.