AA filters

[neil holmes]

I don't mind tangents in discussions like this, but I do not like it when threads leak into each other. If you'd like to continue that discussion, let's go back to that thread.

No it is ok leave both here.....I don't like it either but just as with that thread, I was disagreeing with what you wrote.

In this case correcting something for the record.

In that thread I just disagreed with a general statement you made about a particular lens.

 

[Abe R Ration]

The way I've always thought about it:

• Microcontrast: MTF at relatively fine detail (let's say, 240 lw/ph in the days of film, and perhaps double or triple that today).

I think that would be a rather obscure definition. I'm not sure if there is any "official" definition, but to quote Zeiss' H.H. Nasse:

"micro contrast, i.e. structures, which we can just about see or just cannot see with the naked eye,"

(http://kurtmunger.com/sitebuildercontent/sitebuilderfiles/zeissmtfcurves1.pdf)

Thus I think my use of it is in agreement with this.

 

[JanePete]

At lest that color can be addressed. Resolution preserved.

If you look closely, you'll see that the resolution is not actually preserved. The lines continue to narrow to the point where 36MP does not resolve. You continue to see lines. That's not actual detail. That's false detail. That's a form of aliasing.

So you notice the tiny details of blue but fail to see the resolution loss from the AA filtered camera?

I do see it. Just not at any sane viewing sizes with a 36MP camera. So I get the same resolution as a 24MP camera. Big whoop.

Looks obvious to me the resolution on the 810 is superior. Pretty strange that you area hoping focusing errors would wipe that out.

Look at a corner.

DPReview studio shot from D800 at the corner. Notice how nice and sharp it is.

Same studio shot from D810 in the corner. Blurry.

If dpreview cannot get focus good enough in a test studio to where the D810 wins across an image, how likely are you to do so in the real world?

Are you kidding? There are a number of reasons why the D810 can look blurrier than the D800 in that single test.

Unfortunately not. In many cases, it's impossible to correct moire -- certainly most of the cases I've seen. Actually, that's a pretty good challenge for you. Take the DPReview test image, and try to clean up all the moire all over the place.

In most cases you don,t have to since it is rare to begin with and usually affects on small parts of images.

Aliasing relics are all over your images. They manifest in different ways. Moire is one of the ways, and happens if the pattern happens to be close to the sensor resolution. If the pattern is random, it will translate to higher image noise. If it's regular with a different spatial frequency, it will just look different.

I'm sure most people would want to have the image on the right, without an AA filter, than the mess on the left, with an AA filter.

In contrast, fixing up AA filter is a simple deconvolution in most cases. Doesn't work well for high ISO, but at low ISO, you can correct for AA filter nearly perfectly.

How do you bring back resolution that was lost?

In most cases, very little actual resolution is lost. Sharpness is lost. A tool like DxO, which does deconvolution calibrated to the AA filter, ought to be able to bring almost all of it back.

In the example above the image on the right without the AA has higher the resolution. Obviously if an image is blurred enough then there will be a loss of resolution.

Nope, it was never solved. AA filters do not solve the issue of moire, they only reduce the appearance of it in exchange for the entire image being blurred.

(Last quote one taken from a different post, but also from Jane): A good AA filter will eliminate almost all moire,

Feel free to show examples that do not significantly reduce resolution.

and only blur the image slightly, and in such a way that it has no effect if the image is downsampled even slightly

Why would someone want to to downsize an image to try and overcome blurring? You buy something in the D800 class, and any other camera without an AA filter, to take advantage of all the resolution it can muster.

(in contrast to lens resolution limits, which translate to worse contrast). It doesn't just reduce the appearance -- it actually gets rid of it. In theory, as technology improves, an optical lowpass filter could get arbitrarily good at reducing aliasing without effecting the image. In practice, designing such a filter would be very hard.

The trend is to get away from the blur filter, and that's a good thing.

 

[Abe R Ration]

Alphoid wroteLast quote one taken from a different post, but also from Jane): A good AA filter will eliminate almost all moire,

Feel free to show examples that do not significantly reduce resolution.

If I may step in... while AA filter reduces resolution (very significantly so if all aliasing were to be eliminated on Bayer CFA camera), with proper processing it can be largely recovered (though with increase of noise) with deconvolution as the point spred function of the filter can be very well know.

The below link is about diffraction blur, but the same principle applies.

http://aberration43mm.wordpress.com/2014/11/07/some-resolution-test-images/

 

[JanePete]

Thanks for your input, Basalite/bford/jane pete, whoever you are this month.

I'm bored silly with the subject, which was very old news two years ago. Technology has moved on, the AA filter is gone forever for serious high res work, and good riddance to it. Not a man jack of you can show a better picture from any camera priced lower than the D810, certainly not the laughable plastic Sigma, but you sure love to talk, talk talk. Have fun with it, I've got money to make.

Just pointing out your hypocrisy and dishonesty where you value using RAWs to make comparisons in one situation, the one that benefits you and your choice of camera, and you don't value it when making comparisons to Sigma's cameras, a camera manufacturer you love to dishonestly bash. Then you all of a sudden use JPGs to disadvantage the Sigmas.

Their cameras are also no less metal than another others on the market. If you had actually used one and/or didn't have a silly bias against Sigma's cameras then you would know that.

 

[JanePete]

Preferring a slightly blurred pic is not a crime to my knowledge, but I'll take my raws straight.

You're not getting them straight. You're getting them with a bunch of fake detail which isn't there added.

[ATTACH alt="Is being able to see and read the small text in the blur filter sample on the right "fake detail?" "]812729[/ATTACH]
Is being able to see and read the small text in the blur filter sample on the right "fake detail?"

 

[JanePete]

There's a trend towards cameras without AA filters. Can anyone explain to me why this is a good idea?

This thread makes it obvious why this is a good idea. It has nothing to do with DxOMark or other benchmarks as my original post suggested. It appears that:

1. Pixel level detail is very obvious in 100% crop review photos.
2. Aliasing is a nuanced and complex issue. It requires advanced mathematics to understand. It's clear that the majority of people don't understand what it is, where it comes from, or how it effects their images.

Just requires some decent vision.

Optimizing for pixel-level detail is clearly the winning strategy to sell more cameras.

Why wouldn't you optimize for that? Pixels make up the picture.

It results in slightly worse photos, but not enough to tarnish brand names.

Most photographers would disagree.

On the other hand, explaining the benefits of not having aliasing relics in images is not something you can reasonably do for typical consumers, or even most photographers.

That's because they are not visible in most cases.

If anyone is interested:

1. If you don't have an AA filter, aliasing relics will show up in most of your images. In most cases, the effect will not be as obvious as moire, which only effects a tiny portion of images. It will range from more noise, to fake detail, to incorrect colors, to many other visual artifacts.

You are seeing far, far, far more artifacts than most other photographers.

1. The effect of AA filters is fundamentally different from anything that can be accomplished in postprocessing. Most aliasing relics cannot be removed in postprocessing. On the other hand, if the effect is that the image looks different (e.g. texture where there shouldn't be texture, or similar), but it's not obvious, you might not care.

And you can't add detail that has been lost to the blur filter.

1. AA filters do not effect contrast. They only effect resolution. Having an AA filter is like shooting with e.g. a 24MP instead of a 36MP one, not like having an unsharp lens.

Most people would rather use all those pixels though.

By the way, that's affect in both cases. The effect is the lower resolution.

Didn't you also say previously that resolution is not affected, only sharpness?

1. While AA filters are not perfect, good ones will eliminate almost all aliasing relics.

While ruining detail overall.

 

[Mark S Abeln]

Unfortunately with Bayer sensors, antialias filters have to be strong enough to prevent color artifacts on the red and blue color channels but this excessively harms the green channels which have twice the resolution and much of the luminance information.

 

[gaussian blur]

There's a trend towards cameras without AA filters. Can anyone explain to me why this is a good idea?

Without an AA filter, if you shoot with a sharp lens, you get moire. Moire is visible at any viewing size, and ugly. Artifacts like this creep into your photos (scroll to the bottom):

Moire affecting most scenes is rare, and typically affects small areas,

It affects any area where there's high spatial content at or near Nyquist and it can't be removed without adversely affecting real detail.

Moire affecting most scenes is rare, and typically affects small areas.

Whether it's rare or not depends on what type of photography someone does. Some people might never see it and others will frequently see it, depending on many factors.

Rare for most people and most scenes.

You can't make any generalizations, especially without referring to a particular camera. Whether it's rare or not depends on many factors.

whereas the use of an AA filter degrades the entire image of every image. There is no reasonable purpose for using AA filters. Even the moire reasoning doesn't hold up since they do not eliminate moire entirely.

Completely wrong.

AA filters bandlimit high frequencies that are beyond the limits of the sensor which can't be accurately resolved anyway (and this isn't only about cameras either). Not doing that will result in alias artifacts, or false details not originally in the subject. Moire patterns is just one example.

Read a book on signal theory sometime.

Look up the difference between theory and practice sometime. In every digital photographic case where the AA filters are ditched resolution of the entire image is noticeably increased.

No it isn't. What's increased are alias artifacts, or false detail that wasn't actually there.

What you're claiming is mathematically impossible and no amount of your arguing is going to change math & physics.

Something like all the new Nikons omitting AA filters is bad for most users.

Why? Clearly the entire image of every image is benefitting. For someone that is fishing for the usually tiny effects of moire I find it ironic then that you miss the effect of a blurred image on the entire image for every image.

The benefit comes at a price of alias artifacts. That's why.

What artifacts would those be? You mean actually being able to clearly see the pixels that make up the image? That's a good thing.

Nope. You're twisting things again. If you actually read what I wrote, you'd see that I said what kind of artifacts, and seeing individual pixels is a bad thing, which is why high DPI displays where pixels are too small to be seen are becoming common.

Nyquist of a 36 mp sensor is high enough so that an AA filter is not as important as with a lower resolution sensor, however, there can still be situations where aliasing can be an issue. There's no getting around that. There is no perfect AA filter and tradeoffs must be made.

I never mentioned anything about a "perfect AA filter."

Comments like that is further proof you don't understand anything about signal theory or aliasing.

As they are implemented, they serve no reasonable purpose. In practice with digital cameras they clearly reduce resolution. there's no getting that.

AA filters do not reduce resolution. They bandlimit what can't be resolved, which is a requirement for a properly implemented sampling system. Otherwise you get aliasing artifacts.

Since no perfect AA filter exists, there are tradeoffs to be made. It's just the way it is. We live in an imperfect world.

 

[Alphoid]

That's from a company whose product is stripping AA filters. I like how you posted it without attribution. If you believe that, I've got a few things to sell you. I assume you believe weight loss and skin cream before/after pictures too?

Look at the DPReview samples if you'd like to see what actually happens. The D800/D810 you posted is a good example.

 

[JanePete]

There's a trend towards cameras without AA filters. Can anyone explain to me why this is a good idea?

Without an AA filter, if you shoot with a sharp lens, you get moire. Moire is visible at any viewing size, and ugly. Artifacts like this creep into your photos (scroll to the bottom):

Moire affecting most scenes is rare, and typically affects small areas,

It affects any area where there's high spatial content at or near Nyquist and it can't be removed without adversely affecting real detail.

Moire affecting most scenes is rare, and typically affects small areas.

Whether it's rare or not depends on what type of photography someone does. Some people might never see it and others will frequently see it, depending on many factors.

Rare for most people and most scenes.

You can't make any generalizations, especially without referring to a particular camera. Whether it's rare or not depends on many factors.

I can, and I have. One only has to look at the photos of cameras without blur filters.

whereas the use of an AA filter degrades the entire image of every image. There is no reasonable purpose for using AA filters. Even the moire reasoning doesn't hold up since they do not eliminate moire entirely.

Completely wrong.

AA filters bandlimit high frequencies that are beyond the limits of the sensor which can't be accurately resolved anyway (and this isn't only about cameras either). Not doing that will result in alias artifacts, or false details not originally in the subject. Moire patterns is just one example.

Read a book on signal theory sometime.

Look up the difference between theory and practice sometime. In every digital photographic case where the AA filters are ditched resolution of the entire image is noticeably increased.

No it isn't. What's increased are alias artifacts, or false detail that wasn't actually there.

What you're claiming is mathematically impossible and no amount of your arguing is going to change math & physics.

Right, so the manufacturers like Leica are simply fooling photographers and what photograPhers are seeing is a figment of their imagination.

Something like all the new Nikons omitting AA filters is bad for most users.

Why? Clearly the entire image of every image is benefitting. For someone that is fishing for the usually tiny effects of moire I find it ironic then that you miss the effect of a blurred image on the entire image for every image.

The benefit comes at a price of alias artifacts. That's why.

What artifacts would those be? You mean actually being able to clearly see the pixels that make up the image? That's a good thing.

Nope. You're twisting things again. If you actually read what I wrote, you'd see that I said what kind of artifacts, and seeing individual pixels is a bad thing, which is why high DPI displays where pixels are too small to be seen are becoming common.

Well, there is obviously a flip side to that that you have deemed as a negative and yet affects the entire image.

We don't see individual pixels in most print sizes and yet they affect image quality.

Nyquist of a 36 mp sensor is high enough so that an AA filter is not as important as with a lower resolution sensor, however, there can still be situations where aliasing can be an issue. There's no getting around that. There is no perfect AA filter and tradeoffs must be made.

I never mentioned anything about a "perfect AA filter."

Comments like that is further proof you don't understand anything about signal theory or aliasing.

I understand what my eyes can see. Perhaps you should tell Leica, and other camera manufacturer, they don't understand either.

As they are implemented, they serve no reasonable purpose. In practice with digital cameras they clearly reduce resolution. there's no getting that.

AA filters do not reduce resolution. They bandlimit what can't be resolved, which is a requirement for a properly implemented sampling system. Otherwise you get aliasing artifacts.

Since no perfect AA filter exists, there are tradeoffs to be made. It's just the way it is. We live in an imperfect world.

First you say "AA filters do not reduce resolution" and then you say "no perfect AA filter exists, there are tradeoffs to be made?"

 

[sigala1]

The internet meme is that image quality is determined by only three factors:

1. Sharpness
2. How much "bokeh" there is in the out-of-focus areas
3. How little noise there is at high ISO.

No one ever talks about any other technical aspects of image quality. The three things above are the only things they know because it's the only thing anyone ever talks about on the internet.

Removing the AA filter, thus, appears to increase image quality based on #1 above, while not having any obvious negative efect on #2 and #3.

(Additionally cameras are judged on useability, but the only useability anyone cares about is whether the camera can autofocus on birds in flight.)

 

[Dukhat]

There's a trend towards cameras without AA filters. Can anyone explain to me why this is a good idea?

This thread makes it obvious why this is a good idea. It has nothing to do with DxOMark or other benchmarks as my original post suggested. It appears that:

1. Pixel level detail is very obvious in 100% crop review photos.
2. Aliasing is a nuanced and complex issue. It requires advanced mathematics to understand. It's clear that the majority of people don't understand what it is, where it comes from, or how it effects their images.

Just requires some decent vision.

Optimizing for pixel-level detail is clearly the winning strategy to sell more cameras.

Why wouldn't you optimize for that? Pixels make up the picture.

It results in slightly worse photos, but not enough to tarnish brand names.

Most photographers would disagree.

On the other hand, explaining the benefits of not having aliasing relics in images is not something you can reasonably do for typical consumers, or even most photographers.

That's because they are not visible in most cases.

If anyone is interested:

1. If you don't have an AA filter, aliasing relics will show up in most of your images. In most cases, the effect will not be as obvious as moire, which only effects a tiny portion of images. It will range from more noise, to fake detail, to incorrect colors, to many other visual artifacts.

You are seeing far, far, far more artifacts than most other photographers.

1. The effect of AA filters is fundamentally different from anything that can be accomplished in postprocessing. Most aliasing relics cannot be removed in postprocessing. On the other hand, if the effect is that the image looks different (e.g. texture where there shouldn't be texture, or similar), but it's not obvious, you might not care.

And you can't add detail that has been lost to the blur filter.

1. AA filters do not effect contrast. They only effect resolution. Having an AA filter is like shooting with e.g. a 24MP instead of a 36MP one, not like having an unsharp lens.

Most people would rather use all those pixels though.

By the way, that's affect in both cases. The effect is the lower resolution.

Didn't you also say previously that resolution is not affected, only sharpness?

1. While AA filters are not perfect, good ones will eliminate almost all aliasing relics.

While ruining detail overall.

I didn't realize that we were all taking detail-less train wrecks of photos until just a couple of years ago. It's amazing what cam be learned on the internet.

 

[JanePete]

That's from a company whose product is stripping AA filters. I like how you posted it without attribution. If you believe that, I've got a few things to sell you. I assume you believe weight loss and skin cream before/after pictures too?

If you don't clarify, or better yet, use the quote function, then I have no idea what you are referring to.

Look at the DPReview samples if you'd like to see what actually happens. The D800/D810 you posted is a good example.

I'm not the one that posted the d800/810 sample.

 

[JanePete]

There's a trend towards cameras without AA filters. Can anyone explain to me why this is a good idea?

This thread makes it obvious why this is a good idea. It has nothing to do with DxOMark or other benchmarks as my original post suggested. It appears that:

1. Pixel level detail is very obvious in 100% crop review photos.
2. Aliasing is a nuanced and complex issue. It requires advanced mathematics to understand. It's clear that the majority of people don't understand what it is, where it comes from, or how it effects their images.

Just requires some decent vision.

Optimizing for pixel-level detail is clearly the winning strategy to sell more cameras.

Why wouldn't you optimize for that? Pixels make up the picture.

It results in slightly worse photos, but not enough to tarnish brand names.

Most photographers would disagree.

On the other hand, explaining the benefits of not having aliasing relics in images is not something you can reasonably do for typical consumers, or even most photographers.

That's because they are not visible in most cases.

If anyone is interested:

1. If you don't have an AA filter, aliasing relics will show up in most of your images. In most cases, the effect will not be as obvious as moire, which only effects a tiny portion of images. It will range from more noise, to fake detail, to incorrect colors, to many other visual artifacts.

You are seeing far, far, far more artifacts than most other photographers.

1. The effect of AA filters is fundamentally different from anything that can be accomplished in postprocessing. Most aliasing relics cannot be removed in postprocessing. On the other hand, if the effect is that the image looks different (e.g. texture where there shouldn't be texture, or similar), but it's not obvious, you might not care.

And you can't add detail that has been lost to the blur filter.

1. AA filters do not effect contrast. They only effect resolution. Having an AA filter is like shooting with e.g. a 24MP instead of a 36MP one, not like having an unsharp lens.

Most people would rather use all those pixels though.

By the way, that's affect in both cases. The effect is the lower resolution.

Didn't you also say previously that resolution is not affected, only sharpness?

1. While AA filters are not perfect, good ones will eliminate almost all aliasing relics.

While ruining detail overall.

I didn't realize that we were all taking detail-less train wrecks of photos until just a couple of years ago. It's amazing what cam be learned on the internet.

We? Clearly image resolution is not a concern for *you.*

 

[Dukhat]

There's a trend towards cameras without AA filters. Can anyone explain to me why this is a good idea?

This thread makes it obvious why this is a good idea. It has nothing to do with DxOMark or other benchmarks as my original post suggested. It appears that:

1. Pixel level detail is very obvious in 100% crop review photos.
2. Aliasing is a nuanced and complex issue. It requires advanced mathematics to understand. It's clear that the majority of people don't understand what it is, where it comes from, or how it effects their images.

Just requires some decent vision.

Optimizing for pixel-level detail is clearly the winning strategy to sell more cameras.

Why wouldn't you optimize for that? Pixels make up the picture.

It results in slightly worse photos, but not enough to tarnish brand names.

Most photographers would disagree.

On the other hand, explaining the benefits of not having aliasing relics in images is not something you can reasonably do for typical consumers, or even most photographers.

That's because they are not visible in most cases.

If anyone is interested:

1. If you don't have an AA filter, aliasing relics will show up in most of your images. In most cases, the effect will not be as obvious as moire, which only effects a tiny portion of images. It will range from more noise, to fake detail, to incorrect colors, to many other visual artifacts.

You are seeing far, far, far more artifacts than most other photographers.

1. The effect of AA filters is fundamentally different from anything that can be accomplished in postprocessing. Most aliasing relics cannot be removed in postprocessing. On the other hand, if the effect is that the image looks different (e.g. texture where there shouldn't be texture, or similar), but it's not obvious, you might not care.

And you can't add detail that has been lost to the blur filter.

1. AA filters do not effect contrast. They only effect resolution. Having an AA filter is like shooting with e.g. a 24MP instead of a 36MP one, not like having an unsharp lens.

Most people would rather use all those pixels though.

By the way, that's affect in both cases. The effect is the lower resolution.

Didn't you also say previously that resolution is not affected, only sharpness?

1. While AA filters are not perfect, good ones will eliminate almost all aliasing relics.

While ruining detail overall.

I didn't realize that we were all taking detail-less train wrecks of photos until just a couple of years ago. It's amazing what cam be learned on the internet.

We? Clearly image resolution is not a concern for *you.*

It's amazing how many people can prove themselves to be dumb as bricks.

Thank you for showing that you are just one of the many.

 

[JanePete]

There's a trend towards cameras without AA filters. Can anyone explain to me why this is a good idea?

This thread makes it obvious why this is a good idea. It has nothing to do with DxOMark or other benchmarks as my original post suggested. It appears that:

1. Pixel level detail is very obvious in 100% crop review photos.
2. Aliasing is a nuanced and complex issue. It requires advanced mathematics to understand. It's clear that the majority of people don't understand what it is, where it comes from, or how it effects their images.

Just requires some decent vision.

Optimizing for pixel-level detail is clearly the winning strategy to sell more cameras.

Why wouldn't you optimize for that? Pixels make up the picture.

It results in slightly worse photos, but not enough to tarnish brand names.

Most photographers would disagree.

On the other hand, explaining the benefits of not having aliasing relics in images is not something you can reasonably do for typical consumers, or even most photographers.

That's because they are not visible in most cases.

If anyone is interested:

1. If you don't have an AA filter, aliasing relics will show up in most of your images. In most cases, the effect will not be as obvious as moire, which only effects a tiny portion of images. It will range from more noise, to fake detail, to incorrect colors, to many other visual artifacts.

You are seeing far, far, far more artifacts than most other photographers.

1. The effect of AA filters is fundamentally different from anything that can be accomplished in postprocessing. Most aliasing relics cannot be removed in postprocessing. On the other hand, if the effect is that the image looks different (e.g. texture where there shouldn't be texture, or similar), but it's not obvious, you might not care.

And you can't add detail that has been lost to the blur filter.

1. AA filters do not effect contrast. They only effect resolution. Having an AA filter is like shooting with e.g. a 24MP instead of a 36MP one, not like having an unsharp lens.

Most people would rather use all those pixels though.

By the way, that's affect in both cases. The effect is the lower resolution.

Didn't you also say previously that resolution is not affected, only sharpness?

1. While AA filters are not perfect, good ones will eliminate almost all aliasing relics.

While ruining detail overall.

I didn't realize that we were all taking detail-less train wrecks of photos until just a couple of years ago. It's amazing what cam be learned on the internet.

We? Clearly image resolution is not a concern for *you.*

It's amazing how many people can prove themselves to be dumb as bricks.

Thank you for showing that you are just one of the many.

What's "dumb as bricks," respectfully explaining why you disagree with someone or simply calling them "dumb as bricks?"

 

[Reilly Diefenbach]

Most of you artifact whiners are looking at monitors that simply aren't up to the job of decently representing 36 or 24MP. 720p laptops or smartphones/tablets for the great majority of you. Take some full size D810 jpgs down to the store and plug a USB drive into a 75" UHDTV and wonder where the "artifacts" went.




A7r

 

[Alphoid]

JanePete wrote:
Why wouldn't you optimize for that? Pixels make up the picture.

Because image-level qualities -- such as accurate color -- are much more important than pixel-level features. Aliasing effects image-level features. An AA filter only negatively effects the image when viewed at zoom. The blur of an AA filter will have zero effect if the image e.g. fills a 1080p monitor, or for something like your 36MP sensor, a 4k monitor.

It results in slightly worse photos, but not enough to tarnish brand names.

Most photographers would disagree.

That was exactly my point.

1. AA filters do not effect contrast. They only effect resolution. Having an AA filter is like shooting with e.g. a 24MP instead of a 36MP one, not like having an unsharp lens.

By the way, that's affect in both cases. The effect is the lower resolution.

No. It's not. An unsharp lens will, in most cases, effect microcontrast visible even at normal viewing sizes. Resolution will go down, but when viewed at e.g. 1080p, the details will be a little bit less sharp as well. An antialiasing filter will only effect detail visible at or near 100% zoom. Viewed at 1080p, it will have no effect. An unsharp lens has a complex blur which reduces contrast even at normal viewing sizes.

Try this in photoshop. Take 36MP image. Open it in a second one. Zoom in to 400%. Run a 2 pixel box blur over it in one window. In a third window, zoom in to 400%, and run a Gaussian blur over it until the effect is similar. Now downsample the image to 50% or less using Lanczos downsampling. See the difference? Not all blurs have the same effect. Once you zoom out beyond 50%, a 2 pixel box blur has no effect on the image (assuming Photoshop does the math correctly). A Gaussian blur reduces overall contrast.

That's the same as the difference between lens blur and AA blur.

Unless you're zoomed in to 100%, an AA filter has no effect on the overall image. An unsharp lens does.

 

[Alphoid]

The image you're posting is marketing drivel from a company which removes AA filters. There's a D300 which is a bit out-of-focus, and has a misprocessed image on one side. There's a D300 without AA filter properly focused, and properly processed. It's snake oil.

Which wins?

The proper focus.