Optical filter or no optical filter (Page 3): I too appreciate your explanation. While I made a lucky guess and chose the 5Ds, it would have been nice to have had your reassuring...

Ugly Hedgehog - Photography Forum

Home Active Topics Newest Pictures Search Login Register ☰

Main Photography Discussion

Optical filter or no optical filter.

Page <<first <prev 3 of 7 next> last>>

Oct 8, 2018 09:59:22 #

dkguill Loc: Elkhart, IN

aflundi wrote:

Leaving off the AA filter is bad engineering. Every engineer (and scientist, and mathematician) with a background in sub-sampling systems (like a bayer filter sensor) knows that. Too many look at the pixel details seeing higher acutance without the AA filter and think that's a good thing. They should be zooming out to where the pixel patterns are clearer and evaluate the image there. In my experience, the image actually looks sharper and better with an AA filter.

A simple example is computer text or graphics. If you look at the pixels of an aliased text string or line, you'll see higher acutance than with anti-aliased text or lines. However, the anti-aliased text and graphics look *much* better when you back away from the pixel details. The faster computers today spend significantly more CPU cycles rendering text and graphics with anti-aliasing, and that's primarily why text and graphics today looks vastly superior to the older non-anti-aliased renderings. Camera images are the same.

The idea that it can be left off is based on the idea that sensors with high pixel counts can out-resolve the lens, and therefore the lens can be used as the AA filter rather than rely on dual birefringment layers at the sensor. My experience is that even lenses know for their softness are sharp enough to produce aliasing artifacts on high density sensors.

You can use diffraction as an AA filter, but it requires the aperture to be at about f/29 to make aliasing artifacts go away. That may work for very static photographs like landscapes or stills.

Many say they use their D800E/D810/D850/Z7/5DSR/etc. with little problem. I can't for the life of me figure that out, as my experience with the D810 and D500 leave me with moire, false colors, and false details way, way, way too often. I can only guess that I'm not as shaky as they are, since motion blur also works as an effective AA filter.

Personally, rather than purchase the D850 without AA filter, I choose to buy a used D800 with AA filter, and that is now my most used camera since after 4 years of dealing with my D810's artifacts, I decided that the presence of the AA filter is more important than the other awesomely cool things in the D810 and D850. Almost exclusively based on the AA filter, I also won't be purchasing a Z7 (without AA), but do expect to get a Z6 (with AA).

Leaving off the AA filter is bad engineering. Eve... (show quote)

I too appreciate your explanation. While I made a lucky guess and chose the 5Ds, it would have been nice to have had your reassuring explanation when I made my guess.

Thanks...

Oct 8, 2018 10:15:04 #

billnikon Loc: Pennsylvania/Ohio/Florida/Maui/Oregon/Vermont

camerapapi wrote:

This could be a simple discussion so I hope and I ... (show quote)

I believe the camera companies have had the last say on the matter. Most if not all current professional digital camera's have removed the AA filter. Enough said.

Oct 8, 2018 10:32:18 #

camerapapi Loc: Miami, Fl.

billnikon wrote:

I believe the camera companies have had the last say on the matter. Most if not all current professional digital camera's have removed the AA filter. Enough said.

While I cannot discuss your claim I am sure you will understand that there are more cameras in use with the AA filter than cameras without it. As I said, I am not experienced enough to discuss which is better but I do not hear many complaints from those who use cameras with the filter.
I began this post out of curiosity because I have never had an issue with sharpness using my cameras and applying a little bit of sharpness during editing.

Nikon already has my request to give me a reasonable explanation about all this and I expect them to tell me that if a camera without the filter is better why they keep on manufacturing cameras without it.

Check out People Photography section of our forum.

Oct 8, 2018 11:07:09 #

aflundi Loc: Albuquerque, NM

camerapapi wrote:

... I have talked to several photographers that own cameras like the D800e and they seem to be very happy with their cameras.

Most photographers don't know what to look for, and even mistake false detail from aliasing as real detail. (The higher acutance, for instance, is in fact an aliasing artifact.) For many kinds of photography, the aliasing artifacts stay at the pixel level and are not particularly noticeable when the image is viewed as a whole rather than peeping at the pixels. Some types of photography, like sports and wildlife, are dynamic enough that it's virtually impossible not to have enough motion blur to act as a very good AA filter.

Quote:

... I remember the Nikon D100. It was an excellent tool EXCEPT that the AA filter was too strong. The files required more sharpening and many photographers soon became disappointed in the camera. When I bought my D70s, that had a much thinner filter, I could immediately see an improvement. Now, with the D70s a small amount of sharpening was all those files needed. Modern AA filters are also very thin and require a very small amount of sharpening. ...

By virtue of the fact that you have a sub-sample system with Bayer filters means you'll always need to sharpen in post whether you have an AA filter or not in order to recover from the sub-sampling. However, if you do have an AA filter, the sharpening can work better because the information is more smoothly distributed between neighboring pixels. Without an AA filter, much of the information is irretrievably lost and will necessary yield blotchy and incorrect sharpening.

Dual-layer, birefringent AA filters are very efficient, but they're also intrinsically always too weak. That's why you can get moire even with an AA filter, thick or thin. That moire though is weaker relative to the real signal than if you remove the AA filter, so you're better off. If you can get moire or other artifacts, the AA filter is not strong enough. Manufacturers would need to go to 4-layer, birefringent AA filters to be as strong as they should be, but I don't know of any willing to do that. 2-layer AA plus lens softness plus diffraction plus motion blur is often enough to remove aliasing artifacts, but the manufacturers have been producing lenses that are just awesomely sharp, so aliasing artifacts are more common than perhaps expected.

With lower pixel counts, the AA blurring represents a larger fraction of the total image, so while it is essential for those sensors to have AA filters, one could easily develop a distaste for them. With higher pixel counts, the perceived loss of resolution from using an AA filter becomes becomes less significant as the blur is a smaller fraction of the total image, but the artifacts can be, in the case of moire, by just as destructive as with lower pixel counts. That's why despite popular trends, it really a bad idea to leave the AA filter off high pixel count sensors.

Oct 8, 2018 11:13:18 #

delkeener Loc: SW Rhode Island, USA

The OP is not discussing a lens filter. The AA filter is on the sensor.

mcmama wrote:

I only have filters for protection. I never have them on when I shoot.

Oct 8, 2018 11:23:38 #

aflundi Loc: Albuquerque, NM

billnikon wrote:

I believe the camera companies have had the last say on the matter. Most if not all current professional digital camera's have removed the AA filter. Enough said.

I agree with you in the sense that consumers can only choose from the list of products available.

However, rest assured that it wasn't the engineers that made the decision to remove AA filters. They know better.

The marketing department gave in because they can't be successful as a company if they don't sell enough. In this case, a very large fraction of consumers have dropped the ball by not understanding the issue, yet being convinced they do. Textbook Dunning-Kruger effect. It happens.

Fortunately, the mistake of removing AA filters isn't so fatal as to never yield good images. Unfortunately, the mistake of removing AA filters isn't so fatal as to never yield good images, so it's a mistake that's very hard to correct.

Oct 8, 2018 12:01:40 #

aflundi Loc: Albuquerque, NM

For anyone that doesn't understand why AA filters are necessary, here's a simple thought exercise:

Imagine a small white detail that is a small part of the scene you want to capture. The light from that white detail goes through the lens and images onto the sensor. If that light lands on a blue photosite, it'll be blue in the final image. If it lands on a red photosite, the image will record it as a red detail. There's no way for the sensor or anything in the process train behind the sensor to know that it's actually white.

I've been told by vocal advocates of AA removal that the camera can do some magic that figures out that it is actually white. *Yes*, that magic is called an AA filter!

With an AA filter, the light from the white detail is split just in front of the sensor by the AA filter so that four photosites will receive the light, one blue, one red, and two green which means that all the information needed to reconstruct that detail to have the color white is there.

If you are photographing stars at night, without an AA filter you'd get red, blue and green stars, and no white stars, for this very reason, EXCEPT that the atmosphere is turbulent and is constantly bending the light paths so that atmosphere is acting as a strong AA filter which shifts the light from any one star over multiple photosites and thus recording the information needed to accurately determine the star's color.

Check out The Dynamics of Photographic Lighting section of our forum.

Oct 8, 2018 12:05:39 #

Apaflo Loc: Anchorage, Alaska

aflundi wrote:

Leaving off the AA filter is bad engineering. Every engineer (and scientist, and mathematician) with a background in sub-sampling systems (like a bayer filter sensor) knows that.
...

Absolutely true!

There is also good evidence that Nikon's engineers are aware of that, and the decision to produce so many models without the AA filter is strictly a marketing decision.

Remember when the D800 and D800E were first announced, and the production numbers for the D800 were to be in the tens of thousands while the D800E numbers were miniscule? That is what engineering assumed would be the much lower demand for the technically less able D800E that did not have an AA filter. But to their astonishment customers did not understand, and the high demand was for the D800E.

Customers read how the AA filter blurred the image! That was all they understood, and they didn't want the filter.

Corporate Nikon sells cameras. They shifted immediately to what customers wanted. The non-AA cameras are "better" in that the demand is higher.

And of course here we are years later with millions of web reviews and technical reports touting the advantages of no AA filter. Yet in theory, as studied by every engineer, and demonstrated by testing on every sampling system of any kind, the AA filter is better!

Oct 8, 2018 12:19:58 #

Architect1776 Loc: In my mind

billnikon wrote:

I believe the camera companies have had the last say on the matter. Most if not all current professional digital camera's have removed the AA filter. Enough said.

I thought the d5 and 1dx MII had the filters.

Oct 8, 2018 12:28:57 #

Silverman Loc: Michigan

camerapapi wrote:

This could be a simple discussion so I hope and I ... (show quote)

I have the Nikon D3300 with No AA filter, my images seem to look very sharp, especially with the 35mm or 50mm 1.8g lens

Oct 8, 2018 13:04:47 #

Rich1939 Loc: Pike County Penna.

I am under the impression that high density sensors are one of the deciding factors that manufacturers consider when taking a decision to not use an AA filter.
The denser pixel arrangements of a full frame 36MP, or higher, sensor increases the number of “steps” in an angled line while at the same time reducing their count, effectively creating an AA filter affect.
With a crop sensor, manufactures can eliminate the AA filter for sensors which are smaller numerically because in combination with the smaller physical size that also results in a dense pixel arrangement.
Empirically speaking, it works. There are millions of happy photographers, amateur and professional, that swear by their filter-less cameras.
Maybe technically speaking that shouldn’t be, but then technically speaking bumble bees shouldn’t be able to fly.

Check out Printers and Color Printing Forum section of our forum.

Oct 8, 2018 13:28:46 #

billnikon Loc: Pennsylvania/Ohio/Florida/Maui/Oregon/Vermont

Architect1776 wrote:

I thought the d5 and 1dx MII had the filters.

Correct on the D5, don't know about the Canon.

Oct 8, 2018 13:28:55 #

chrisg-optical Loc: New York, NY

This article explains the difference D800/D800e - they both have AA filters but the 800e is "tuned" a bit differently using polarizers in the filter sandwich to reduce the blurring effect of the AA filter. Only the D810 and D850 dispense with the AA filter completely.

https://www.bhphotovideo.com/explora/photography/hands-review/discerning-differences-between-nikon-d800-and-d800e?BI=572&gclid=Cj0KCQjwgOzdBRDlARIsAJ6_HNkQnJy7Dz4I0o12xJRGBeaTwKf3MGQUnHXb3oP_79nosE2B-tzpnS4aAtsKEALw_wcB

Oct 8, 2018 13:41:53 #

aflundi Loc: Albuquerque, NM

Rich1939 wrote:

...Empirically speaking, it works.

It works kind-of. It produces something some people are OK with, but the artifacts are there. Sometimes the artifacts are very bad. People with poor technique are more likely to find AA-less sensors acceptable since focus-problems, motion-blur, etc. lessen or eliminate aliasing artifacts by doing what the AA filter would have done.

However, there's no reason for the artifacts, and the supposed advantage of removing the AA filter is an illusion.

Quote:

There are millions of happy photographers, amateur and professional, that swear by their filter-less cameras.

This has already been covered. Go back and read the parts of the thread you skipped. Look up Dunning-Kruger Effect.

Quote:

Maybe technically speaking that shouldn’t be, but then technically speaking bumble bees shouldn’t be able to fly.

Saying bumble bees shouldn't be able to fly is like saying AA filters aren't important. People make up all kinds of stuff.

Bumble bees *do* fly, and sensors without AA filters *do* have artifacts.

Oct 8, 2018 13:52:46 #

aflundi Loc: Albuquerque, NM

chrisg-optical wrote:

... D800/D800e - they both have AA filters but the 800e is "tuned" a bit differently using polarizers in the filter sandwich to reduce the blurring effect of the AA filter. ...

That's not quite right. The D800 and D800E both have dual birefringent layers, but the D800E does *NOT* have an AA filter. The birefringent layers are arranged on the D800E in a way where the 2nd undoes exactly what the 1st does and thereby turns the two layers into nothing more than an inert layer which is only there to maintain the same optical path length as the D800 but with no AA.

Page <<first <prev 3 of 7 next> last>>

If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.

Check out Drone Video and Photography Forum section of our forum.

Main Photography Discussion

Home | Latest Digest | Back to Top | All Sections

UglyHedgehog.com - Forum