Sharp thinking: Nikon creates selectable strength low-pass filter

Nikon has patented a technology that can electronically adjust a camera’s low-pass (AA) filter based on the situation. By using a liquid crystal panel, the AA filter can either be turned on and off, or set to ‘normal’ or ‘high’ intensity. The first design could allow for a digital SLR to have its AA filter turned off at the press of the button. The second design would have a mild anti-aliasing effect for stills, and a stronger effect to eliminate moiré in movies. The company suggests still more could be achieved by mounting two such filters back-to-back.

The patent claims the entirely electronic design means a two-mode low-pass filter can be implemented in a way that doesn’t decrease the reliability or durability of the camera (rather than a mechanical system that switched different filters into place).

How it works:

The designs are essentially a development of the filters used in the D800 and D800E. The standard D800 has a low-pass filter that utilizes a property called birefringence – where a substance splits light depending on the light’s polarization. The thickness of the filter defines how much the light is split, so the thickness of the filter can be tuned to match the sensor it’s placed in front of (Blurring the light across the width of one pixel cuts out the high frequencies that would cause moiré).

The D800E features a second birefringent plate of the same thickness, set so that it re-combines the light split by the first plate – cancelling the effect.

The latest design can be an extension of the system used in the D800E, but instead of having two birefringent plates that cancel each other out, there’s an extra layer that can selectively combine or subtract their effects.

The Nikon patent inserts a liquid crystal layer between the two plates. A liquid crystal layer can be made to rotate the polarization of light or just let it pass straight through, depending on whether an electrical charge is applied to the layer. This ability to change the polarization of the light as it passes between two polarization-sensitive plates lets you decide whether the effects of the two plates is additive or subtractive.

In the simplest system described in the patent, the first layer separates the light before it hits the liquid crystal. With the liquid crystal turned off, the polarization of this split light is rotated through 90 degrees, causing the light to be recombined by a second birefringent plate. However, if the liquid crystal is turned on, it doesn’t rotate the light, and the second filter exaggerates the splitting effect and increases the blur.

The result is a low-pass filter with two modes. In one instance you could have a filter that can either be turned on or off (offering a high-res mode if you’re shooting landscapes or situations without repetitive patterns).

Design 1: AA filter on/off

Here, the liquid crystal is turned off meaning the light reaching the second plate is the opposite polarization from when it hit the first one – cancelling the two plates’ effects.

This allows a single point of light to be focused on a single point on the sensor. This is the equivalent to having no low-pass filter.

[Photo credit: Japan Patent Office]

By turning on the liquid crystal, the polarization of the light is unchanged and the effect of the two plates is combined.

As a result, light lands on two different points on the sensor, blurring the image and acting as a low-pass filter.

[Photo credit: Japan Patent Office]

Alternatively, you could design the filter with a stills mode (blurring light by the width of one pixel to avoid moire in full-res images), and a video mode that splits the light over a greater area (to prevent moiré at the much lower sampling frequencies used by video).

Design 2: AA filter normal/high

Here, the front birefringent element is thicker which, when the liquid crystal is off, has the same effect as a standard AA filter.

[Photo credit: Japan Patent Office]

With the liquid crystal on, light is split even further, providing a stronger AA effect ideal for video recording.

[Photo credit: Japan Patent Office]

Putting two such assemblies back-to-back would further expand the capabilities of such a system. These could include a camera with off, stills and video modes, or allow the horizontal and vertical filtering to be separately adjusted to take into account how the sensor is sub-sampled when shooting video.

Researchers in Tokyo develop high-speed subject tracking system Aptina's Clarity+ sensor tech promises 1EV improvement Bell Labs creates lensless single-pixel camera Fujifilm harnesses silver for new touch-screen technology



For a photo-specific device, its best to just leave the AA filter out.. the only reason for this complexity is to accommodate the photo-video market, which I prefer not that they just leave out. i.e. for the D800e and upcoming 54mp D4x.. just market it as a photo taking device.


A general comment: We all parse on each and every bit of news regarding which camera is “best”, the weight vs. feature, will this change be useful, the new low pass, the next hi-tech change. I for one rarely print, had a darkroom back-in-the-day, photography was a full time gig, now I shoot, post, email and occasionally print. The changing, “newer-better” tech has little bearing. Am I alone in that95% of my work exist as digital? Do many of you print, frame or sell your work? Its been a strange trip from the time and patience it took to bring photos to perfection, jiggling trays to Instagram, where I can see some excellent work and post my own and get feedback in seconds. AA filters? Gold Awards? Pros and cons of RAW’ It’s hard to buy a truly bad camera these days, the differences are small, many lenses are made by the same manufacture, pocket size this, lens heavy that, other than full time working professionals, what does it matter to the 99% whose photos never leave the screen?


I find your comment intriguing. I just finished reading the article and my first thought was “what is left besides the photographer’s eye?”
Once upon a time, when I processed Cibachromes although I preferred b/w, each print, which took hours to produce was for me a masterpiece. I worked hard at it. I manipulated colour, was precise in my camera exposure, or…played with the exposure to get the right effect.There is room for experimentation these days- but every now and then when I am asked to teach I wonder, how do you teach basics any more. Every digicam is a computer. Every computer can be improved upon. Images are better and easier than ever. I revel from time to time on how easy it is to capture a moment in time- still… the more we improve, the harder it is the claim an image as “our own” and not that of the digicam.


In the beginning there was the D800 for $3000. So Nikon stripped the AA filter and tge D800E was born. Because less is more, Nikon charges $3300 for the D800E. The next logical step would be a D800F with a configurable AA filter for $5000. E la nave va.

Dave Luttmann

They didnt strip the AA filter out. Looks like you should read up on the topic before you post


Yay, when this technology trickles down to the dxxxx series, there will be no more complaints of soft pictures…

Rob Bernhard

[[there will be no more complaints of soft pictures…]]


Thank you for that. I now have to clean coffee off my monitors. 🙂


Wow, this will bring Nikon to leadership position, hahaha. How about lower prices on the better stuff. At Nikon, every buton or hinge added, doubles the price.


Haven’t electronically-controlled ND filters been used before in compact cameras in lieu of actual aperture blades to reduce diffraction? I remember hearing this. Of course, Nikon’s proposed system is much more advanced.


Couldn’t they just take Panasonic’s approach to video, scaling down the entire frame instead of line skipping? No moire at all on my GH2, and it’s infinitely sharper and more detailed than my old D7000 and current D7100…


Two polarizers: how much light loss there is?

Sounds like a good idea which is really not needed. And even less in the future with ever sharper sensors.


I think AA filters don’t make sense anymore. Just have small enough sensels to be past Nyquist for any lens and do analog/digital reduction in resolution. Easy. As I account for Bayer filters, a 5um sensel is at Nyquist for a lens resolving 50lppmm — which few do corner-to-corner. 5um is just 14MP APS-C, or 36MP FF — i.e., the D800. A little higher sensel count just makes the AA issue disappear.


The D7100 pixels are 3.9 um. That’s the major reason it doesn’t suffer from lack of an AA filter.

Ivan Azzopardi

It seems to be nice but what about the weight would be that of the D800 or D700 .I prefer them a bit lighter as I can use more heavier sharper lenses. The lighter the better for travelling.

Nigel Wilkins

Do you really think you’ll notice the weight of an AA filter?? It’ll be less than the weight of a larger diameter lens cap!

Ivan Azzopardi

Nigel I am not saying that the filter will effect the weight of the camera but the camera that will be introduced with this filter for eg D900 will be the same weight of th D800


I wish the tech was able to actually adjust the strength of the AA filter.

Still, a very exciting development.


Unnecessary complication I guess; just one more thing that can break down and that costs extra money…right software can do same trick just fine and most people (who can afford such camera) probably already have it.

Ken Phillips

Actually, you want to quash moiré at the source … software has no clue as to what image created the values on a given spot on the sensor. Of course, if you are diffraction limited, you no longer need an AA filter (you’ve effectively got one!)
Personally, I’d rather have it “in or out” selectable (flip up like the mirror?) than have several more layers in front of the sensor.


Less time editing = more time shooting.


Good idea, probably not suited to cameras used in varying temperatures, LC properties will change with temperature, causing AA control failure.

Dimitris Servis



Outstanding. You’ll be able to see all those oil spots on the sensor really clearly now. 😉


It may require oil change every 25k shots 😉

Hubertus Bigend

I’m not in the least sure that I understood the principle correctly, but wouldn’t such a device, no matter to what specific setting it was configured in a specific moment of time, always somehow act like a polarizing filter screwed onto the lens, procducing more or less unnatural images?


It’s Nikon so you have a nice green color cast anyway.


I was wondering if this was going to happen… awesome it is.

Model Mike

Totally awesome… and totally marginal benefits.


sounds need in theory, but if Nikon needs as long as a D16-85 VR f4 to take this in production, the design is absolete before we can see it working in a actual real camera.
Soon the mirrorles cameras will take over.


And this idea can’t work in a mirror less camera because….?


bla bla bla


Yeah sure mirrorless cameras are going to take over just as soon as vendors figure out how to sell them in large quantities. Right now, they aren’t exactly flying off the shelves.

But poor sales is only a minor sticking point for mirrorless cheerleaders.


A little more crispness in my D800 files would be a bonus but I can’t help thinking that the effect of every additional translucent layer and surface placed between the subject and image is increasingly detrimental, especially to saturation and dynamic range.

Improved saturation, dynamic range and contrast have the effect of improving clarity and therefore also contribute to the impression of sharpness and resolution

Paul Hodgson

James, as a recent D800 users but a long time Nikon user I’ve found that turning up the sharpening on my D800 to 5 results in a sharper image. However, it’s a compromise, I must first process my images in either ViewNX2 or NX2 which I’m happy to do. On a personal experience level I find that shooting raw the file never appears overly sharp, just enough to make the difference.

Using Lightroom or another RAW converter I don’t find the same output. With some other software applying sharpening later seems sometimes to create oddities which then require yet more time to correct.One thing I will say about D800 raw files, they’re enormous. Couple that with a computer system not set up for speed added to software that’s actually slow (NX2) and you’re in it for the long haul! Me, I need a better speedier computer but still prefer processing my RAW files through Nikon software.


Stronger AA for reducing moiré in video???
This is as ridiculous as it gets. Take a lesson from the Panasonic GH2 and their like and do some serious binning instead of heavily subsampling the sensor. 😛


Subsampling requires far more powerful hardware than currently available at, say, Nikon. For example, with the 41 Mpixel sensor of the Nokia 808, the bandwidth is some 1 Gbps.

No wonder very few cameras (GH2 to a certain degree, Nokia 808, Nokia Lumia 1020) use true subsampling when shooting video.

All in all, a configurable, video-friendly (that is, one that can be configured to be VERY strong) AA filter is definitely welcome.


If Nikon and Canon are struggling with 2k video, imagine what will happen when 4k and 8k video comes. The future belongs to the electronics companies: Sony, Panasonic and Samsung.


Is it useful at all? As of today, a D800E resolution is very difficult to make out unless you use very expensive lenses.
I think the weakest link in the chain is the quality of not so expensive lenses.
this tech will make sense in a 10000$ camera for professionals who do big prints.
However, for the 2000-3000$ buyer this tech will be useless as most of them will not own lenses which can take advantage of no AA filter.
Unless lens tech advances so that quality comes at an affordable price, this is not so useful.
Come to think of it, 50 year old lenses have similar quality to todays lenses optically.


You’ve answered your own criticism. Nikkor H series and AI lenses are available at very reasonable prices. The ones that I own (24, 28 (H), 50 and 35-70mm) out-resolve the D800 sensor. These lenses even out-resolve the Nikon 1 sensor.


Neat idea… looking forward to see how it works under real life conditions…


Just one more reason to charge more $$$, we need sensors with more DR!


Greatly needed by the 12MP FF of years ago but not so much today. What baffles me are MF cameras that didn’t have such filters especially with their bigger pixels and larger lenses. Anyway, still a good patent that might be useful.


call me a pessimist but it just sounds like something else to go wrong within a camera. Don’t get me wrong it is an exciting and ground breaking development. i just keep thinking of all those old cars we see with hardly any electronics and the newer modern cars that are continually developing problems in their electrics.


On the other hand if we fear failures within evolution we would still be riding horses today.


True we need to evolve, i’ve just learn’t never to buy a product as soon as its out let others find the flaws and problems.


Or not even that… the horses can die, bite, and kick you off…


I agree entirely with this statement.
One more thing to go wrong, more complexity within the optical path to degrade image quality, one more menu setting or switch on the camera to confuse you.
In the old days, you could pick up just about any film SLR and you would automatically know how to use it, you would know exactly what every lever and knob does. Camera manufacturers should concentrate on simplifying and making the controls more intuitive rather than piling on more features and settings.

Just my opinion.

Nigel Wilkins

Or the alternative view…in the old days, you didn’t have the choice.

You can do the same with digital once you’ve set it to your liking. Just because you can adjust something, doesn’t mean you have to.


A small but worthwhile evolution it would be.

Though i still don’t see any need, by miles, to give my D800 more resolution as it is. What a beast 🙂


it’s not about res, but extra sharpness, no ?


“it’s not about res, but extra sharpness, no ?”

And also lack of moire / aliasing in videos – something DSRL’s are suffering from ATM…

Der Steppenwolf

A novel approach but how log will we wait for this to come out ? And will it be just another one of those patents that Nikon has but we are yet to see anything of..?

Source Article from