Do NEX cameras have AA filter?

" Does detail (no matter how it is quantified) have a spatial frequency (or a temporal one for the matter?)."

Sure. Think of the little black bars on a white background which you see in lens test charts. Think also of the various frequencies in lines/mm which you see in MTF charts.

Incidentally, there could still be moire even with an AA filter, because the designers have to balance reduction of moire against reduction of desired detail. Some designers, like those of the Kodak SLRn and some Leicas, chose to eliminate AA filters entirely in the interests of detail rendition.

As you mentioned, you can do some AA processing after the signal is captured, but doing it in JPG processing is no good for people who use Raw, which includes many (probably most) who are serious about IQ.

PeteGanzel said:

I suppose I could set up a moire inducing subject and then put the filter in the light path behind the lens. Not sure that would prove anything as I can purposely produce moire patterns depending on subject and magnification with all the cameras I have and they supposedly have an AA filter.

Pete

--
http://www.pbase.com/pganzel

Click to expand...

I believe this is not dissimilar to the protective glass and the IR filter behind it as on the Leica M9... The oblique rays originating from the far protruding rear elements of some rangefinder lenses (wide angles) cause purple fringing near the corners if not also smearing. Lack of offset microlenses and corrective inherent software could well be the reason why their own 16mm wide angle perform so poorly after the two-thirds of the sensor field..

Bob
--
BobYIL

While the clothes in the portrait shot are good, the skin texture is in my opinion not so good. There is fine texture in luminance only, and a lack of colour texture.

Compare this for instance

http://www.flickr.com/photos/rytterfalk/4918239396/sizes/o/in/set-72157624652380587/

BobYIL said:

I believe this is not dissimilar to the protective glass and the IR filter behind it as on the Leica M9... The oblique rays originating from the far protruding rear elements of some rangefinder lenses (wide angles) cause purple fringing near the corners if not also smearing. Lack of offset microlenses and corrective inherent software could well be the reason why their own 16mm wide angle perform so poorly after the two-thirds of the sensor field..

Click to expand...

Bob:

It is even more similar to just about every sensor filter I've looked at..a clear piece of glass in front of the IR absorbing glass. Wouldn't the IR filter alone provide similar protection? I'm assuming an AA filter can use a couple of techniques for blurring such as a finely etched surface treatment or matte coating or suspended particles of the right size. Of course at the level of blurring needed with these tiny sensels, the filter would appear as a clear piece of glass.

Pete

http://www.pbase.com/pganzel

RussellInCincinnati said:

BlueSkies: Contact these folks if interested in removal - they claim that all dSLR have the AA filter, and also explain reason for AA filter and benefits of removing:
http://www.maxmax.com/hot_rod_visible.htm/

Well we know that a blanket claim that all "DSLRs" have an AA anti-aliasing filter is wrong, hence ill-advised.

Click to expand...

Correct, but the majority do.

RussellInCincinnati said:

Consider the well-known fact that the new Pentax 645D has no AA filter.

Click to expand...

The pentax 645D is a MEDIUM FORMAT camera, not 35mm. Plus virtually all Medium Format cameras do not have an AA filter. However, an AA filter can be added by the user if desired.

At this point in time I see zero hard evidence that the NEX cameras do not have an AA filter. With the NEX series, being a consumer camera, it would surprise me if they didn't. It is possible, however, that they have a weak AA filter.

--

The greatest of mankind's criminals are those who delude themselves into thinking they have done 'the right thing.'

• Rayna Butler

D Cox, what does your sample photo tell us about whether or not the Nex has an AA anti-aliasing sensor filter?

Ultimately the original poster just wants to know if the Nex 3/5 14.5 megapixel sensor is being held back, resolution wise, by an AA anti-aliasing filter. You folks can argue until blue in the face about what filter is "technically" or nominally over the Nex sensor. But there's just no need to argue, because John Bean has shown us that there's no AA filter worth worrying about on the Nex.

Also the ease of producing raw photos with fairly high contrast, single-pixel-wide and even single pixel features on the Nex also suggests there's no AA filter worth worrying about. Consider my photo in the next message, made with a decades-old, single-coated 28mm enlarging lens that is not optimized in any way for digital cameras, or distance photography, and sits only 21mm away from the Nex sensor. Criminy, you can see moire patterns near the upper right corner of that photo. How much less AA filter can the original poster want?

The only logical argument with this point will have to include your judgement that, if the original poster sends in a Nex to have it's AA filter removed, the original poster will get back a camera with noticeably higher resolution than an unaltered Nex.

I've been doing some research on these darn optical low-pass filters (OLPFs), which you may think of purely academic interest, but here goes. OLPFs or antialiasing filters are not just diffusers, like Softars or the like. They consist of 2 or more layers of birefringent crystalline material (something with 2 different indices of refraction in different directions). One layer splits each original image point into a pair of points horizontally, another layer does this vertically, so each original point is turned into 4 points (not just a bigger blurrier spot where the point was). The birefringent material may indeed be quartz, or it may be another mineral.

The OLPF is then laminated to an IR-blocking filter behind it. So if you remove your OLPF, you had better put an IR blocking filter in front of the lens, unless you specifically want to have an IR camera. Also, if you remove the AA filter you will remove the dust-removal shaker mechanism.

We know from the spec. that the NEX has an OLPF (or AA filter--same thing). It may be a weak one, because as pixel density increases (and resolution with it), the need for an AA filter decreases. Obviously, a camera designer would like to have the weakest AA filter he can get away with, in order not to degrade resolution more than necessary. With a birefringent material, all you have to do to reduce the spreading of a point into 2 points is to make the layer of crystalline mineral thinner.

I really had to look hard to see the moire in your picture, which is blown up enormously. Remember that this is a consumer camera (in Sony's mind, anyway) and few of the intended users are likely to be making huge enlargements.

Incidentally, I did run across a post on a video forum indicating that the NEX-5 does have a reputation for being relatively prone to moire.

It tells me that the AA filter is not the main cause of poor fine detail.

D Cox said:

While the clothes in the portrait shot are good, the skin texture is in my opinion not so good. There is fine texture in luminance only, and a lack of colour texture.

Compare this for instance

http://www.flickr.com/photos/rytterfalk/4918239396/sizes/o/in/set-72157624652380587/

Click to expand...

If you want to show the moire, you need a repeating fine pattern that exceeds the Nyquest limit/2, I guess.

Russell, how come your photos always have a brownish cast to my eyes? Do you post-process all of them before posting?
--
Gary W.

KM Legacy said:

The OLPF is then laminated to an IR-blocking filter behind it. So if you remove your OLPF, you had better put an IR blocking filter in front of the lens, unless you specifically want to have an IR camera. Also, if you remove the AA filter you will remove the dust-removal shaker mechanism.

Click to expand...

KM:

The anti dust mechanism is just a frame around the sensor filter and can be put back around whatever protective filter you place in substitution.

It has a piezo element on the side that provides the shaking.



Pete

--
http://www.pbase.com/pganzel

KM Legacy said:

I've been doing some research on these darn optical low-pass filters (OLPFs), which you may think of purely academic interest, but here goes. OLPFs or antialiasing filters are not just diffusers, like Softars or the like. They consist of 2 or more layers of birefringent crystalline material (something with 2 different indices of refraction in different directions). One layer splits each original image point into a pair of points horizontally, another layer does this vertically, so each original point is turned into 4 points (not just a bigger blurrier spot where the point was). The birefringent material may indeed be quartz, or it may be another mineral.

Click to expand...

Thanks for this KM. Just checked and the NEX sensor filter is birefringent, which means it not a piece of ordinary glass. Curiouser and curiouser....

Pete
--

http://www.pbase.com/pganzel

Don't know how you could have made an explanation, of how birefringent AA anti-aliasing filters work, much clearer. Can now see how camera designers can "dial in" the spread of a lens image point into a set of 4 points increasingly distant from each other, by adjusting the thickness of a layer of birefringent material over the sensor.

Published Sony specs say there's a low-pass filter on the Nex. If by that they mean an AA anti-aliasing filter, which it sounds like you're pretty sure it is what they mean, then I guess that's it, the Nex does have an AA filter.

You have also made it pretty clear (hah) that it's real easy for Sony to have an arbitrarily weak AA filter. Thus am undeterred from saying to the original poster, that based on how much false color aliasing is evident in raw photos from the Nex with sharp lenses (and in test chart photos by John Bean, and even with 28mm wide-ish angle simple lenses near the corners), that one would be unlikely to notice, or get their money's worth from, having a Nex's AA filter removed.

Gary, it would be great if you could tell me if this particular photo looks brownish on your screen. Have gone to great lengths to get the color balance accurate before posting this sample. On my system, the white painted gutter pipes are indeed white, and the skin tones are both plausible and quite as I remember them from just yesterday.

I thought maybe the sensor was glued to the vibrator, or at least held by some adhesive less permanent than glue. Does it just snap in and out.

If you want to fool with removing those filters, you are braver and more dextrous than I am. Incidentally, I found a business who advertises Sony Alpha OLPFs for $100/each. Obviously it does not cost Sony that much, or the original NEX buyer, but all the replacement parts needed to build a camera usually cost more in total than the finished camera.

According to the Photoshop dropper, the drain pipe, the white squares on the table cloth, and the cloth on the guy's knee all have a slight blue cast.

What colour was the light? Blue sky or grey sky?

KM Legacy said:

I thought maybe the sensor was glued to the vibrator, or at least held by some adhesive less permanent than glue. Does it just snap in and out.

Click to expand...

The filter is sealed to the sensor with double stick foam tape and held by the frame. Apparently the tape is flexible enough to allow the filter to vibrate seperate from the sensor.

You have to peel off the filter, reusing the tape for the new filter.

Pete

--
http://www.pbase.com/pganzel

D Cox: According to the Photoshop dropper, [in the photo above] the drain pipe, the white squares on the table cloth, and the cloth on the guy's knee all have a slight blue cast.

Well we're kind of splitting hairs here. Recall that my original question was, does the photo above look brownish, which it certainly doesn't in my Google Chrome v12 browser. It is true that my color probe shows the white drainpipe as a touch blue, perhaps I went a bit overboard in trying to avoid a "brownish look". So below is yet another version of the photo, a bit less blue, so that the white, branched drainpipe checks out as neutral gray.

My question and concern remains, do these photos look brownish?



What colour was the light? Blue sky or grey sky?/

It was an overcast grey sky. You may notice that this scene was surrounded by large, non-white surfaces. And has white objects that were not painted with professional color calibration reference paints. Making it, as usual, hard to declare what the "right" color balance "should" be.

RussellInCincinnati said:

Gary, it would be great if you could tell me if this particular photo looks brownish on your screen. Have gone to great lengths to get the color balance accurate before posting this sample. On my system, the white painted gutter pipes are indeed white, and the skin tones are both plausible and quite as I remember them from just yesterday.

Click to expand...

White with a bluish tint is common for many "white" items -- a bit of blue is often added to make things look "brighter". So, I'm not expecting white things to be perfectly neutral. But the eyedropper tool is a good idea. When I hover over the gutter, the green component is higher than the red, with the blue highest of all. The blue doesn't bother me so much.

It isn't just this photo. I've been reading your messages for a while, and I've been thinking that perhaps you live in the Matrix or something. ;-) (You know, in the movie, everything had kind of a green cast while you were inside the Matrix.)
--
Gary W.