Nikon D800 is outed with pics and specs. 36 MPX

[Mulis Pictus]

I have in the end decided to get the D800 and should be available in Tokyo in April for a shoot out against any of the 40 mp class backs. :-) whatever comes out it should be good fun!

Interesting. Why the D800 and not the E version? I am just curious because I am having hard time to decide myself.

[BernardLanguillier]

Interesting. Why the D800 and not the E version? I am just curious because I am having hard time to decide myself.

A few reasons:

- I don't believe the small extra resolution is worth the extra time required to deal with moire. I know that I will not have the time to scan images for moire and will end up finding it after printing large images which will generate a significant cost,
- Since I intend to keep stitching anyway, this extra detail is even more irrelevant,
- The difference of price in Japan is nearly 600 US$ which corresponds to the money I just spent on a Sigma 8mm circular fisheye which I find to be much better value,
- I still have philosophical doubts about the relevance of capturing images that can mathematically not be a faithful reproduction of the scene due to the presence of false colors and false details,
- I intend to use the D800 for video also and there are widespread concerns about video on AA filter less devices (unconfirmed so far though),
- For family reasons, I need this camera quickly and the 3 weeks early shipment of the D800 will actually make a difference for me,
- Finally I know from my D3x experience that proper sharpening manages to remove most of the initial softness one can feel when opening images in a suitable raw converter,

We may find out that Nikon did an outstanding job with moire control on the D800E and that there really are no issues with moire. Worth case I'll get one in a few months as a back up. 

Cheers,
Bernard

[BernardLanguillier]

I believe van Gogh had similar concerns.

I think he was on average way to stone to understand the word mathematics in the first place... 

Cheers,
Bernard

[BJL]

- I still have philosophical doubts about the relevance of capturing images that can mathematically not be a faithful reproduction of the scene due to the presence of false colors and false details,

And I have philosophical doubts about using philosophical doubts about not wanting to lose any information as a reason for accepting false or misleading information. It is like a diner in a restaurant picking up eating a scrap of food that fell on the floor to avoid wasting any part of the wonderful food.

That is, I agree, but mostly I think that the philosophical claims on either side (see Sean Reid's essay for the other side) can not decide it: we need to look at how filtering and post-processing options affect the sort of pictures and prints that we are interested in. (Or, "Can I just brush that dust of the food I dropped, or will it taint the taste of the meal?")

there are widespread concerns about video on AA filter less devices

Local video guru Graeme Nattress seems to agree, with motion handling and compression an extra issue not facing still photography:
http://www.luminous-landscape.com/forum/index.php?topic=62537.msg503875#msg503875

[Ray]

A few reasons:

- I don't believe the small extra resolution is worth the extra time required to deal with moire. I know that I will not have the time to scan images for moire and will end up finding it after printing large images which will generate a significant cost,
- Since I intend to keep stitching anyway, this extra detail is even more irrelevant,
- The difference of price in Japan is nearly 600 US$ which corresponds to the money I just spent on a Sigma 8mm circular fisheye which I find to be much better value,
- I still have philosophical doubts about the relevance of capturing images that can mathematically not be a faithful reproduction of the scene due to the presence of false colors and false details,
- I intend to use the D800 for video also and there are widespread concerns about video on AA filter less devices (unconfirmed so far though),
- For family reasons, I need this camera quickly and the 3 weeks early shipment of the D800 will actually make a difference for me,
- Finally I know from my D3x experience that proper sharpening manages to remove most of the initial softness one can feel when opening images in a suitable raw converter,

We may find out that Nikon did an outstanding job with moire control on the D800E and that there really are no issues with moire. Worth case I'll get one in a few months as a back up. 

Cheers,
Bernard

Very wise decision, Bernard  . A major advantage of the lack of an AA filter in MFDBs was the reduced cost of a system that was already ridiculously expensive. However, the D800E has no such advantage. The cost of a lack of an AA filter is greater. One is paying more for an absence of something. Sounds a bit crazy to me  .

Another issue which I confess I don't really understand, because I'm not as smart as some of you very technical guys, is how a signal can have it's vertical resolution diminished in one instance, then restored in the next instance. I'm curious as to what's actually taking place here. Can anyone elaborate?

[BJL]

... how a signal can have it's vertical resolution diminished in one instance, then restored in the next instance. I'm curious as to what's actually taking place here. Can anyone elaborate?

As indicated by Nikon's illustration here
http://www.dpreview.com/previews/nikonD800/page3.asp
the beam is split by one filter layer and then reunited by another. Talk of "diminished resolution" is misleading --- it is just that the light from one point is temporarily spread over several locations, which is similar to the state of the light from a single point on the subject as it travels through a lens before reaching the focal plane. The strange and somewhat costly split-reunite strategy is, I am guessing, in order to keep as much of the optical path and design the same in the two models, to avoid having to make other more expensive design changes in what is expected to be a far lower volume alternative. And that "lower volume" is the most obvious explanation for the price difference.

P. S. to John Camp below: it might have been my post you are recalling; I have been shopping that hypothesis in various places. Also, for $300, I will provide a UV filter smeared with a small, scientifically determined amount of vaseline, to convert the D800E back to a D800.

[John Camp]

Very wise decision, Bernard  . A major advantage of the lack of an AA filter in MFDBs was the reduced cost of a system that was already ridiculously expensive. However, the D800E has no such advantage. The cost of a lack of an AA filter is greater. One is paying more for an absence of something. Sounds a bit crazy to me  .

I saw an answer to this somewhere that made some sense to me; it may have been on DPR. For cost reasons, the D800E does not actually have fewer filter elements than the D800 (as you note in the bottom half of your question.) This may be the case because simply leaving the filters out entirely may have required other adjustments in the camera that would be more expensive that putting them in, in a different orientation. At the same time (this other answer suggested) Nikon expects to sell significantly fewer D800Es, and so the per-unit cost of the altered machine rises, given the additional machining work, etc.

Here's a question that I would like somebody to address: WOuld it be possible for, say, a high-end filter maker to make a blurring filter that would fit on the end of the a lens, that would have the same effect as an AA filter? So that you could buy a D800E and add a filter if you thought it was necessary?

[John Camp]

And here's another question. Does moire occur with any regular repeating structure that's of the proper size to generate it, or does the structure have to have a rectangular form? That is, would a series of regular, closely-spaced, curved forms, like the scales on an insect wing, also generate moire?

[seanesopenko]

Here's a question that I would like somebody to address: WOuld it be possible for, say, a high-end filter maker to make a blurring filter that would fit on the end of the a lens, that would have the same effect as an AA filter? So that you could buy a D800E and add a filter if you thought it was necessary?

Has anybody had experience with these things?  They screw in front of the lens and provide different levels of AA filtration.

[Fine_Art]

It should have no AA filter. On the rare occasion that you are dealing with a fabric or similar that will cause the problem the photographer is smart enough to know that moving closer or farther away will remove the unfortunate frequency. Alternatively a "pantihose filter" also does the trick.

The unfortunate downside to this is the manufacturer has to sell you less when they want to sell you more.

[BJL]

... the photographer is smart enough to know that moving closer or farther away will remove the unfortunate frequency.

If that moving is possible, and if the photographer can either anticipate the moiré in advance, or can detect it on review (zoom to 100% pixels on LCD, pan around for a minute) and has the opportunity to move and reshoot. This would not work with the example posted recently of aliasing in roof tiles, where the distance is too great for a change of camera position to help, nor for situations where the photographer cannot move easily or when there is only one opportunity to get the shot.

So sometimes, some photographers need the camera to get it right the first time.

[BernardLanguillier]

I have philosophical doubts about anyone who has philosophical doubts concerning the ethics of inducing false colours and false details by using a camera without an AA filter and yet who champions stitching.

That actually is an interesting comment worth discussing in depth but I will unfortunately not have enough time today.

There is in fact in my view key philosophical differences between:
- an image whose every pixel is potentially nothing but a digital artifact even if the macro image looks real,
- an image blurred because of an AA filter,
- an image whose corners are blurred because of the limitations of a lens,
- a stitched image with possible stitching artifacts,
- an image blurred because of slow shutter speed,
- an image blurred because of poor optics,
- ...

They are all imperfect projections of reality and play differently with the additional filter called our senses. My view is that some are less fake than others but I would already reach a high level of philosophical enlightment if we could agree that they all are. The rest will be a matter of how fake we want our images to be.

But I do apologize for the usage of arrogant words like "philosophy"! :-) nonetheless, this philosophical argument is only one of the reasons why my next DSLR is probably going to be a D800.

Cheers,
Bernard

[seanesopenko]

I've always liked stitching not for more resolution but for control of the projection and therefore the composition.  Lenses are either fish-eye or rectilinear.  Digital processing makes it much easier to use other projections such as cylindrical.  There are also stitched images where the camera is moved down a street, for example.  This type of stitching gives an image not unlike a painting in that it's a flat field rather than a one eyed image drawn out by a cyclops.

[Michael LS]

Just some observations...

Human nature never ceases to surprise and amaze.

Over the years, and on many a thread, I've seen photographers lament,
complain, and emote, to the point that some even send their
cameras away to have the AA filter removed. I've read their pleas,
suggestions and fervent hopes for camera makers to offer a next gen
model "without an AA filter".

Now, with the D800e, Nikon has come across with exactly what they've
so ardently asked for and dreamed of. What do they do? Well, I won't
speak for all, but from what I've read at LuLa, and other sites, many
are now running the other way, in rabid fear of the demon moire!!

  Gotta luv it, I guess.

[Josh-H]

What do they say.....

'Be careful what you as for.. because you just might get it.' 

[BernardLanguillier]

Bernard, no need for apologies, I hope you enjoy your new camera.

I am looking forward to using some lenses with it that are really not a good match for DX bodies. The 85 f1.4 AF-S and my new Sigma 8mm fisheye should deliver their full potential.

Cheers,
Bernard

[Ray]

As indicated by Nikon's illustration here
http://www.dpreview.com/previews/nikonD800/page3.asp
the beam is split by one filter layer and then reunited by another. Talk of "diminished resolution" is misleading --- it is just that the light from one point is temporarily spread over several locations, which is similar to the state of the light from a single point on the subject as it travels through a lens before reaching the focal plane. The strange and somewhat costly split-reunite strategy is, I am guessing, in order to keep as much of the optical path and design the same in the two models, to avoid having to make other more expensive design changes in what is expected to be a far lower volume alternative. And that "lower volume" is the most obvious explanation for the price difference.

As someone who stopped using protective UV filters years ago because of the possibility of unwanted reflections, as well as the additional cost, I'm a bit concerned psychologically about the presence of two unnecessary pieces of glass in the path of the signal, that are both actually altering the signal but serving no ultimate purpose.

Such concerns may well be entirely groundless, but one can't help wondering about the effects of even the slightest misalignment due to imperfect quality control.

In other words, is the difference in crispness and that magical 3-dimenional quality likely to be as great as it might have been without those two unnecessary pieces of glass in the path of the signal?

I would have been happier if the initial design of the camera physically excluded any AA filters and the premium was charged to those who prefer the model which included the AA filters.

[hjulenissen]

And here's another question. Does moire occur with any regular repeating structure that's of the proper size to generate it, or does the structure have to have a rectangular form? That is, would a series of regular, closely-spaced, curved forms, like the scales on an insect wing, also generate moire?

If you have no AA filter, you will have aliasing for any scene/lense/stand/... combination that project light with sufficient high spatial frequency onto the sensor.

This aliasing may look "good", "bad" or "just the same" for any one viewer in a particular setup. If the scene contains highly regular structures (typically, but not necessarily man-made) of high contrast and/or color-contrast, this aliasing will typically give false-frequency content that can be quite visible, that most people dislike, and that can be hard to remove.

-h

[Mulis Pictus]

A few reasons:

- I don't believe the small extra resolution is worth the extra time required to deal with moire. I know that I will not have the time to scan images for moire and will end up finding it after printing large images which will generate a significant cost,
- Since I intend to keep stitching anyway, this extra detail is even more irrelevant,
- The difference of price in Japan is nearly 600 US$ which corresponds to the money I just spent on a Sigma 8mm circular fisheye which I find to be much better value,
- I still have philosophical doubts about the relevance of capturing images that can mathematically not be a faithful reproduction of the scene due to the presence of false colors and false details,
- I intend to use the D800 for video also and there are widespread concerns about video on AA filter less devices (unconfirmed so far though),
- For family reasons, I need this camera quickly and the 3 weeks early shipment of the D800 will actually make a difference for me,
- Finally I know from my D3x experience that proper sharpening manages to remove most of the initial softness one can feel when opening images in a suitable raw converter,

We may find out that Nikon did an outstanding job with moire control on the D800E and that there really are no issues with moire. Worth case I'll get one in a few months as a back up. 

Thanks for sharing your reasons.

I am quite tempted to get the E version. I think I will wait to see more samples. Hopefully LR and other tools will get support for D800(E) soon after the cameras start shipping.

My thoughts are:

- one cannot get rid of a moire even with the AA filter anyway. See current Lloyd's blog for example, the D3x test pictures with the roofs have a lot of it. The bad thing about moire is, that once you start looking for it, you find it more easily :-)

- I am not sure about the false details, are the blurred details more real than the aliased ones? Probably not, sometimes maybe better looking and sometimes not. Deconvolution is great, but has trouble with noise and cannot reconstruct the signal completely. It might introduce slight halos as well

- I will be (probably) not doing any wedding or studio work with clothes, so the risk of loosing critical shot because of moire is quite low

- don't hear much complaining from MFDB users about moire

- OTOH the price difference is indeed not negligible, nearly 20% here as well right now

It is a bit difficult to choose, but having the choice is great and I am happy Nikon offers both versions ;-) I would prefer to have the E version without any low pass filter though.

[hjulenissen]

- one cannot get rid of a moire even with the AA filter anyway. See current Lloyd's blog for example, the D3x test pictures with the roofs have a lot of it. The bad thing about moire is, that once you start looking for it, you find it more easily :-)

You cannot remove it completely, but you can have less moire at the price of reduced sharpness.

What kind of scenes and print sizes do you do where the difference in sharpness is so important?

-h