I'm finding this a bit odd wading into the Nikon forum. but anyway just found this on another site & found it interesting.
I'm finding this a bit odd wading into the Nikon forum. but anyway just found this on another site & found it interesting.
Well that settles it for me.
the 'extra detail' provided by the D800E is rubbish!
it simply looks to be oversharpened but with a hint of moire to boot!
fuzzier looking pics have never looked so good
When Nikon starts using Foveon sensor designs in the cameras with weak or no AA filter, then I may redress that option again one day.
OK.. a change of heart.
RG has downloadable raw files from which you can make your own assessment of the differences between the two.
I didn't initially see that, and also partly missed that he processed his raw files using Photo Ninja.
I read that he did, but it seems that the majority of the images were processed with this Photo Ninja, and not CNX2.
I loaded those NEF images into CNX2 and got much better looking results than Rob seems to have got.
Firstly the sharpening on that street corner image is horrible on the D800E. . the halos in some areas are too strong, compared to similarly processed images in CNX2.
BUT!!.... in CNX2 you need to turn off the in camera sharpening and only use the USM tool that CNX2 has on offer!!
No halos, much lower moire on the un-moired D800E image and the moire reduction tool in CNX2 takes care of the issue in the image well enough.
Now that I've revisited the article, I've kind'a changed my mind back toward the D800E(unless a better camera is announced of course )
If I were a birder with some of those high quality Nikon super teles, I'd still be weary of moire being present on very fine bird feather details tho!
For the most part, the D800 images can be USM'ed to a very good level to regain that 'lost detail'.
Although it must be noted that not all detail lost in the D800 image will be recovered compared to the super duper fine detail that the D800E can provide.
I've been trying to find this moire of which everyone speaks. Nothing, nada, zip, zippo, zero so far. Tested with 200/2, 300/2.8, 400/2.8, 600/4.
i just can't seem to get any happening. The detail is mind boggling however. I can crop like never before.
Here's an example of a BIG crop using a 24-70, handheld, 1/80, f6.3 on a dull overcast day. The feather detail is just amazing, and is the closest thing to moire I can produce so far. i am not saying I won't get any in the future but at this stage its a non event.
Last edited by Sylvia; 07-05-2012 at 3:29pm.
This is good, and it seems that choice of raw conversion software is partly to blame(possibly!!).
While you may be no good at inducing moire Silvia, this doesn't mean it's not going to happen(as you've already pointed out).
Once it does happen, it happens and has to be selectively processed out, otherwise possible other issues may manifest themselves in the image.
I downloaded Rob Galbraith's images(his two raw files of the street corner) opened them in ViewNX2(fast!!) and CaptureNX2(less fast ) and the moire he's presented was at least half as much and easily dealt with in CNX2's moire reduction section.
ie. I'm concluding that from Rob Galbraith's testing .. he's using the wrong software(some beta raw converting software called Photo Ninja).
The point of the exercise tho is that once you got it in your images, it's in the images, and while it's definitely true that you can capture 'more detail' natively from the D800E images with a lower sharpening routine than you can with a D800, the D800 image loaded into CaptureNX2 had no moire at all!
I'd love to show the crops I've rendered of Rob's freely available raw images, but he has a copyright issue on them that I'm not going to infringe upon.
The real question is not so much whether you can see the moire or not, it's more about is the risk of getting it worth the hassle.
I processed the D800 image from RG to a level that is pretty close to what the D800E achieved in terms of detail and edge contrast, and it still produced no moire at all.
The other issue I noted in that street corner image comparison is a 'pseudo moire effect' where I wouldn't have expected it.
Rob has pinpointed it out with his crop link to the sign.
The D800E's rendering of the sign is a coloured mess. Looks like moire, but doesn't behave like moire.
(well CNX2 doesn't recognise it as moire, anyhow).
Using CNX2, the shoes on the billboard cleaned up perfectly with a medium amount of moire removal, and no colour was lost anywhere else on the rest of the image.
Trying max moire removal level only served to reduce the saturation of the red channel, and still not remove the colour artefacts on the writing on the blue sign.
D800 has the same level of detail as the D800E, when appropriately sharpened but none of that weird colour artefact on the blue sign.
Strange part is that traditionally, moire is an issue when the sensor encounters very high frequency repeating patterns in the scene, but where those patterns are of a repeating nature.
The writing on that sign is not really of a repeating nature, where the micro fine details are different enough that they shouldn't cause moire in the image ... but they did!
I still have a little while yet before I finally get my camera, and my decision will eventually be made after more comparisons are made by respected reviewers.
So far, from what I've seen, the D800E is probably worth the extra effort that could be involved.
So far, the moire issue doesn't appear to be anywhere near as drastic as the D70 was in it's day.
The detail from both the D800 and D800E is incredible. Croppability is amazing to say the least!
Here is one from the D800 + 500 f4 VR handheld and cropped to various amounts:
This is the original @ f5.6
Cropped to the same size as it would be in DX mode, ie about 66% of the original image or similar to using a 750mm lens instead of the 500mm I used.
Vertical crop for variation.
Super close crop. This is about 35% linear crop of the original, or similar to using a 1500mm lens, instead of the 500 I used.
Super duper, no holds barred, close up shot. This equates to about a 15% linear crop of the original, or, if you like, taking a photo with a 3500mm lens instead of the 500mm I used.
Last edited by Lance B; 09-05-2012 at 3:57pm.
My PBase site: http://www.pbase.com/lance_b
Considering the size of the bird, thats just unreal how much you crop down to and still have alot of detail. Even the last image still looks good to use probably even better after some dodging and burning perhaps.
Photographer & Retoucher at L'Obsession Secrète
Great example Lance. I always wanted a 3500mm lens.
welcome to the world of medium format digital cropping! well almost
OK, So it's a case of extra work with CNX2 vs no work at all with the d800?
I'm liking the Lorikeet very much.
In any case, the shot of the sign to my eyes it was obvious both had moire, and while the 800e looked much more severe, the image was still much sharper.
If it is a case of photo ninja software being next to useless with a d800e, that's not necessarily a bad thing?
wonder what the sign would be like with CNX2 applied instead.
What I've seen(at least in this image) is that if you use the high setting in the moire reduction tool, you lose saturation in the red channel, but the coloured lettering in the sign is not fully elliminated
It certainly gets better, but considering the red channel loss, it may not be worth the effort.
Without affecting the rest of the image too much, I found that the best method to reduce the colouration of the writing on the sign, which looks like a moire effect, was to use the Selection Control Point, reduce the radius so that the effective edit area is only on the sign itself, make sure that the centre is located on the lettering on the sign, and use the LCH editor to reduce saturation level to zero.
This has the effect of keeping the sign board blue, but the annoying(and to me obvious) yellowy red lettering is rendered white(as I expect it should really be).
The very slight effect on the blue on the sign board is minimal, and to my eyes acceptable, and I suspect that you could process that slight colour change back in again if you were picky.
The point is, that I haven't seen any really efficient way to process moire out of the image without resorting to detailed editing techniques(which of course means more work).
That is, the moire reduction tool, can't always easily or effective be used, which minimises it's usefulness in a batch process(if batching was important).
BUT!! on another note:
I opened the same images again with Fastone's FSViewer and noted that the images rendered in FSViewer were almost completely unaffected by moire.
Where the images (I downloaded) had some moire in the shoes on the advertisement sign using both VNX2 and CNX2, in FSViewer there is zero moire in the shoes!
There is still the issue of the moire effect on the lettering of the blue parking sign tho, but it does look slightly less intrusive than it does via CNX and VNX.
This is due to the fact that FSViewer uses the embedded jpg file to display the image, and doesn't interpret the raw file data to render an image.
If you switch FSViewer into raw file mode, not only does it slow down to glacial pace(much sower than either CNX or VNX), it also renders quite an 'obscene looking'(IMO) image, but it has moire moire than ever before
Much more moire than CNX and VNX. And the rest of the image is quite ordinary looking to boot.
I suppose one theory as to why this can be happening is that the camera may be processing some of it's own moire reduction into the file.
FSViewer is quite a handy program to have access too for many reasons, but my main reasoning for it's use is for when I process images and want to use it for other reasons.
If I process a raw file using CNX or VNX, not many other image software recognise the edits in the image, as they may be trying to render the raw files from the raw data.
But as CNX and VNX will alter the embedded jpg file in the raw data, FSViewer renders the Nikon software edited raw file 'properly'(what I believe to be properly, anyhow).
Not a vital bit of info, I suppose, and most folks will almost certainly be tied to their choice of software.
While I'm also tied to mine, I also keep other handy tidbits of software for various reasons too.
(eg. FSViewer is my portable/mobile image viewing software, kept on a USB stick, for when I need to display images on someone else's PC. Not many other software will run off a USB stick).
This is only really important if you are viewing your images via multiple brands of image viewing software.
They render differently under each set of conditions.
That is, while one fellow says this moire issue is a furphy and "I can't see it", doesn't necessarily mean it doesn't exist or won't be a problem.
The software used to render the image is what's affecting their judgement here, not the actual degree of the issue.
FWIW tho, the moire issue is no where near what I was expecting it to be, going with my D70s experience.
D70s had a weak AA filter fitted to it, which displayed 'sharper' images than the D100, which it received it's sensor from, but then again having these sharper images, it did have a tendency to produce moire a lot more efficiently too.
So while the D100 looked to be less sharp, when you got moire in your D70s files, it was hard if not impossible to remove it all, if you could!
My suspicions are that Nikon has added a moire reduction in it's image processing on the D800E(possibly the D800 too???).
On the basis of how FSViewer renders the raw files in either raw file mode and embedded preview mode, and the wild difference it provides, I think this may be the only explanation.
my current opinion on the two cameras and editing workflow. While there may be an extra bit of work to be done on a D800E image over a D800 image, it may not be overly complicated as it stands now.
CNX2 can effectively cut moire to a point where it's not going to be too bothersome.
If you are seriously pedantic, then there will be some cases where moire had to be locally edited out on a per image basis.
For a working pro, this may not be ideal. For a hobbyist(like me) who has the 2-5mins per image to edit out the aberration in the image, it's actually quite easy.
As time goes by, and if moire is seen to be more of an issue on more images from more complainers, then maybe the software writers will provide more efficient and accurate tools to eradicate the problem easily.
You know I would be interested to see a crop of a bugs eye say from a fly or dragonfly from a D800e to see if moire would be effected in that area giving the texture & wide range of colour the eyes from a bug can produce at the best of times.
I have yet to see a good example surprisingly from either model.
A bugs eye may or may not induce moire depending on lighting I guess. It's no so easy to predict, but the conditions for inducing moire are fairly easy to find.
Nikon's own site highlighted the conditions, with at least one example that I can remember, from Nikon themselves
If you can't see the moire in that shot ... you have to ask yourself .... do you really need a D800E?
The 800E is a hands down better output IMHO. Detail is critical for my work, cannot wait for it to turn up..
Rob's site has some great examples, a very interesting read.