Pages: [1]   Go Down

Author Topic: Imaging resource samples D800, D800E and Pentax 645D (Moiré)  (Read 7620 times)

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto

Hi,

Imaging Resource published samples from both the D800 and the D800E. One of the test images I checked earlier with the Pentax 645D had a lot of quite unexpected Moiré. I downloaded and compared the three corresponding images.

The Pentax shows a lot of moiré on Samuel Smith label, the D800 none and the Nikon D800E some.

I have two screen dumps from Lightroom, first is with "landscape" sharpening on all. On the second one I applied a deconvolution sharpening I would guess is adequate for AA-filtered images.

Of course it can be argued that all images could be sharpened in the same way, but AA-filtered images need more sharpening, it is part of the equation.

To put things in perspective, very few landscape photographers ever complained about moiré, and if a photographer doesn't have problems with a large pixel pitch MFD (like old 25 MP backs) it is highly unlikely that he/she would run into problems with the D800E.

On the other hand, the image is a fine illustration that moiré can show up in unexpected places.

Image order: Pentax 645D, D800 and D800E, left to right.

Best regards
Erik
« Last Edit: May 10, 2012, 12:41:24 am by ErikKaffehr »
Logged
Erik Kaffehr
 

ckimmerle

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 441
    • http://www.chuckkimmerle.com
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #1 on: May 10, 2012, 12:32:00 am »


On the other hand, the image is a fine illustration that moiré can show up in unexpected places.


Actually, it's not all that unexpected that a printed label would suffer moiré as it's made from uniformly placed dots. Still, the photos are great examples of the potential pitfalls of not having an AA filter. The Kodak 14n, which I used for work for more than three years, would show moiré in areas I would never have thought.

I know that it can be corrected in post, but there seems to be so little sharpness difference between having, and not having, an AA filter that using one seems a no brainer.
Logged
"The real voyage of discove

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #2 on: May 10, 2012, 12:53:02 am »

I'm looking forward to the day when we have 300MP sensors which are lens-limited or undergo diffraction softening at almost all apertures and, therefore, don't exhibit moire - not to mention the fact that large files are considerably easier to post-process, when even a 'sharp' edge is several pixels wide.

Until then, an AA filter that can be flipped in or out would be great.
Logged

Bernard ODonovan

  • Full Member
  • ***
  • Offline Offline
  • Posts: 220
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #3 on: May 10, 2012, 07:53:14 am »


My earlier observations of the same test:

http://www.luminous-landscape.com/forum/index.php?PHPSESSID=338d734c5b16eb1699e0d2082cb50a3f&topic=65927.msg527505#msg527505


Personally I would prefer the D800E model.

The higher frequency pattern on the 645 is obviously related to the differences in pixel spacing, distance etc so a different subject could result in a flip round between the non AA models, but it is interesting to note that in that case the higher resolution did not help the 645.

I think the costs of AA are too high on image quality, for higher ISO's the camera is having to work with a messier image and so quality drops off quicker. The D800E holds much better into higher ISO

Incidentally @ ISO 50 the D800 has less DR on the high EV side, but the comparisons I did hold true for ISO 100 and the massive improvement in bright white handling of the D800E which should interest Photogs who probably worry too much about moiré. This should hopefully make them reconsider the cost of having an AA filter to overall imaging capability.

Many thanks



Logged

theguywitha645d

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 970
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #4 on: May 10, 2012, 10:59:36 am »

Interesting. I wonder how much of the moire in the 645D vs the D800s is to do with image scale--the labels are not the same size. And how much to do with the difference in pixel pitch/MTF. Still, after a year of shooting with the 645D, moire is not really an issue for me--I have yet to see it in my images and I have mostly been shooting landscapes and a little portraiture in the studio.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #5 on: May 10, 2012, 11:57:41 am »

Hi,

My guess is that if we see moiré or not depends on the relation of the pixel pitch and the pitch of the pattern in the subject. So if we don't see moiré it doesn't say it will not be a problem. There are also plenty of samples of moiré on OLP filtered sensors.

I'd suggest that the Pentax has higher MTF at Nyquist, most of what I have seen would indicate that.

Best regards
Erik


Interesting. I wonder how much of the moire in the 645D vs the D800s is to do with image scale--the labels are not the same size. And how much to do with the difference in pixel pitch/MTF. Still, after a year of shooting with the 645D, moire is not really an issue for me--I have yet to see it in my images and I have mostly been shooting landscapes and a little portraiture in the studio.
Logged
Erik Kaffehr
 

dturina

  • Full Member
  • ***
  • Offline Offline
  • Posts: 152
    • Picasa gallery
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #6 on: May 16, 2012, 02:05:41 am »

I don't know about the rest of you, bu I can't really see any significant gain in sharpness in D800E over D800. It's all within a margin of error, which probably means that the AA filter is very thin and really cuts in only at Nyquist.
Logged
Danijel

kers

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4388
    • Pieter Kers
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #7 on: May 16, 2012, 05:26:42 am »

I don't know about the rest of you, bu I can't really see any significant gain in sharpness in D800E over D800. It's all within a margin of error, which probably means that the AA filter is very thin and really cuts in only at Nyquist.

For me coming from a D3x I like very much that the D800e takes away the pixel level fuzziness and so adds to detail and the contrast in the detail.
You will see that more clearly if you sharpen the images above a little more.
The benefit is only seen (as the moiré) with very good lenses in the sharp areas of the image, but for me it makes a difference.
From what i have seen in samples about moiré it is not going to be a big problem, but if i would be shooting tissue i would certainly go for the D800
here- a bit oversharpened , just to show what i mean
« Last Edit: May 16, 2012, 05:33:32 am by kers »
Logged
Pieter Kers
www.beeld.nu/la

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #8 on: May 16, 2012, 06:38:49 am »

To put things in perspective, very few landscape photographers ever complained about moiré, and if a photographer doesn't have problems with a large pixel pitch MFD (like old 25 MP backs) it is highly unlikely that he/she would run into problems with the D800E.

I recently got an Aptus 75. I see aliasing effects in almost all of my pictures, however not of the ugly moire pattern type (haven't seen that yet in a landscape picture, but I guess it will come sooner or later :-)), but of the "christmas light" false color type. I have noted though that cameras with AA filter are not immune to this problem either.

Here's a small 100% crop from a nature scene on my aptus (a bit of sharpness lost due to JPG conversion). The magentas and blues in that blåslav(?) (English please) are aliasing artifacts which is partly due to CFA and partly to the lack of AA filter. I've noticed that when something is slightly out of focus or heavily diffracted the same false color issues does not occur, so it is definitely an AA filter thing, although the sharpest demosaicing algorithms does make it worse.

As said this occurs to various extent in almost every picture. I'm not very disturbed by it though for my use, but I'm surprised that this false color issue does not get more reports. If it does not occur often on the D800E it must be because lenses are not sharp enough :-)
« Last Edit: May 16, 2012, 06:41:25 am by torger »
Logged

kers

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4388
    • Pieter Kers
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #9 on: May 16, 2012, 06:52:46 am »

I recently got an Aptus 75. I see aliasing effects in almost all of my pictures, ...

That looks rather nasty indeed.. every day christmas is a bit too much.
The d800E is a made a little different than other cameras without an AA filter, because it is in fact a modified D800. (It has probably thicker glass than usual in front of the cells)
Because of that maybe I have not seen such strong examples in published photographs of this phenomena- also Moiré seems to be less strong than expected- and probably extra sharpness too
Logged
Pieter Kers
www.beeld.nu/la

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #10 on: May 16, 2012, 07:00:24 am »

That looks rather nasty indeed.. every day christmas is a bit too much.
The d800E is a made a little different than other cameras without an AA filter, because it is in fact a modified D800. (It has probably thicker glass than usual in front of the cells)
Because of that maybe I have not seen such strong examples in published photographs of this phenomena- also Moiré seems to be less strong than expected- and probably extra sharpness too

I will be evaluating raw converters more later on too, I use RawTherapee (just got .mos files working, woohoo!) and Emil's "Amaze" demosaicing algorithm is one of the sharpest I've ever come across. But in this case it might be a disadvantage. I shall look into Capture One and see what results I get there. Maybe I'll push more often to f/16 too (the example above is f/11, Aptus pixels are large), the extra DOF wouldn't hurt.

This IQ180 jpeg file is also a good example of color (and pattern) moire http://en.wikipedia.org/wiki/File:JacquelineMegaw.jpg (watch full size on the fur in the in-focus parts to see it. No, it is not visible on the breasts :-) .)
Why fashion photographers would not want an AA filter is beyond me...
« Last Edit: May 16, 2012, 07:05:18 am by torger »
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #11 on: May 16, 2012, 09:09:06 am »

I will be evaluating raw converters more later on too, I use RawTherapee (just got .mos files working, woohoo!) and Emil's "Amaze" demosaicing algorithm is one of the sharpest I've ever come across. But in this case it might be a disadvantage.

I agree about the "Amaze" algorithm, it extracts very sharp images, and increasing its number of false color suppression steps works very well.

Quote
I shall look into Capture One and see what results I get there. Maybe I'll push more often to f/16 too (the example above is f/11, Aptus pixels are large), the extra DOF wouldn't hurt.

This IQ180 jpeg file is also a good example of color (and pattern) moire http://en.wikipedia.org/wiki/File:JacquelineMegaw.jpg (watch full size on the fur in the in-focus parts to see it. No, it is not visible on the breasts :-) .)
Why fashion photographers would not want an AA filter is beyond me...

Perhaps it's not a question of not wanting, but a lack of availability. Not only would the cost of a large AA-filter be significant, I suppose it would also work bettter with retrofocus lens designs (especially the ones which are optimized for having an additional AA-filter in the optical path).

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

torger

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3267
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #12 on: May 16, 2012, 09:21:39 am »

I agree about the "Amaze" algorithm, it extracts very sharp images, and increasing its number of false color suppression steps works very well.

Thanks for the false color suppression tip, that made wonders with the above example. I shall look and see if there are any bad side effects though.
« Last Edit: May 16, 2012, 09:28:05 am by torger »
Logged

dturina

  • Full Member
  • ***
  • Offline Offline
  • Posts: 152
    • Picasa gallery
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #13 on: May 16, 2012, 10:33:11 am »

For me coming from a D3x I like very much that the D800e takes away the pixel level fuzziness and so adds to detail and the contrast in the detail.
You will see that more clearly if you sharpen the images above a little more.
The benefit is only seen (as the moiré) with very good lenses in the sharp areas of the image, but for me it makes a difference.
From what i have seen in samples about moiré it is not going to be a big problem, but if i would be shooting tissue i would certainly go for the D800
here- a bit oversharpened , just to show what i mean

The difference is still incredibly tiny; I can get more difference by setting only slightly different processing parameters in Rawtherapee. The idea that you're going to get some noticeable increase of resolution for landscapes with D800E is, well, overly optimistic. Based on this comparison I'd probably take the D800, just because it's a cheaper version of the same camera. I'm not concerned by moire at all.
Logged
Danijel

Lightsmith

  • Full Member
  • ***
  • Offline Offline
  • Posts: 197
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #14 on: May 25, 2012, 09:40:20 pm »

As Nikon has illustrated in its D800 Technical Guide the Moire effect is most apparent at larger apertures. It becomes a possible trade-off between a small aperture and the image degradation from diffraction or Moire. That said the only time I have encountered Moire is with clothing (ties, lace veils, pin-stripped shirts, etc.). Maybe a herd of zebra clumped together might be a problem but I am hard pressed to think of nature photographs where this might be encountered. In any case by shooting at f8 and smaller apertures the problem is likely to disappear.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Imaging resource samples D800, D800E and Pentax 645D (Moiré)
« Reply #15 on: May 26, 2012, 03:21:03 am »

Hi,

What the test shots illustrate is that Moiré can pop up unexpected. Stopping down beyond optimum aperture can be seen as bad technique. The OLP filter is an advanced optical device designed to reduce aliasing to reasonable levels. Stopping down limits resolution (or rather MTF at Nyquist) becuase of increased diffraction.

Anyway, Moiré indicates that the sensor generates fake detail. Fake detail can actually enhance an image, think of the feathers of a distant bird. If detail in the feathers is below Nyquist you cannot see any detail. Aliasing can create fake detail, at a lower frequency that may be visible and look quite natural.

Best regards
Erik

As Nikon has illustrated in its D800 Technical Guide the Moire effect is most apparent at larger apertures. It becomes a possible trade-off between a small aperture and the image degradation from diffraction or Moire. That said the only time I have encountered Moire is with clothing (ties, lace veils, pin-stripped shirts, etc.). Maybe a herd of zebra clumped together might be a problem but I am hard pressed to think of nature photographs where this might be encountered. In any case by shooting at f8 and smaller apertures the problem is likely to disappear.
Logged
Erik Kaffehr
 
Pages: [1]   Go Up