Ad
Ad
Ad
Pages: « 1 ... 3 4 [5]   Bottom of Page
Print
Author Topic: Nikon D800E - Amazing Resolution ...but "Houston, We Have A Problem!"  (Read 19672 times)
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3868


« Reply #80 on: June 07, 2012, 11:50:06 AM »
ReplyReply

Or perhaps more likely, some of you guys have never photographed standard test charts.  Grin

Well that wouldn't be me then Wink . The color moiré that you see is the product of the regular sensel grid and the laser engraved pattern of the focus screen. Apparently they are both repetitive, but at different spatial frequencies, hence the moiré. You'l see similar effects between the sensel grid and the LCD on the camera, but they are all depending on a certain level of detail magnification, so it's hard to predict how it's going to look, or even if it shows.

Cheers,
Bart
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1704


« Reply #81 on: June 07, 2012, 05:24:03 PM »
ReplyReply

... The aliasing artifacts themselves are not going to help and boost the resolution, also because aliasing is by definition larger than Nyquist.

Two sines, both "fit" with the suggested samples.

If the (potentially filtered) signal that enters the point-sampler is allowed to take the shape of both sinuses, we cannot from the samples decide which it actually was, and therefore we cannot faithfully recreate the waveform.

If we filter the signal, or can make assumptions about the signal such that one of those sinuses are allowed, but the other is not, then we could recreate the waveform from the samples. A given set of samples generally corresponds to an infinite set of possible waveforms, but pre-filtering reduce that set to (ideally) exactly one waveform.

In other words: what you call "aliasing" can be just as relevant information as anything else, provided that the Nyquist criterion is satisfied. Usually, this is done with a lowpass filter that tries to limit frequencies >= fs/2, but it can also be done by a bandpass filter operating at fs/2 <-> fs by making a appropriate test-target if there is no OLPF. In that case we could be able to test the lense performance in another spatial frequency range.

Yet another way to paraphrase: some frequencies wrapping over into other frequency bands does no harm if there are no other signals in that band to interfer with, and if we know which frequencies wrapped where.

There are practical obstacles that I am aware of (no image sensor is a point-sampler, a lense does not have a wavelength-independent PSF), and surely some that I am not. It seems to me that you don't get the theoretical point that I am making, and I think that our discussions would be more fruitful if we are able to agree on those.

-h
« Last Edit: June 07, 2012, 06:04:25 PM by hjulenissen » Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8939


« Reply #82 on: June 08, 2012, 12:12:47 AM »
ReplyReply

Well that wouldn't be me then Wink . The color moiré that you see is the product of the regular sensel grid and the laser engraved pattern of the focus screen. Apparently they are both repetitive, but at different spatial frequencies, hence the moiré. You'l see similar effects between the sensel grid and the LCD on the camera, but they are all depending on a certain level of detail magnification, so it's hard to predict how it's going to look, or even if it shows.

That's probably all quite true, Bart. No dispute there. One has to be at a certain distance from a target in relation to a certain size of repetitive detail, and such distance should also be in accordance with the focal length (or magnification) of the lens, so a bit of experimentation is required in order to see the effect.

However, such color moire and circular artifacts around straight B&W lines on a test chart are also apparent through an optical viewfinder when, presumably, the sensel grid is not involved.

This was the method I used in the old days when my main camera was the Canon 20D, which doesn't have autofocus fine tuning, and when I used to buy lenses from 'bricks & mortar' camera stores and was worried about QC variation and pleased that I could easily return a lens without questions asked, within a week of purchase.

Being aware of differences between artificial and natural, real-world scenes, I constructed a test target for my purposes designed as a mixture of real-world and artificial products. The test target, pictured below, consists of a mixture of Norman Koren B&W line test charts, various real-world textures such as old sandpaper, patches of fabric, bunches of twine, all against a background of the texture of natural timber. I even included a few neutral-grey patches.

I admit it's not a pretty sight to behold, but it served its purpose.  Grin

Concerned now, that what I recall seeing a few years ago was merely a peculiarity of the Canon 20D, I retrieved that dusty, old test target from the garden shed, and hung it on the screen-door-latch of my new house.

I used manual focus with my D7000 and Nikkor 24-120/F4 zoom lens, set at 31mm and F5.6.

This was a quick and dirty shot. No tripod. As you can see, I didn't even crouch down far enough to get the door frame vertical. The point is, this shot was taken when I saw the maximum degree of circular artifacts on the artificial part of the test target. The degree of moire was not as great as I recall seeing several years ago, but that is probably due to the fading of the dye-based test target prints, printed on my Epson 1290 as I recall. If anyone reading this post wishes to try this procedure for accurate autofocussing, I'd recommend printing the line chart on premium gloss, or semi-gloss paper, and placing it in bright light outside.
Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3868


« Reply #83 on: June 08, 2012, 06:15:00 AM »
ReplyReply

However, such color moire and circular artifacts around straight B&W lines on a test chart are also apparent through an optical viewfinder when, presumably, the sensel grid is not involved.

Ah, but they are not just straight lines, they are a repetitive pattern of lines. That periodic signal can aliase with any periodic structure in the focusing screen, indeed no sensor influence required.

Cheers,
Bart
Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3868


« Reply #84 on: June 08, 2012, 07:00:27 AM »
ReplyReply

If the (potentially filtered) signal that enters the point-sampler is allowed to take the shape of both sinuses, we cannot from the samples decide which it actually was, and therefore we cannot faithfully recreate the waveform.

But that's where you lose me, I don't understand what you mean by that. The Blue line is what can be reconstructed if the higher frequency detail is not resolved (beyond Nyquist or low-pass filtered). The blue line could originate from any signal frequency, even real detail with that frequency.

So what is needed is more sub-pixel samples (effectively a higher sampling frequency, although with an area aperture sampling device, not point sampling). It's not the lower frequency alias that helps, but the denser sampling. As said, the aliasing, when identified as such (which may be hard because it looks like any other real detail), can help in pin-pointing the amount of sub-pixel displacement of the samples. However, as the literature references show, the more successful SR approaches are spatial domain and sub-pixel sampling based approaches. PhotoAcute apparently uses a warping variation which reveals clues about the displacement of the sub-samples, and thus doesn't suffer from lens distortions as much as some other solutions. Such spatially variant solutions do require some extra processing power.

Quote
If we filter the signal, or can make assumptions about the signal such that one of those sinuses are allowed, but the other is not, then we could recreate the waveform from the samples. A given set of samples generally corresponds to an infinite set of possible waveforms, but pre-filtering reduce that set to (ideally) exactly one waveform.

Yes, but we cannot recreate the higher frequency signal from the lower frequency one, because it might well be the correct waveform (or an alias from a number of possible higher frequencies). Sub-pixel spatial displacement is a requirement to achieve Super Resolution, aliasing is not (unlike what the Wiki page suggests). In practice we will have a mix of both in our sub-images, because the AA-filters in our cameras are not perfect.

Quote
Yet another way to paraphrase: some frequencies wrapping over into other frequency bands does no harm if there are no other signals in that band to interfer with, and if we know which frequencies wrapped where.

But that's the problem. Human vision is pretty good at picking out aberrant information when it doesn't fit a pattern, but an automatic system has no clue about what to expect or not. All data is seen as relevant, even noise.

Quote
It seems to me that you don't get the theoretical point that I am making, and I think that our discussions would be more fruitful if we are able to agree on those.

Indeed, but I'm afraid I still do not get your point. My point is that aliases and real detail share the same output signal after being quantized. Only by sub-sampling (thus resolving higher spatial frequencies than the Nyquist frequency of a single image) can original detail be identified, and aliasing in the sub-images eliminated. My test above should illustrate that, only where the sub-sampling increased the Nyquist frequency (in the horizontal direction thus mostly benefitting vertical feature orientations) will the aliasing artifacts be reduced.

And this all has only partial impact on single shot lens tests with or without OLPF.

Cheers,
Bart
Logged
Johnphoto
Newbie
*
Offline Offline

Posts: 38


« Reply #85 on: June 08, 2012, 02:49:06 PM »
ReplyReply

I have both cameras and I must say that the D800E is in a class of its own in comparison with the D800. I am getting edge effects between dark and bright areas on the D800 files. It shows like an extra contour of 2 pixels around rocks contrasting against a bright sky behind it. It is very significant and disturbing. No such effects are appearing on the D800E. I believe that it comes from the resharpening of the image after the anti aliasing? Do you guys have any other idea?
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1704


« Reply #86 on: June 08, 2012, 03:52:34 PM »
ReplyReply

Yes, but we cannot recreate the higher frequency signal from the lower frequency one, because it might well be the correct waveform (or an alias from a number of possible higher frequencies).
If the test chart is guaranteed to not contain the low frequency one, then we can be pretty certain that such a frequency did not enter the camera sensor, dont you think?

If we have en equation with N solutions, if we can eliminate (N-1), then we are left with only one.

-h
« Last Edit: June 08, 2012, 03:57:33 PM by hjulenissen » Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3868


« Reply #87 on: June 08, 2012, 05:58:42 PM »
ReplyReply

I have both cameras and I must say that the D800E is in a class of its own in comparison with the D800. I am getting edge effects between dark and bright areas on the D800 files. It shows like an extra contour of 2 pixels around rocks contrasting against a bright sky behind it. It is very significant and disturbing.

Hi,

My first guess based on you description would be that it's a sharpening artifact, because the sensor of the D800 doesn't produce such contours by itself. It's hard to do any better than a guess without an actual image sample/crop that illustrates your issue.

Cheers,
Bart
Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3868


« Reply #88 on: June 08, 2012, 06:05:22 PM »
ReplyReply

If the test chart is guaranteed to not contain the low frequency one, then we can be pretty certain that such a frequency did not enter the camera sensor, dont you think?

That's correct, a high pass filtered target projection, or one only containing frequencies above Nyquist, will produce only aliasing artifacts or a grey image.

Cheers,
Bart
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1704


« Reply #89 on: June 10, 2012, 02:27:50 AM »
ReplyReply

That's correct, a high pass filtered target projection, or one only containing frequencies above Nyquist, will produce only aliasing artifacts or a grey image.

Cheers,
Bart
Which brings me back to the starting-point for our discussion:
Wondering if lensezone.de will use a D800E to test new lenses. Aliasing can be used as a measure of lense sharpness if the input signal is known beforehand.

-h
If the test-chart contains energy from fs/2 to fs, and the sensors response to such signals is known, then any attenuation should be due to lense-flaws. (assuming perfect focus, no camera movement etc)

-h
« Last Edit: June 10, 2012, 02:29:47 AM by hjulenissen » Logged
Pages: « 1 ... 3 4 [5]   Top of Page
Print
Jump to:  

Ad
Ad
Ad