What is this pattern?

[Stagecoach]

The below image has had its 'exposure' increased by +3 in ACR. No other adjustments including sharpening or noise reduction in ACR.

In PS, capture sharpening was applied at 0.3 300% where the pattern just became visible (no other adjustments made). In applying capture sharpening again at 0.3 300% it became far more prominent.

I came across this when I was playing to see if a few images could be made usable, what has caused this pattern?

1/200, f/5, ISO1600


Grahame

[xpatUSA]

The below image has had its 'exposure' increased by +3 in ACR. No other adjustments including sharpening or noise reduction in ACR.

In PS, capture sharpening was applied at 0.3 300% where the pattern just became visible (no other adjustments made). In applying capture sharpening again at 0.3 300% it became far more prominent.

I came across this when I was playing to see if a few images could be made usable, what has caused this pattern?

1/200, f/5, ISO1600


Grahame

My guess is internal reflections off the microlenses to the lens elements and back.

Only a guess, just sayin' ...

[Dave Humphries]

Hi Grahame,

EXIF shows Nikon D800, is that correct?

High ISO, but then I see a street lamp lit, so presumably it was still quite dark (EXIF says 06:30)

Was this a 'straight' capture?
e.g.
Not shot through any glass (like a car window)?
Nor a reflection shot in a mirror?
Any filter on the lens?

How hot was it?

Thanks,
Dave

[Stagecoach]

Hi Grahame,

EXIF shows Nikon D800, is that correct?

High ISO, but then I see a street lamp lit, so presumably it was still quite dark (EXIF says 06:30)

Was this a 'straight' capture?
e.g.
Not shot through any glass (like a car window)?
Nor a reflection shot in a mirror?
Any filter on the lens?

How hot was it?

Thanks,
Dave

Hi Dave,

It was the D800 with a Tamron 28-75mm.

It was shot 10 minutes before sunrise, the sky was just starting to show dark cloud definition but still quite dark.

It was a straight shot, not through glass or reflected, there was a UV filter fitted.

Temperature was 'cool' around 18 deg C.

The UV filter is not a 'well known brand' but has always been on this lens since purchase and I have never had any concern about it (but my images are not normally exposed so poorly). I will do some tests with and without it on a -3 'underexposed' image for elimination purposes.

The shots, 4 off, were part of a burst sequence where the flash had failed to fire.

[Manfred M]

This definitely looks like a moiré pattern. I wonder if Ted is onto something and we are seeing the effect of the Bayer array and microlenses interacting. Patterns like these need two different patterns in close proximity for it to show up.

[proseak]

I've seen this before once or twice. It was caused by raising the exposure in ACR of a heavily underexposed pic.

I used DxO to create a .dng which didn't show the same effect.

[Stagecoach]

Further investigation;

I have now taken some underexposed (-3 in camera) shots at ISO100/1600, with/without the filter fitted and also some correctly exposed shots of the same scene.

What I can confirm is that I can replicate the 'pattern' on all of the images, including the well exposed ones.

It appears that repeated actions of 'Capture Sharpening (0.3 300%)' in PS are causing it and it has nothing to do with the fact that the image was underexposed. But applying a single action of 'Capture Sharpening' at 0.3 max 500% does not produce the pattern.

Repeated actions of 'Creative Sharpening (4.0 40%)' will not replicate the pattern.

Edit : Further test has confirmed that it is not lens specific.

[James G]

I've seen this before once or twice. It was caused by raising the exposure in ACR of a heavily underexposed pic.

I used DxO to create a .dng which didn't show the same effect.

I too have encountered this pattern (I shoot with Canon kit), and would agree with Peter that is associated with a significantly underexposed image.

Speciofically, I have got it when processing night sky/astrophotography images and 'pushed' the exposure up too much while trying to bring very faint objects more into the image. From memory it is not mitigated when images are stacked, appearing when the image is 'pushed' in later post processing.

I 'assumed' it was something along the lines that Manfred suggested. I am slightly surprised to see it in what is essentially a dawn/twilight shot.

That said, Grahame's experiment attributing it to 'sharpening' does fit in respect of the nightsky shots too.

[Stagecoach]

Peter, James

The results here suggest this pattern is not due to an image being underexposed

No 1


No 2


Grahame

[James G]

Grahame I agree, I had only experienced this with my night shots..

I've checked back and had a quick 'play' with a single capture from Dec 28 2017 when I was shooting the moon and Orion.

I will post some images later (TinyPic not working or really slow )

In summary though,

1) Working in Adobe ACR I increased the exposure on a raw frame (+3) but could not really see the 'pattern' in the background (lots of noise though). When I converted to jpeg, the pattern became obvious though quite faint!

2) Still in ACR, I then 'sharpened' the image and the pattern became very clear, in line with your experience.

3) Moving the sharpened image into Photoshop CC and further sharpening emphasised the pattern further, not a surprise at this stage.

To be honest, given the nature of the capture and knowing I was pushing my kit, I assumed that I needed to go for longer exposure times during capture... which did/does work .

I've also checked some images I took at sunset a couple of weeks back, (Moon high in the sky with a tanker on the near horizon, at sea), to see if I could replicate but no 'success'. The image was underexposed by 2 stops since I was deliberately bracketing at the time.

My images were captured with a Canon 5DsR which has no high pass fFilter, but I have seen this same patterning before with my 7Dmk2 which does have one, (also when taking night shots).

Trouble is I don't think it 'proves or disproves' Manfred's suggestion it is a moire effect since I has shooting at the 'limits' of each camera's capability.

One thing does seem certain though, ife the pattern presents, post processing only makes it worse

[odds]

...The results here suggest this pattern is not due to an image being underexposed

Some years ago Michael Reichman demonstrated (in an article on Luminous Landscape) how a moire or moire-like pattern could show in an image displayed on a computer screen even if there was no moire in the capture. If I remember correctly, it was a combination of hardware pixel density or resolution on screen and in the image file.

I wonder if the pattern you get will show up in print.

Edit: For those who have access to Luminous Landscape (it is now a pay site) see https://luminous-landscape.com/nikon...l-impressions/ The demonstration of screen-moire does not work anymore, clicking on the demonstration image returns a 404 Not found error instead of the second image, but the explanation seems intact.
--
Odd S.

[davidedric]

Not really following the thread (not relevant to me), but I'm pretty sure that LuLa is accessible to non-subscribers - you just can't post.

Dave

[Manfred M]

Some years ago Michael Reichman demonstrated (in an article on Luminous Landscape) how a moire or moire-like pattern could show in an image displayed on a computer screen even if there was no moire in the capture. If I remember correctly, it was a combination of hardware pixel density or resolution on screen and in the image file.

I wonder if the pattern you get will show up in print.

Edit: For those who have access to Luminous Landscape (it is now a pay site) see https://luminous-landscape.com/nikon...l-impressions/ The demonstration of screen-moire does not work anymore, clicking on the demonstration image returns a 404 Not found error instead of the second image, but the explanation seems intact.
--
Odd S.

I do feel that Michael was onto something with respect to moiré, but he appears to be looking at a slightly different manifestation; where a photographed pattern reacts with the pattern of pixels on a computer screen. Here resizing the image even slightly will make this effect go away. I see moiré patterns quite often in the Photoshop opening screen that displays large thumbnails.

Grahame's examples are a bit different. The pattern is fixed and repeating, so that suggests something different is happening. The pattern is virtually identical in all his examples, which suggests that something is happening in-camera. The only repeating patterns that are in close proximity are the Bayer filter and the sensor elements themselves. The effect seems to only become pronounced when a specific software setting is used in sharpening. As sharpening enhances the local micro-contrast where dark and light regions in the image meet, so one should not be surprised that a particular sharpening setting enhances the issue.

[Manfred M]

Not really following the thread (not relevant to me), but I'm pretty sure that LuLa is accessible to non-subscribers - you just can't post.

Dave

Not anymore Dave. Lula went subscription a number of years back and only subscribers can access the articles. At $12US per year it is a real bargain, even as a subscription only site.

[Stagecoach]

I have come across an image moire affect before where it comes and goes when small changes are made to the image screen size.

If I have time today I'll see if a D300 image does the same and also see if I can 'narrow' down how specific it is to a sharpening settings range.

[James G]

"Curioser and Curioser" as Alice did say.

I've just noticed a 'similar' background pattern when I'm pre-processing images using Adobe Camera Raw.

I got a lot of dust bunnies on my sensor recently when I was at an airshow. I'm slowly working my way through the captures and I am using the spot removal tool to execute the bunnies ! What I just noticed when I use the 'visualise' option is a pattern to the noise grouping in the background as shown here :-

[IMG][/IMG]

So? maybe something associated with the way sharpening masks are generated?

Oh yes, and I've had the sensor cleaned since the Old Warden shoot!

[Stagecoach]

This pattern of mine has opened a right can of worms

Following on from my post 9 where I was convinced that I could replicate this pattern on any image, whether severely underexposed or exposed perfectly, by adding successive amounts of sharpening in PS at 0.3 300% (typical capture sharpening amounts but not something you would normally do) I opened another image and could not make it appear.. Unfortunately I then had to spend the day under my wagon doing what I had hoped would be an easy job

So, further investigation this evening has revealed the following;

a) If I open an image in ACR that has the "Enable Profile Corrections" Checked in the "Lens Corrections"/Profile tab I can replicate the pattern on any image.

b) If I open an image in ACR that has the "Enable Profile Corrections" Un Checked in the "Lens Corrections"/Profile tab I can not replicate the pattern on any image.

c) If I open an image in ACR that has the "Enable Profile Corrections" Un Checked in the "Lens Corrections"/"Profile" tab and go to the "Manual" tab and put an amount of + or - in from the default 0 with the 'Distortion Correction' slider I am able to replicate the pattern when opened in PS and repeatedly sharpened at 0.3 300%.

A couple of weeks ago when I quick edited some 500 images I did them solely in ACR and at that time set new 'defaults' that included camera profile corrections/CA removal which I normally have turned off.

So, it appears that the "Distortion" function within the Lens Profile Corrections is determining whether this pattern can be made visible or not. I am getting the same with profiles for both the 28-75 and 70-200 which ACR automatically recognises.

Now to delve into lens profiles

[billtils]

This pattern of mine has opened a right can of worms

Logically investigated and nicely solved Grahame.

[DanK]

Grahame,

Very helpful. Thanks for sorting this out and posting. In hindsight, the patterns in your images would seem to fit with that.

This will make me a little more careful in using profile correction, which I often do without much thought. Up until now, the one reason I had been a bit careful is that if an image taken with a lens that vignettes substantially is underexposed in general, the profile correction includes a lot of brightening of the corners, which can bring out noise. I don't know Nikon lenses, but of all of my Canon lenses, the only one that has enough distortion and vignetting that I use profile correction often is my 24-105, which is unfortunately also my most-used lens.

I wonder if the profile corrections in other software do something similar.

Dan

[Stagecoach]

It will be interesting now if someone can explain the 'technical' reasons why this is happening when distortion correction is applied

In ACR there is a slider that allows you to adjust between 0 - 200 the amount of distortion correction applied when "Enable Profile Corrections" is selected.



A test of various settings of the slider revealed the following;





The default setting for the slider is 100.

I have also confirmed this is all applicable to images that have been taken with my D300 and using other lenses.