Weird color noise(pattern?) on X100T

[DrCastle]

Hi. I've been using my X100T for about a year by now, and only today when I was checking out my photos I noticed that there's a weird color noises on certain regions of a photo. Please check out the photos below





original photo. f8, iso 200



It looks just fine in this way, but when I see it in full screen on Photos some patterns are visible.




cropped 1. See that wobbly wave-kind of colored stripes?

Cropped 2. That color noise is very apparent on that yellowish wall, and if you look close enough, yellow/blue patterns are seen on those bars too.


cropped 3. need I say more?



The weirdest part is that none of those annoying color patterns are seen in other dark regions of the photo. At first I thought it was just a normal color noise, but then I realized there can't be a color noise on a photo shot in iso200 on a bright daylight.

See? No color noise. So what's the matter with all those patterns/noises on other regions?

I kind of figured that they arise in the surfaces with a fine sift(grid? I don't know) pattern, but i can't understand it any further. Can you explain what's wrong with this situation?



Thank you in advance. Sorry if I was being unclear in some parts.

 

[mbaginy]

That noise is moire. You can adjust, reduce or delete that in post processing.

By the way, all these images were shot with a shutter speed of 1/10,000! Wow! Did you have a special reason for such a shutter speed?

--
Mike
www.mikebaginy.weebly.com

 

[tesilab]

Yes, Fuji told you (or at least implied)that Xtrans will give you sharp images without moiré patterns.

The reason all those cameras used to universally have anti-aliasing screens was to deny you that last bit of resolution so that extremely fine lines in your image wouldn't happen to land on just the green (or red or blue) sensels as they inevitably must.

 

[Compact40]

You can also avoid it by changing your distance from the subject. But the problem is that you usually don't realize the problem until you see it on the computer. But you can learn what sort of subject is prone to moire (fine textures like wires, fences, sweaters), so you could be careful.

 

[nixda]

Yes, Fuji told you (or at least implied)that Xtrans will give you sharp images without moiré patterns.

Fuji said the X-Trans "minimizes generation of both moiré and false colours". They don't claim that it eliminates moiré. It is Fuji's assessment, though, that it eliminates the need for an AA filter.

The reason all those cameras used to universally have anti-aliasing screens was to deny you that last bit of resolution so that extremely fine lines in your image wouldn't happen to land on just the green (or red or blue) sensels as they inevitably must.

Yes, moiré is possible with the X-Trans too, and that's been known from day 1, but one doesn't get it quite as readily as with the standard CFA.

 

[tesilab]

Yes, Fuji told you (or at least implied)that Xtrans will give you sharp images without moiré patterns.

Fuji said the X-Trans "minimizes generation of both moiré and false colours". They don't claim that it eliminates moiré. It is Fuji's assessment, though, that it eliminates the need for an AA filter.

Yes of course. But I see a lot of Fuji boosters claim that it just isn't a real problem on xtrans.

The reason all those cameras used to universally have anti-aliasing screens was to deny you that last bit of resolution so that extremely fine lines in your image wouldn't happen to land on just the green (or red or blue) sensels as they inevitably must.

Yes, moiré is possible with the X-Trans too, and that's been known from day 1, but one doesn't get it quite as readily as with the standard CFA.

Those images seem to contradict you. It is a classic example of producing moire readily. All it takes are some fine lines captured at the correct angle and boom. I'd rather have the bayer CFA pattern, if I have to worry about moire anyway. (Too bad the XA is too cheap a camera and the Fuji GFX is just too expensive).

Sony and Pentax have both experimented to alternative approaches to dealing with the tradeoff. My RX1RII has an electronic LPF (AA filter) with three settings, Off, Medium, High. There's a Pentax with ultrafine IBIS induced vibration to effectively blur the exposure slightly when enabled. One of these approaches, their descendants, or some other solution (a foveo-like successor?) is more likely to gain traction and acceptance against moire than XTrans.

 

[BurpeesAreHard]

Can you explain what's wrong with this situation?

I wouldn't suspect the camera as the source of the moire patterns. Instead, I'd be looking at the resize algorithms that the image went through in post processing.

 

[DrCastle]

No in my phone the Metapho app shows it's shot in 1/420s! I don't know what went wrong when I uploaded it

So it's a moire? Interesting. Thank you!

 

[DrCastle]

So it can be avoided if I carefully convert it from the raw file? Thanks!

 

[Yake]

Adobe's Lightroom has a brush to help eliminate moire.

 

[Yake]

Can you explain what's wrong with this situation?

I wouldn't suspect the camera as the source of the moire patterns. Instead, I'd be looking at the resize algorithms that the image went through in post processing.

The camera is always the source. Use the moire brush in Lightroom to eliminate the moire.

 

[BurpeesAreHard]

So it can be avoided if I carefully convert it from the raw file? Thanks!

Do you have any straight-out-of-camera files from these images? Either raw files or in-camera generated jpgs?

 

[BurpeesAreHard]

Can you explain what's wrong with this situation?

I wouldn't suspect the camera as the source of the moire patterns. Instead, I'd be looking at the resize algorithms that the image went through in post processing.

The camera is always the source. Use the moire brush in Lightroom to eliminate the moire.

In this old example: https://www.dpreview.com/forums/thread/3127070

nobody agrees with you.

 

[tesilab]

Post processing is not the most effective strategy to deal with moire patterns. At best you can use the tools to attempt to mitigate the issue.

If you don't have an adequate low pass filter on your sensor, your best bet is to attempt to avoid exciting moire in the first place by not focusing on very fine patterns. For example, if you are shooting pictures of a bride, avoid focusing on the veil.

 

[Yake]

Post processing is not the most effective strategy to deal with moire patterns. At best you can use the tools to attempt to mitigate the issue.

If you don't have an adequate low pass filter on your sensor, your best bet is to attempt to avoid exciting moire in the first place by not focusing on very fine patterns. For example, if you are shooting pictures of a bride, avoid focusing on the veil.

The OP is shooting buildings, not brides. So that strategy won't work.

 

[baobob]

Do not use ES when unnecessary, in the examples shown what is the point of 1/10000sec?

I have been using X trans sensors now for many years and I can count on the fingers of one hand the occurence of moire actually 2 in many 10000 shots during 4 years...

 

[nixda]

Do not use ES when unnecessary, in the examples shown what is the point of 1/10000sec?

I have been using X trans sensors now for many years and I can count on the fingers of one hand the occurence of moire actually 2 in many 10000 shots during 4 years...

When the EXIF is incomplete for an image and does not contain a value for the exposure time, DPReview shows a default exposure time of 1/10000s, but the actually used exposure time is very likely something else. The OP mentioned it was in fact 1/420s.

 

[sankos]

Sony and Pentax have both experimented to alternative approaches to dealing with the tradeoff. My RX1RII has an electronic LPF (AA filter) with three settings, Off, Medium, High. There's a Pentax with ultrafine IBIS induced vibration to effectively blur the exposure slightly when enabled. One of these approaches, their descendants, or some other solution (a foveo-like successor?) is more likely to gain traction and acceptance against moire than XTrans.

Additionally, Olympus and Pentax have the pixel-shift feature (a variation of image-stacking) which gets rid of moire and gives you overall greater resolution.

 

[sankos]

Also, you can stop down your lens to something like f/11 and more if taking landscape photos -- the lens diffraction which sets in with small apertures can act as a sort of AA filter.

--
Marcin

 

[baobob]

Thx for info, a very odd default value isn't it ?

Anyway most surprised of such moiré I never had such patterns in my long (and huge in terms of shots) Xtrans experience (precisely the X trans pattern is supposed to minimize moiré...)