color artifacts produced by the camera

[hatsnp (NP)]

@Tamas_Papp

Correct, but other things can. AFAIK some cameras with IBIS can move the sensor just a bit to simulate a low-pass filter. Maybe Sony?

Canon Patents IBIS-Powered AA Simulation Tech, Similar to Pentax

This?

[Yogansh_Bhatt]

Smartphones. Some sensors produce a lot of it but because of quad/nona or whatever bayer they have, it is pretty common. Some sony sensors have it a bit more than Omnivision but it’s pretty bad on almost all modern smartphones. Maybe because of the different remosaicing they use. You don’t know it until you’re in dt
You might have a ton of examples if you already shoot RAW on your smartphone.

[priort (Todd Prior)]

Remove moire in ART

Hi everyone, Have you some tricks to remove moire in ART ? Thanks

Demosaicing issue or is something else going on as well?

I’ve been getting familiar with my new camera and ran into something that looks like a combination of discolouring and a moiré pattern in one of my images. After some digging and trying I do believe that this is a false colour demosaicing issue, one close to what is described on RawPedia Here’s a 100% zoomed in view: [false.colour] Switching to AMaZE or AMaZE+Bilinear partially resolves this, even with the false colour suppression steps set to 5 there’s still a remnant of the green visible: …

Moiré nightmare challenge from hell

After some discussion and kind explanations on the IRC channel, I submit to you, the impossible moiré challenge from hell. 20201213_0022.NEF (29.7 MB) I tool this picture in Montpellier, it’s the city town building. I wanted to get out of my comfort zone with some modernist geometries, and once home I notices something very strange in some areas, see if you can spot it: [20201213_0022] The whole thing is a moiré festival, but here is a cut of the most interesting part: [20201213_0022_zomm…

There are raws included in these posts but you might have them already…

[hannoschwalm (Jens-Hanno Schwalm)]

Yes - the city building image is my demosaic nightmare

[leonidas111]

Moiré is following me. Not even a week after the last one!
Tried some diffusion combined with chromaticity blending and I can live with the result, for now.

[Iain]

@hannoschwalm Here are some of my experiments with moiré.

Demosaicing Experiements

I’ve been playing with demosaicing, and since my shooting style seems to promote moire and aliasing I have been especially interested in reducing it. Here it is compared to LMMSE. LMMSE is on the left. [image] it works by first, demosaicing the image twice. On one the green channel is interpolated horizontally, the other vertically. Then the G-R and G-B is interpolated. This gives two very different moire patterns for each image. Then I choose which part of each image has the least chroma …

Detect Moire Filter

I’ve just added a “Detect Moire” filter to Testing → Iain Fergusson. [Detect_moire] It’s a problem I’ve been trying to solve for a while and I was inspired to revisit it by the recent work done to combine demosaicers in Raw Therapee. It’s pretty good at detecting moire on low ISO images, high ISO images require a bit of fiddling to make sure noise is not causing false positives.

New Demosaicing filter

I created a simplified GMIC filter script to isolate the moire removal code It’s conceptually simple and I have commented to explain most of it. Here is the overview, more details in the link. moire_remove_only: bayer_pattern=$1 tile_size=$2 -rgb2bayer[-1] $bayer_pattern,1 # convert greyscale raw image into RGB # demosaic by interpolating green horizontally only --moire_remove_horizontal[0] $bayer_pattern,1 # demosaic by interpolating green vertically only -moire_remove_vertical[0] $bayer…

I have a collection of images with bad moiré I can hunt down if you want them.

[hannoschwalm (Jens-Hanno Schwalm)]

1. Yes - i would like a “moire-collection”.
2. I had a quick look through above links, i now remember those. Not sure how those fit into dt demosaicing model. Maybe we share ideas via mail? Would be nice to start a “new project”

[snibgo (Alan Gibson)]

No one seems to have mentioned that some cameras (eg Nikon D800) have a built-in anti-moiré filter. This introduces a slight blur at the sensels, after the light has been filtered by the Colour Filter Array.

This suggests that a digital fix might be to slightly blur the grayscale image before it is de-mosaiced. I haven’t experimented with this as I use a D800 which doesn’t suffer moiré.

It seems to me (I could be mistaken) that more recent cameras don’t have anti-moiré filters.

[afre (Alan)]

@snibgo

This suggests that a digital fix might be to slightly blur the grayscale image before it is de-mosaiced.

I believe demosaicing algorithms already do this to a certain extent but are hard coded and therefore not adaptable to specific moiré control needs.

@hannoschwalm

Would be nice to start a “new project”

Look forward to the next amazing collab.

[Tamas_Papp (Tamas K. Papp)]

@snibgo

This suggests that a digital fix might be to slightly blur the grayscale image before it is de-mosaiced.

There is no “classical” signal-processing fix for moiré (classical here means non-ML). The signal is literally lost in the frequency domain, you cannot reconstruct it.

You can only mitigate the phenomenon at the cost of degrading the information content of the image. Sometimes this works fine though without apparent loss in image quality, but it is always a judgement call.

@snibgo

recent cameras don’t have anti-moiré filters

Some do, but they are weaker than they used to be, because of the obsession with “sharpness”. Manufacturers don’t release detailed specs about this, because the armchair photographers will immediately denounce the camera at the slightest hint (Canon seems to be immune to this. Good for them). “Having no AA filter” is actually a considered an advantage by pixel peepers.

To be fair, the higher the resolution, the less likely you are to encounter moiré. When sensors were 6-8 Mpixels it was much easier to run into.