When patterns attack

This entry was posted in Digitization, Work Flow and tagged , . Bookmark the permalink.

Up until about a month ago, I’d never heard the term moiré pattern, but I’d definitely seen them in action. You probably have, too. It might’ve looked something like this…


…or maybe this…


A moiré pattern, which may appear as shimmering lines or dots or even swirls, can manifest itself when two similar but not identical patterns are superimposed or two identical patterns are superimposed at an angle. (It’s the visual equivalent of what happens when you’re driving in the car, listening to music, and you hear the song’s beat begin to sync up with your turn signal…but just for a moment.) In images, what causes the interference is not easy to spot — because one of the patterns in question comes from the workings of the camera itself.

A digital camera has a particular pattern to how it senses color. Reds, greens, and blues are captured and translated into pixels in a grid-like Bayer filter pattern. When that sensor grid in the camera lines up in a certain way with a pattern in the object you’re photographing — the pattern on someone’s shirt, for example, or simply the pattern of dots that manifest in offset printing — strange things can happen.

The following images show a black grid partially laid over a Bayer filter pattern.

Black grid of same scale as the Bayer pattern:


Black grid a bit larger in scale.


We’ve found that moiré patterns don’t creep into our camera-based digitization very often — they wouldn’t show up in handwritten materials or photographs, the bulk of the archival material we capture — but when they do, they can be a real problem.

We’ve developed some methods of dealing with the psychedelic-looking color patterns intruding upon our workflow. There are actions we take during the capture process that normally combat the problem pretty well, but we also have a couple of tricks up our sleeves for correcting the problem in post-production. Full disclosure: we exclusively run into problems based in the color channel, rather than the brightness channel (we get dot patterns rather than swirls), so that’s what these methods are designed to address.

During Capture

The front-line defense for making the pattern disappear while a digital image is being taken is to change the relationship of the object and camera. Hopefully, this will bring about a different relationship between the object’s pattern and the camera’s.

1. Orientation: Rotate the object until a test shot or preview shows that the pattern disappears. Here are the grids from example one above, set at a 45 degree angle from each other.


2. Distance: Move the camera or the object so that they are at least a few inches closer or further apart.

After Capture

While it’s better to nip the problem in the bud before you take the picture, there are a couple of ways to improve — but not entirely fix — the image after it’s been made.

1. Photo editor tools: Use a moiré correction tool if your photo editor has one. We use Adobe Bridge’s camera raw editor to crop and rotate images, and we found that it has a Moiré Reduction tool suitable for dealing with minor cases.

2. Manual RGB manipulation in Photoshop: I won’t pretend to describe this process here, but it involves manipulating the values of the RGB channels, and it must be done on a case-by-case basis. It’s not something that makes sense for the average digitization project’s workflow.

There is no one simple fix for moiré interference patterns. As you photograph things with a digital camera, be on the lookout for that tell-tale shimmer of color, and see what you can do to re-position the object as you’re shooting it. Hopefully, you won’t have to bring out the big, complicated guns. Avoiding combat in this case really is half the battle.

This entry was posted in Digitization, Work Flow and tagged , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *