wtf even is infrared photography and why are my photos so messed up looking

cee May 2, 2026
Source

I impulse-bought a full spectrum camera a month ago and I've been really really enjoying using it to take pictures of things I'm normally not interested in. I'm usually a nature and animal guy but I'm doing a lot of city photography lately, and I'm getting a lot of questions, concerns, and comments, so I figure I'll write something to answer all these.

a collection of responses ive gotten after posting IR photos

Infrared photography is when a photo is taken with a camera that can capture infrared light that the human eye can't see under normal circumstances. Usually this means using a camera that has been "converted" to full spectrum - a process that involves taking the camera apart and replacing its built-in infrared filter with a new piece of glass that doesn't block infrared. This kind of renders the camera useless for normal-looking photos, but it enables some really cool things to be captured. When reading about infrared photos you'll hear people mentioning the "nm" of the filter they're using, referring to the wavelength of light the camera's filter is letting in. Usually something like "590nm" (some visible light reaches the camera), "720nm" (almost no visible light), or even up to 800-900nm. The number here generally refers to the shortest wavelength that passes through the filter, so 590nm allows everything "redder" than 590nm.

Cool whatever fine i get it, but why do your photos look like that

The photos I usually post are what I would call "590nm Red/Blue channel swapped", and theres some cool unexpected stuff that makes the images look Like That, so I guess let me explain some stuff and then I'll actually demo how they work together to make the cool colors.

The 590nm Filter

When you see people talk about "nm" in IR photography, they're usually referring to a filter that blocks every wavelength of light shorter than that value (i.e. cutting out any color toward the blue end of the spectrum). In this case, I have a 590nm filter

Photo of me holding a Kolari 590nm filter up to the window, showing the red colored glass

590nm is golden-yellow on the color spectrum, so the filter allows anything redder than yellow through it. Yellow, orange, red, and infrared. Crucially, this means that the data reaching the camera sensor isn't just infrared, it's yellow-ish, red, AND infrared. The full spectrum sensor and the filter interact in two different ways:

  1. Visible red light hits the Red sensor pixels
  2. Infrared light hits all three color channels equally

Because of the way the camera sensor is designed for visible light only, IR light penetrates the dyes of all three sensor channels and the camera sees IR as "bright", not any specific color. And, because the 590nm filter I'm using lets in red visible light, the red channel receives some information. This small slice of visible red light is the key component that creates the deep blue color of leaves and clothing in unedited 590nm photos.

Blue? Wouldn't it make the images red?

This is a cool trick that needs two different things to happen: the camera's white balance algorithm, and the light information that a leaf actually sends through the filter.

White Balance

White balance is pretty simple, and modern photo software is good enough that you almost never have to think about it. A camera's white balance algorithm tries to detect what multipliers it needs to apply to each RGB channel to make some specific off-white color look white in the resulting photo. In IR photography, we can exploit this to create bright and varied colors, even though the light reaching the sensor is basically entirely red.

Let's say we're outside with the full spectrum camera and we set the white balance to a white sheet of paper with the 590nm filter on. The full spectrum camera sees the light coming from the paper as:

  1. Visible red light
  2. Infrared light hitting R, G, and B channels evenly

Basically the camera sensor detects a high brightness, very red color. In order to shift the total color balance of that to a white color, it determines that green and blue should be boosted to equalize the RGB values. Every photo from now on with this WB setting will have blue/green heavily boosted.

Seattle space needle before and after WB. Before showing deep red color cast, After showing blue leaves and yellow sky

So why are leaves blue now

Let's talk about what sorts of colors leaves and grass send to our camera.

Imagine we take a picture of a green leaf. The stuff (chlorophyll and water mostly) in a leaf absorbs red and blue, so only green and infrared reach the filter. The 590nm filter then cuts the green light, so the sensor ends up seeing no red and lots of IR. Since IR light hits all three sensor channels evenly, the camera sees leaves as pure white. After the camera applies white balance (boosting green and blue), the leaves turn a bright blue-cyan color.

spectral curve showing 590nm to NIR range of green foliage, with a cyan preview color

(btw you can play with this simulator at https://spectral.cee.wtf/)

The same concept applies to many other objects like clothing or paint. All this stuff results in a simple fact: objects that show up as blue in the raw 590nm infrared photo have bright infrared, and low visible red.

Yellow is also captured in certain cases. A clear blue sky has a large peak in the blue end of the spectrum, tapers off through yellow, red, and drops off almost entirely in the IR range, producing an olive color after the WB is applied

Blue sky spectral curve, showing the blue peak and the drop off through red and IR

Channel Swap

ok so I have a photo through the 590nm filter, white balance is calibrated to a neutral grey, and I take a picture of some stuff. It's a weird-looking mix of colors - Blue trees, blue highlights, sometimes a pink shirt, and a dumb yellow poop-colored sky.

i90 bridge in 590nm after white balance, with no channel swap, showing blue foliage and olive-colored sky and water

The blue trees are really cool to me, but the gross sky color ruins it, so I play around with the color channels. By swapping color channels in the raw image, I can produce some more natural looking photos without destroying any data in the image. The most common channel swap you'll see is swapping the color values of Red and Blue pixels, creating an image with red or orange highlights and a cyan or blue sky.

i90 bridge in 590nm Red/Blue channel swapped, showing orange foliage and blue-cyan sky

Its worth noting here that I actually really don't like editing my normal photos' colors at all because I think it makes the image unrealistic and more of a digital painting than a photo, but since IR photography lets me work with light that has no real representation in real life, color shifting is cool to me! Its fun for me to decide one color swap to apply to a whole set of photos to represent the vibe, although I almost always go for yellow or orange foliage

i90 bridge in 590nm Red/Blue channel swapped, showing orange foliage and blue-cyan sky

Discussion in the ATmosphere

Loading comments...