Make a red and cyan 3D photo

The red and cyan channels pull apart by depth, so bright parts of the picture float forward through a pair of old 3D glasses. Worked out from your photo, no AI.

Open the editor with this effect

What you get

Made with plain math, not a model

There is no generative model behind this look and nothing was trained on anyone's work. It is a short list of fixed steps running in your browser.

  1. Read the photo: Your photo is read pixel by pixel inside your browser. It is never uploaded, so nothing leaves your device.
  2. Red and cyan split: The image is split into a red copy and a cyan copy, shifted apart by depth so that one plane stays lined up and the rest fringes, like an old 3D print.
  3. Same input, same pixels: Every step is fixed math with fixed seeds. Run it again tomorrow on the same photo and you get exactly the same image.

Red and cyan fringes on every edge, and a picture that seems to lift off the page through a pair of cardboard glasses. That is an anaglyph, the oldest cheap way to show a 3D image, and it still turns up on posters, comics and album covers. Bitgrain builds the look from a single photo of yours with plain image maths, no AI and no upload.

How a real anaglyph works

Your eyes see the world from two points a few centimetres apart, and your brain turns the small difference between the two views into depth.

An anaglyph prints both views on top of each other in different colours. The left eye view goes into the red channel, the right eye view into green and blue, which together make cyan.

Glasses with a red lens over the left eye and a cyan lens over the right let each eye see only its own view, so the brain puts the depth back together.

The method dates back to the 1850s and became famous with 3D comics and films in the 1950s. Green and magenta glasses came later and give slightly truer colour.

How the effect makes depth from one photo

A true anaglyph needs two photos taken side by side. With only one, the effect has to guess depth, and it uses brightness as the guess. Lighter parts are treated as closer and darker parts as further away.

The red channel is shifted one way and the cyan channels the other, by an amount that follows that guessed depth. Whatever sits at the focus depth stays lined up, everything else splits.

The depth map is blurred first, so the shift changes smoothly across a face or an object instead of jumping between neighbouring pixels.

Through glasses, bright subjects on a darker background really do seem to float forward. A dark object close to the camera will sit back instead, because the guess is based on brightness, not distance.

Photos that suit it

A bright subject against a dark background gives the strongest pop, such as a lit face at night or a product on black.

Portraits, statues and flowers work well. Strong edges show the red and cyan fringes clearly even without glasses.

Flat, evenly lit photos have little to work with. Raise the contrast first so there is a clear difference between near and far.

How to do it in Bitgrain

Open the editor from this page. The Anaglyph 3D style is already on the sample. Drop in your own photo.

Choose your Glasses: Red and cyan, Green and magenta, or ChromaDepth, which paints depth as a rainbow from red near to blue far.

Set Depth for how far the colours pull apart, and Focus for which depth stays lined up. For a bolder look without glasses, raise Depth.

Under Advanced, Depth softness smooths the guessed depth, and Contrast, Brightness and Invert adjust the tone. Export PNG when you are happy.

Common questions

Will it look 3D through real glasses?

Yes, with red and cyan glasses the red lens goes over the left eye. The depth is guessed from brightness, so it works best when the subject is brighter than the background.

Is this just an RGB split?

No. An RGB split moves the whole picture by the same amount. Here the shift changes with the guessed depth, so some parts stay lined up and others split, which is what gives the 3D feel.

Does it use AI?

No. Depth is worked out from the brightness of your photo by a fixed algorithm, in your browser. Nothing is generated and nothing is uploaded.

Related tools