In a previous article, I presented fake anaglyphic 3D photos: it was only a quick and dirty manipulation in Photoshop. The result was a bit « flat », and often left you cross-eyed.
■ True fake anaglyphic 3D explained
Going deeper in my research, I found out that a trick is used in movies: the films are shot, and 3D is added later! They are using what is called a depthmap.
The depthmap is a special grey-scale version of the original photo. The levels of grey represent the depth value of the objects. The most famous norm tells that lighter greys are near, darker greys are farther (neutral grey is the 0 position).
From this depthmap that gives a distance from the eyes of the observer, we can use simple mathematical calculus to create a shifted image from the original. Imagine that this shift represents the distance between the eyes...
Let’s say that the original image is the one for the left eye. By computing the shifted image for the right eye, we obtain the basis of anaglyphic 3D, it’s that simple!
Well, it’s not that simple. It takes a lot of time: each image has to be cut out in depth zones, which is an entirely manual operation. Then, the shifted image is computed (the Displace Photoshop plugin is perfect for that), then recombine the « left » and « right » images.
Cutting out the image in zones is so painful that I tried several ways:
- manually: requires a graphics tablet and a lot of patience
- cut-out plugins for Photoshop: I tried them all, but none was precise enough to obtain a correct 3D rendering
- the last solution I’m currently working on is using artificial intelligence: the advances in deep learning these last years are promising, and I’m still studying the subject
You have to know that I spend at least four hours per image for the moment!
■ True fake 3D examples
Now you can see a few examples of my work.
For an explanation of the formats used, please read the chapter after the gallery.
Croc-croc ©AbsurdePhotonOriginal photo of the squirrel
Croc-croc 3D (depthmap) ©AbsurdePhotonThe squirrel depthmap to compute the left & right images
Croc-croc 3D (red and blue anaglyph) ©AbsurdePhotonThe squirrel in colour red & blue anaglyphic 3D
Croc-croc 3D (cross-eyed anaglyph) ©AbsurdePhotonThe same squirrel in anaglyphic 3D without glasses
Croc-croc 3D (animated without glasses) ©AbsurdePhoton3D animated squirrel without glasses
In the meadow ©AbsurdePhotonThe original photo of this beautiful woman in a meadow
In the meadow 3D (depthmap) ©AbsurdePhotonThe depthmap shows the distances: grass, model, background, sky
In the meadow 3D (red and blue anaglyph, black and white version) ©AbsurdePhotonRed & blue anaglyphic black & white version
In the meadow 3D (cross-eyed anaglyph) ©AbsurdePhotonTry to cross your eyes to look at the photo
In the meadow 3D (animated without glasses) ©AbsurdePhotonThe angle is a bit different from the original in this animated version
■ 3D photo formats used on this site
In my photo galleries where you can find 3D versions of images, you can encounter the following formats. Here is how to visualize them:
- the original photo: it is here so you can compare with the other versions
- the depthmap: represents the distance in grey -scale of the objects in the original photo
- red and blue anaglyph: you need red and cyan anaglyph glasses (red for the left eye) to visualize the photo with depth
- cross-eyed anaglyph: for people who don’t own anaglyph glasses. To perceive depth, no need for them: you have to cross your eyes and control them, to try to mix the left and right parts in the « middle ». It requires a bit of training, and can cause strain to the eyes. The advantage is that the colours are not distorted.
For more explanations and training, you can follow this article on Kùla.
Software like StereoPhotoMaker let you visualize these files, and convert them to other formats
- parallel anaglyph: a variant of the above, no need for 3d glasses. Approach your nose near the image on your screen. Look at the image « through » the monitor, staring in the distance, right in front of you. The right and left images should mix in the middle. Be patient, it can take several seconds the first time. It requires a bit of training, and can cause strain to the eyes but less than the cross-eyed method. The advantage is that the colours are not distorted, the backdraw is the image is smaller.
For more explanations and training, you can follow this article on Kùla
- Animated: computing several positions allows to create a video giving the impression of depth without glasses. Videos are in aPNG format: if you don’t see anything moving, please update your internet browser! Only recent releases of Firefox, Chrome and Safari can read them
To create the videos, I’m using a heavily modified script given back to Ugo Capeto’s site for the oldest photos, and my own tools that you can find on GitHub for the latest ones.
Enjoy my photos in 3D!