Grayscale depth map encoding vertical distance information — enables parallax layering and precise Z-space compositing. Essential for stereoscopic and multi-plane effects.
In compositing, we constantly work with depth data — but the classic depth map only captures the Z-axis, meaning the distance from the camera. However, as soon as you deal with vertical parallax effects, focus shifts, or spatial layering, you quickly realize that you also need to know where something is positioned vertically within the frame. The Y-depth image steps in precisely here: a grayscale map that doesn't store depth distance but rather the position on the vertical axis — from the bottom edge of the image to the top.
Practically, this means you get additional channel information alongside your standard depth and position passes. While the classic depth map tells you "this object is 5 meters away," the Y-depth image tells you "this object is at 65% of the frame height." This is essential, for example, when simulating a parallax shift during camera movement — different vertical positions will then shift visually at different speeds, greatly enhancing the sense of depth. You need this precision, especially in 3D compositing or when integrating CG elements into live-action, to achieve natural-looking movements.
Application and Workflow
In editing, the Y-depth image is typically exported from the 3D render as a separate channel — similar to position passes or object IDs. You pack it into your EXR sequence and then have direct access to this vertical data in Nuke or After Effects. You often combine it with standard depth and Z-position for complex reprojection operations or for controlling defocus effects that should vary depending on vertical position. A classic example: you have a scene with multiple figures at different heights. With a Y-depth image, you can define a focus area that not only considers depth but also accounts for the fact that the upper figure is positioned higher and might become out of focus faster than the lower one.
Some rendering engines don't generate Y-depth images automatically — in such cases, you extract them yourself from your position passes. This takes a minute of work and immediately pays off later in compositing. The purpose is to decouple spatial operations from pure Z-depth space and bring true 3D coordinate information into compositing. This saves you hours of masking work and gives you a significantly better grip on complex layering.