Saw this cross BlenderNation this week: True Depth, the Blender add-on that turns a flat image into 3D geometry, shipped a V3 update that’s a genuinely bigger jump than the version number suggests. Worth stopping on, because the underlying trick — camera projection from a single reference — is one every matte painter and environment artist already knows by hand. This just automated the part that used to eat an afternoon.
What it does
True Depth runs a monocular depth-estimation model (it’s built on the second generation of DepthAnything) over any image you feed it, generates a depth map, and displaces it into real geometry — no photogrammetry rig, no LiDAR scan, no reference set of overlapping shots. One image in, mesh out. All of it runs locally, on your own GPU, not through a cloud API.
V3 adds three modes on top of the original flat-relief displacement:
- Flat Relief — the original mode: a 2.5D displacement straight off the depth map, good for quick paintover-style bumps.
- Perspective — builds an actual camera-projected 3D scene from the image, stretched back into depth. This is the one that matters for environment work.
- 360° HDRI — takes an HDRI environment and projects it onto a dome, giving you a rough lighting/staging setup in one click.
It also now claims print-ready output for anyone wanting to take a depth relief into physical form, though that’s the least interesting part of this for a games or film pipeline.
Where this actually lands in a pipeline
Camera projection onto rough geometry is not a new idea — it’s how matte paintings have been integrated into live-action and CG shots for decades, and it’s how a lot of previz environments get built when there’s no time to model anything real. What’s new is that the depth pass, which used to be either hand-painted or reconstructed from multiple photos, now comes out of a single image in seconds, locally, without sending anything to a third-party server.
That matters for the boring but real reason: a lot of studio pipelines still won’t let you run concept art or location photography through a cloud AI service, full stop. A tool that does the estimation on your own machine sidesteps that conversation entirely, which is probably a bigger deal for adoption than the depth quality itself.
The honest use case is previz and pitching, not production geometry. If an art director hands you a location photo or a concept painting and asks “can we get a camera move through this by Friday,” True Depth’s Perspective mode gets you a plausible, lit, movable-camera version of that image in an afternoon instead of a week of manual projection setup and geometry sculpting. That’s genuinely useful — it’s the difference between showing someone a flat image and showing them a shot.
What it isn’t is a replacement for actual environment modelling. Monocular depth estimation has a hard ceiling: it infers depth from a single vantage point, so anything the camera didn’t see — the back of that building, what’s actually behind the foreground tree — doesn’t exist. Push the camera more than a little off the original angle and you’ll see the stretching and the holes where occluded geometry should be. That is the physics of working from one image instead of a scanned or modelled scene. Anyone who’s tried to fake depth with a single parallax-mapped card in a game engine will recognise the limit immediately.
The actual value
The right way to think about this is as a faster front end to a process environment artists already run informally — project a photo onto a rough proxy to sell a camera move before committing to a build. True Depth just compresses the “generate a usable depth pass” step from hours to seconds and does it without a cloud round-trip. Worth having in the toolbox for pitching and blocking, not worth mistaking for a shortcut around modelling the thing for real.