Metaball

Depth points scattered into a volumetric density field and raymarched as an isosurface — blobs that genuinely fuse where they meet, with capsule connectors, cast shadows and ambient occlusion.

DEPTH ✦ family

← All the FX

Family
DEPTH ✦
Availability
Part of the full library — see pricing
Controls with options
4

A metaball is a field, not a sphere

Give every point in space a number that falls off with distance from a centre, then draw the surface where that number crosses a threshold. One centre gives you a sphere. Nothing surprising yet.

Now add a second centre and SUM the two fields. Far apart, each one crosses the threshold on its own and you see two separate spheres. Bring them closer and the gap between them starts receiving a contribution from both — and at some separation the sum in the middle crosses the threshold too, so the surface closes over the gap and the two become one shape with a waist.

Nothing was joined. No geometry was welded. The merge is simply what adding scalar fields does, and it is why metaballs melt into each other in a way that no arrangement of actual spheres can imitate. It is also why the melt behaviour and the size are two different controls: one is the falloff, the other is the threshold.

Why the connectors have to exist

Summed spheres have a hard limit. Past a certain separation the contribution in the gap never reaches the threshold no matter how the falloff is shaped, so the blobs simply come apart — you cannot get a ball-and-stick structure out of spheres alone. That is a property of the maths, not a shortcoming of the implementation.

Capsule connectors add an explicit link between neighbours, stamped into the same field, so structure survives across distances where the fields alone would have separated. Which points get linked is re-rollable independently of where the points sit, which means the structure can be varied without moving anything — a different lattice out of the same scene.

Raymarching, and the second march

There is no mesh here to rasterise. The surface is found by marching a ray out from the camera through the density field until it crosses the threshold, which is how implicit surfaces are drawn and why the shape can change every frame without anything being rebuilt.

Cast shadows and ambient occlusion come from marching that same field again — toward the light for shadow, around each surface point for occlusion. Doing it in the field rather than compositing it afterwards is why the shadows land correctly inside the crevices where blobs merge, which is precisely where a screen-space fake gives itself away.

The camera orbits with real parallax for the same reason: the geometry exists, so moving the viewpoint is just moving the rays.

Scale by depth, and the fields

Because every blob came from a pixel with a depth, each one can be scaled by its own distance. Turn that up and the far end of the scene resolves into a fine lattice while the foreground merges into solid mass — a single control that produces two quite different materials in one frame, ordered by the actual geometry of the photograph.

Six displacement fields then push the underlying points, with Curl Flow again the one that folds without tearing. Four materials and three environments finish it, including an image-based environment lit by the picture the blobs came from.

When you’d reach for it

When you want the photograph rebuilt as substance rather than as points. Point Cloud keeps the picture legible as a cloud of samples; Metaball melts those samples into continuous volume, which is a much bigger departure from the source and a much stronger read as an object.

Every option, from the app

Generated from the registry the app publishes, so this list is whatever the current build actually ships — not a copy of it.

Range
Field
TurbulenceCurl FlowRadialWindGravityVortex
Material
ClayPearlBrushed MetalLiquid Metal
Environment
StudioImageHorizon

Open the app to try it; the free effects work with no signup. Open V3KTR · all the FX · how it works