Fields

A field has a value at every point in space, worked out on the spot rather than stored anywhere.

What a field holds

Most data in DNA is stored. A collection holds a fixed list of rows, a raster a fixed grid of pixels. A field holds no values at all. It holds a rule for producing a value anywhere it is asked.

Ask that rule at a single point, at a million points, or across a whole surface. Only the points actually used are evaluated.

Because a field is a rule, it costs almost nothing to carry and has no resolution to run out of. Zoom into a noise field indefinitely and it never goes blocky, because the value is recomputed for wherever you are looking.

Flavours

A field always returns something at each point. What it returns depends on its flavour.

• Scalar: a single number at every point. Density, height, displacement strength, a mask, how much of anything.

• Vector: a direction and length, an arrow, at every point. Use it to push, flow, or steer. Wind, currents, forces.

• Colour: an RGBA colour at every point. This is how you texture and paint without unwrapping UVs. Noise, gradients, checkerboards, voronoi cells and sampled images all arrive as colour fields.

• Distance: at every point, how far you are from a shape's surface and whether you are inside or out. The maths behind smooth blends, rounding, hollowing, and the Analytic shapes pipeline.

• Multi-channel (record): many channels bundled into one field. A material field carries albedo, roughness, metallic and normal, all sampled together at the same point.

• Morph: a continuous blend between two different things. A morph field melts one shape into another, a curve into a volume or a mesh into a blob, evaluated continuously so the in-between states are real surfaces.

Sources and consumers

Fields come from generator nodes: noise, gradients (Scene values), patterns, a Sphere's distance, a sampled image. They also come from the The Expression language node, where you write your own per-point formula.

They also come from other fields. Derive takes a field in and hands back an operator over it: its gradient, its edges, its normal, its curvature or laplacian, the divergence, curl or magnitude of a vector one, or the hue, saturation or luminance of a colour one. The result is another field, so it composes with everything on this page. Set its Disposition to Capture and the result is added to the input as a named channel instead of replacing it, which is one of the ways a plain field becomes a multi-channel one.

Once you have a field, feed it to something that samples it.

• A colour field on a shape's fill paints across the surface, point by point.

• A scalar field wired into a displacement gives every vertex its own value to move by.

• A vector field fed into a simulation gives each particle the arrow under it to drift along.

• Distance fields combined with booleans carve, merge and smooth solid forms.

The rule travels with the field, so the same noise can drive colour on one node and displacement on another in lock-step. Both are asking the same question and expecting different answers.

Field vs collection vs raster

• A collection is stored rows: a definite list of points, shapes or geometry you can count and edit. See Collections.

• A raster is stored pixels: a fixed grid, resolution and all. See Rasters (images).

• A field is a stored formula. No rows, no pixels, only a value waiting to be asked for.

Use a field for anything continuous and resolution-free. Use a raster or a collection when you need concrete, countable, exportable data.

See also

• Collections

• Rasters (images)

• Distance fields & volumes

• The Expression language