Background

Background [Bg] is one of Fusion's fundamental Generator Nodes. Unlike MediaIn it has no required image input: the node creates its own image and therefore often starts a separate branch of a composition.

The simplest mode is Solid Color. Color defines RGB while Alpha defines the transparency of the generated fill, making the node useful for title plates, mattes, light passes, color cards and technical test images.

The captured Resolve 21 Inspector shows five visible Type choices: Solid Color, Horizontal, Vertical, Four Corner and Gradient. The first modes provide quick predictable fills while Gradient exposes the full custom-gradient control set.

Horizontal and Vertical create two-color transitions along the corresponding axis, while Four Corner uses four corner colors. These are efficient choices when a full custom-gradient editor is unnecessary.

With Type set to Gradient, a separate Gradient Type appears. The Resolve 21 Reference Manual documents six forms: Linear, Reflect, Square, Cross, Radial and Angle. These are spatial mappings of the same gradient data rather than six unrelated color sources.

Start and End define the gradient's spatial direction and scale in the Viewer. If the gradient points the wrong way, diagnose these handles before compensating by shuffling color stops or adding a downstream Transform.

Gradient Colors is a separate color-stop bar. Stop colors, alpha values and positions determine what is interpolated along the selected geometry; adding more stops does not change Gradient Type itself.

Interpolation Space determines the color space used to calculate colors between stops. This can materially affect saturated transitions and must not be confused with Source Color Space metadata on the Image tab.

Offset moves the gradient relative to Start/End. Repeat controls what happens when that shift extends beyond the base range: Once, Repeat and Ping-Pong create different behavior for animated bands and looping motion-graphics backgrounds.

Sub-Pixel controls precision when gradient edges become visible in repeat or animation. Higher precision can cost render time, so judge it on real motion at target resolution rather than treating the highest setting as automatically better.

The Image tab controls properties of the generated image rather than mask geometry. The captured Resolve 21 UI shows Process Mode, Width, Height, Pixel Aspect, Auto Resolution, Depth, Source Color Space and Source Gamma Space.

The current capture labels automatic size inheritance as Auto Resolution. The Resolve 21 Reference Manual describes the common generator behavior under Use Frame Format Settings: width, height and pixel aspect follow the composition. The article relies on the semantics; always trust the target-build label shown in Inspector.

When automatic frame-format inheritance is disabled, Width and Height define Background's own canvas. These are not Rectangle Width/Height and not Transform Size; they change the resolution of the image produced by the node.

Depth controls pixel precision. Float processing matters when downstream math must preserve HDR or super-white values above 1.0, negative values or repeated color operations without early clipping; it is a technical range decision, not merely a Viewer-quality switch.

Source Color Space on common Generator Image controls tags or passes color-space metadata for a linear workflow. The Resolve 21 manual explicitly distinguishes this from an actual conversion such as Gamut, CST or OCIO, so do not assume that selecting metadata transformed the RGB values.

Source Gamma Space / Curve Type handles gamma metadata and can expose Remove Curve or Log/Lin behavior in relevant modes. Treat this as a separate signal-state layer from Type and Gradient controls.

Pre-Divide / Post-Multiply is premultiplied-alpha handling. It is not a generic edge enhancer; evaluate it on genuinely translucent pixels and compare soft edges over contrasting backgrounds.

Background's only external input is the optional blue Effect Mask. The mask does not bring RGB into the node and does not replace canvas size; it limits the coverage of the image that Background already generated.

Settings controls such as Apply Mask Inverted and Multiply by Mask alter how the connected mask is applied. Evaluate these with the Alpha Viewer and a real composite rather than judging only one RGB frame.

Global In/Out defines the time range in which the generator produces an image. Outside that range a Background should not be treated as an infinite source, which matters in templates and when diagnosing a layer that suddenly disappears.

A reliable working order is simple: decide which RGBA pixels Background should generate, set canvas/depth, apply shape coverage with a mask, and only then decide how the finished layer is composited with Merge. This keeps generation, shape geometry and compositing as separate problems.