sText

sText [sTxt] is an advanced character generator designed for the Fusion Shape environment. Unlike Text+, its result is not immediately a normal RGBA image; it remains Shape data until a downstream sRender.

The topology matters: sText has no external Shape or Image input. The node generates its own glyph geometry from Styled Text, Font, Style and the rest of the text controls.

Shape Output can feed sGrid, sDuplicate, sTransform, sChangeStyle, sMerge, sBoolean and other Shape tools, or go directly to the orange Input1 on sRender.

sRender is the vector/raster boundary. Before it, the typography is still editable Shape geometry and style; after it, the result is a regular 2D image with alpha for Merge, Glow, Blur and other image tools.

Blackmagic describes sText as a Shape version of the 2D Text+ tool whose controls are largely the same. One explicit exception is important: sText does not support gradient colors.

The Text tab controls content and typography: Styled Text, Font, Style, Color, Size, Tracking, Line Spacing, vertical and horizontal anchors/justification, Direction, Line Direction, Emphasis and Write On.

The captured Text panel shows two anchor layers per axis: an icon row selects the anchor mode while a separate numeric slider refines the anchor position. Treating both as one generic Anchor hides useful behavior.

V/H Justify aligns content while Anchor establishes the reference point. Dynamic titles often drift when these concepts are mixed and the text string changes length.

Write On uses Start and End to define the visible text range. It is not a downstream raster mask; it changes which part of the glyph sequence is produced by the generator.

Tab Spacing provides selectable tab stops plus Position and Alignment. It is suitable for structured text, but it should not be treated as a full table-layout system.

Advanced Controls include Reading Direction, Force Monospaced, font kerning, ligatures, split ligatures, Stylistic Set and Font Features. These are text-shaping controls, not Shading layers.

Manual Kerning and Manual Positioning are local corrections. Global Tracking should solve global rhythm while manual adjustments should be reserved for specific pairs or characters.

Layout controls placement of the text object separately from font metrics. The captured panel shows Type = Point, Center X/Y/Z, Pivot X/Y/Z, Size, Perspective, Rotation Order and X/Y/Z rotation.

Text Size and Layout Size are not interchangeable. Text Size changes the base typographic scale while Layout Size scales the already placed text object.

Center moves the entire layout. Pivot establishes the center of layout-level rotation and scale, so a poor pivot often looks like an unintended orbit during animation.

Perspective and X/Y rotations can create a 3D-like presentation, but sText remains a Shape-system vector tool. It does not become Scene3D geometry and is not a replacement for Text3D.

The Transform tab works inside the text object. The captured panel shows Transform = Characters, Spacing, Offset X/Y/Z, Rotation Order, X/Y/Z, Adapt Width to Angle, and collapsible Pivot, Shear and Size groups.

With Transform set to Characters, offsets, rotation and size operate at character scope rather than moving the entire title. Use Layout or a downstream sTransform when the whole object should move.

Adapt Width to Angle is best judged together with angular transforms. It is not a universal quality switch.

Shading Elements provide up to eight independent appearance layers. The captured panel shows Select Element 1-8, Name, Enabled, Sort By and the Properties section for the selected element.

The captured first element is named White Solid Fill. The name is a user-facing label for the current shading layer, not a separate node type or hard-coded preset.

Sort By offers Priority, Distance and Z Position. Priority is the straightforward 2D stacking mode; the other modes matter when spatial ordering is part of the design.

Opacity applies only to the selected Shading Element. It is not a master opacity for the entire sText node and it is not the Blend control of a later Merge.

Blending defines how the selected shading layer combines with the other text shading layers. Composite is the captured baseline.

Appearance selects the visual role of the active shading element, such as fill/outline-like representations. It styles existing glyph geometry rather than changing the text string or font.

Override Color Fonts controls whether embedded color-font information is replaced by the active shading settings. It may appear to do nothing with ordinary monochrome fonts.

The color swatch and Red/Green/Blue/Alpha belong to the selected shading layer. Alpha defines authored transparency and Opacity attenuates that layer again, so lowering both compounds transparency.

Position/Priority and Offset X/Y/Z move only the current shading layer. This is the correct level for shadow-like or extrusion-like offsets that should not move the whole text object.

Rotation, Pivot, Shear and Size on the Shading tab are also per-element transforms. Controls with similar names on Layout and Transform have different scopes.

sText is strongest when typography must remain vector data for Shape operations: grids, duplication, boolean work, geometric combinations and one final sRender.

Text+ is often more convenient for ordinary title-over-video workflows because it outputs an image directly and supports gradient-capable text shading. Choose the tool based on pipeline requirements.

For quality control, inspect vector structure before sRender and raster edges after sRender. A fitted Viewer cannot replace a 100% check at the target resolution.

Animated Write On, character transforms and shading offsets must be reviewed in playback, not only on a single still frame.

When the text content changes length, re-check anchors, justification, layout center and shading offsets. A design tuned to one word may not remain stable for a longer string.

The captured Inspector panels are UI evidence for one saved state, not universal factory defaults. Additional advanced controls should be treated as documented controls rather than guessed menu states.