Exporting Variable Fonts
Compile the designspace into a variable font and inspect its interpolation, layout, and hinting at representative locations.
Save the native project, then choose File > Export Font As… and select a Variable TT or Variable PS/CFF2 profile. Use Customize… to review its settings before export. See The Export Process for destination and batch-export controls.
Variable Fonts
A variable font packages a designspace for use at runtime. It does not store the complete native FontLab source: it stores the default design and the variation data needed by the target representation.
Variable TT uses TrueType variation data, including gvar, and is a common delivery route. Variable PS stores variable cubic outlines in CFF2. Choose a flavor supported by the intended applications, then test FontLab’s exported file there; support for ordinary static OpenType does not establish support for both variable flavors.
Before export, check compatible glyph structures, usable master locations, axis identities, the default/main master, and named instances. Then inspect the output's key relationships:
fvardescribes axes and named locations.avarcarries applicable axis-coordinate mapping.STATprovides style-attribute information.- Variation metrics and supported MVAR/GPOS data preserve relevant metric and positioning changes.
User coordinates are mapped and normalized for the compiled model. A design-range extension in the source is not, on its own, evidence that the exported font has an anchored and usable extended range.
Visual kerning and varying anchor data need the proper variable-output build path, which rebuilds positioning from master data; recall Kerning, GPOS, and the kern Table. A fixed handwritten snapshot of one master's positions does not become varying GPOS automatically. Check substitutions, marks, and kerning at intermediate locations, not just at named instances.
Component retention, coordinate precision, and hinting depend on the output flavor and profile. TrueType point coordinates commonly require integer representation, while CFF-based paths have different precision and hinting rules. Retain fractional source work and judge the actual conversion instead of preemptively rounding every master without a reason.
Keep a static reference export when it helps compare a named variable instance with the intended design. Check family/style names, including typographic style name ID 17 from Naming Fonts and Families, in the consumer as well as the outline result.
Sigma and bracket layers
For the editing workflows, see The Sigma Tool and Bracket Layers.
Variable TT and Variable PS can export Sigma trajectories. The exporter checks sampled output against the source before publishing the font. Turn off Variations > Export Sigma in the export profile to omit Sigma corrections from the export copy and use ordinary linear interpolation. Native saves retain the authored trajectories regardless of this export option.
Bracket layers export as alternative glyph designs selected by conditional substitutions. The two layers at a break do not become two font-wide masters at the same location. Check the lower side, the boundary itself, and the first location above it, then test text with kerning, marks, ligatures, and optional features. The compiled boundary uses the variable-font coordinate precision, so also inspect the switch in the intended rendering application.
A font can contain both Sigma glyphs and bracket-layer glyphs. Export currently rejects a glyph that combines bracket switching with Sigma curves, including an affected composite. Do not assume that support for both features in a font means every combination within one glyph is supported. Resolve that combination in an export copy or generate the intended static instance.
On import, FontLab can reconstruct recognized early design switches as ordinary bracket layers and verified Sigma trajectories as editable paths. Sigma recognition supports CFF2 outlines and both quadratic and cubic TrueType outlines. Existing layout references and component references are retained; a separate source glyph may remain when another feature or component still needs it. Switches limited to particular languages, contextual rules, and other unsupported conditional layout retain their ordinary representation instead of becoming global bracket layers.
Reopening a binary is not a substitute for keeping the native project. Reconstruction depends on the compiled structure being recognized and verified. Keep the VFC or VFJ source, and compare outlines and layout after any export–import–export workflow.
Review Hinting and Component Diagnostics
If a glyph’s masters are incompatible for visual-hint preparation or conversion fails, variable export can continue with warnings and omit unsafe visual hints for the affected glyph. Read the diagnostics and inspect the exported result instead of treating file creation as evidence that every intended hint survived. Variable components referencing bracket glyphs must be decomposed before variable export.