How to latin filter edge cases and quirks reference
- Step 1Audit Latin coverage first — Run the Character Coverage Map. The Latin Basic + Latin-1 row should be at 100% before filtering. If lower, the source font is missing characters and the filter result will inherit that.
- Step 2Check for hybrid scripts — Some specialty fonts ship Greek + Latin in one file (academic/scientific typography). Latin Filter drops the Greek; if you need both, use the Smart Subsetter with a custom union of latin + greek presets.
- Step 3Verify metrics integrity — Some fonts lock OS/2 ascent/descent values that depend on glyphs outside Latin. Filtering doesn't change the metrics, but if your design depended on the original glyph extremes, validate with the Metrics Analyser.
Frequently asked questions
Why does my filtered font look smaller per em?+
If the source font's largest glyphs were outside Latin (e.g., a CJK font with Latin support), filtering removes them but keeps the original UPM and bbox. The filtered font may render at a different em-size compared to a properly recompiled font.
What if the source has zero Latin coverage?+
The filter throws a clear error: 'Subset would be empty.' This catches the case where you accidentally fed a Devanagari-only or pure-emoji font through Latin Filter.
Does it strip the Latin numerals?+
No — 0–9 are in Basic Latin (U+0030–U+0039) and are preserved by definition.
Will OpenType features survive?+
Mostly yes for Latin-only features (kern, liga, smcp). Features that depend on script-specific glyphs (Cyrillic-specific locl) are reduced or dropped along with the glyphs they reference.
Privacy first
Every JAD Font tool runs entirely in your browser using opentype.js and the wawoff2 WASM Brotli encoder. Your fonts never leave your device — verified by zero outbound network requests during processing.