PDF Fonts: Embedded, Subset and Missing Explained

A PDF is meant to look the same everywhere, and fonts are the main way that promise can be broken. When a document appears with different spacing, strange characters or boxes on another computer, or a printer says "fonts are not embedded", the reason is in the font list of the file. This guide explains what that list means.

Embedded

An embedded font has its glyph shapes stored inside the PDF, in a font file of one of a few types: Type 1, TrueType, or CFF/OpenType. A viewer draws the page with that stored data, so the document does not depend on the fonts installed on the computer. This is the normal and safe state.

Subset

Most embedding tools store only the glyphs that the document actually uses, which keeps files small. A subset font has a name with six capital letters and a plus sign in front, such as ABCDEF+Calibri. A subset is fine for viewing and printing; its limit is that you cannot add new text in that font with just the embedded part. A full embedding contains the whole font and is rarer.

Not embedded

If the file only names a font, the viewer looks for a font with that name, and if it does not find one it substitutes another. Spacing, line breaks and special characters can change. This is what the warning in the PDF Font Checker is about.

The 14 standard fonts

Early PDF defined 14 fonts that every viewer was expected to have: Courier, Helvetica and Times each in regular, bold, italic and bold-italic variants, plus Symbol and ZapfDingbats. A PDF may reference them without embedding. That is allowed in ordinary PDFs, and the checker lists them as "standard 14, not embedded" rather than as an error. Archive and print standards, PDF/A for example, require every font to be embedded; the PDF/A Checker deals with that side.

Type 3 and composite fonts

A Type 3 font draws its letters with PDF drawing commands inside the file. It always travels with the document, but it cannot carry hinting and some programs produce poor text from it. A composite (Type 0) font is the usual form for TrueType and OpenType fonts with large character sets; its real font program is attached to a descendant, and the checker follows it.

ToUnicode and copying text

Whether you can copy or search text depends on a ToUnicode map, which says what character each glyph stands for. A font can be embedded perfectly and still give unreadable copied text if the map is missing or wrong. For legacy-encoded Bangla, see Bijoy vs Unicode.

Reading a font report

• "Declared on pages" says which pages list the font; "selected" says which pages actually switch to it.
• Look first for fonts marked not embedded and not in the standard 14; those are the ones that can change the layout.
• A report cannot tell whether an embedded subset includes every glyph needed, or whether a font licence allowed embedding.

Fixing missing fonts

Go back to the program that made the PDF and turn on "embed all fonts" when exporting or printing to PDF, or install the missing font and export again. If you only have the PDF, a PDF editor can sometimes re-embed from an installed font, but results vary. For a print job, tell the printer which font problems you found. Related print preparation is covered by the PDF Bleed Adder, and the general state of a file is shown by PDF Information.