PDF fonts not displaying correctly — why the same file looks different on every machine

Updated 13 August 2026

A PDF font changes because the font was never embedded in the file. When a viewer cannot find it, it substitutes one installed locally — spacing and stem weights drift, so the same document renders differently on every machine. Embedding, flattening or PDF/A locks the appearance.

Why PDF fonts display wrong: embedding, subsets and viewer substitution — why a file renders differently on two machines, and how to lock its look.

1. Open the file's document properties in any viewer and read the Fonts list.
2. Look for the word Embedded beside each font — and for six-letter prefixes like BAAAAA+, which mark an embedded subset.
3. Any font without Embedded is being substituted from whatever that machine has installed.
4. Re-export with fonts embedded, or lock the current appearance by flattening the file.

What a viewer does when the font is missing

A PDF names the fonts its text is drawn with, and may or may not carry the font programs themselves. When a font is not embedded, the viewer substitutes from what is installed locally — it has to draw something. The substitute carries different widths and stem weights, so spacing drifts, lines re-wrap, and a table that fitted its column stops fitting. Nothing is corrupt: the document outsourced its appearance to the reader's computer.

Embedded, subset, absent — how to tell in ten seconds

Read the Fonts tab in document properties. Embedded: the font program travels inside the file and every machine renders it identically. Embedded subset: only the characters the document uses are carried, marked by a six-letter prefix such as BAAAAA+SFNS-Regular — the normal case. Not embedded: substitution is happening, and the page you see is not the page the author saw.

Why one typeface becomes thirty font objects

Subsetting also explains a Fonts list that looks alarming. Some producers split one typeface across many subset objects — Chrome's print-to-PDF can emit thirty-plus font objects for three typefaces. A long list is not evidence of damage; the mechanics are in how PDF editors keep the same font — or quietly fake it. The consequence that does bite is narrower: a subset may not contain a character you later type.

Fixes that travel with the file

The clean fix is upstream — re-export from the source with font embedding enabled, a checkbox in most word processors. Converting a Word, PowerPoint or HTML document straight to PDF here embeds Cyrillic, Greek and extended-Latin text automatically — see converting non-Latin text to PDF for what is and is not covered. When the source is gone, two routes lock what you have. Flatten PDF bakes the current content so appearance stops depending on installed fonts; the cost is editability, so keep the unflattened copy. PDF/A is the archival direction: everything needed to render the page travels inside the file, which is why archives and courts ask for it. Rendering pages to JPG freezes appearance too, at the cost of every selectable word.

Editing a file whose fonts are already unreliable

Editing is where substitution turns from cosmetic to obvious, because a new word in a fallback face sits beside old words in the real one. UnboundPDF's editor reuses the fonts embedded in the document itself, so edited lines match their neighbours — and when a font cannot be reused safely it says so rather than substituting silently. The full fallback order is in the same-font guide; if the click will not open an edit at all, the cause is one of seven.

When the problem is not the font at all

Two lookalikes worth separating. A scanned page has no fonts — only pixels shaped like letters — so what reads as a substituted typeface is the scanner's rendering of the original print; OCR is the route to real text there. And if the letterforms are correct but soft, the problem is resolution or compression: that is why is my PDF blurry.

Scope

A font that is neither in the file nor on your system cannot be embedded after the fact, by any tool. Flattening fixes appearance at the cost of editability. When an embedded subset lacks the glyph you typed, the editor embeds a replacement and names it instead of printing empty boxes. All of this runs in your browser; verify the no-upload behaviour in the PDF Privacy Lab or on Compare.

Frequently asked questions

Why does my PDF look different on another computer?

Because at least one font is not embedded in the file. The viewer falls back to a font installed on that machine, and the substitute has different widths and stem weights — so lines re-wrap, spacing shifts, and the page reads as subtly wrong.

What does the BAAAAA+ prefix on a PDF font mean?

It marks a font subset. PDFs usually embed only the characters the document actually uses, and the six-letter prefix identifies that subset. It is normal, not damage — but it is why a character the document never used can be missing when you edit.

How do I stop a PDF font from changing?

Embed the fonts at export time in whatever produced the document. If the source is gone, flatten the file so the appearance is baked into the page content, or convert to PDF/A, whose whole point is that everything needed to render the page travels inside the file.

Can I just install the missing font?

That fixes it on your machine only. The file still renders differently for everyone else, because the fix lives on your system rather than inside the document. Only embedding travels with the file.