A text-encoding mismatch can make subtitle characters render incorrectly even when their timing is right.
Why this happens
If the subtitle timing follows the dialogue but accented letters become odd symbols, the issue is different from a missing language track. Text subtitle files store characters using an encoding, and the player needs to interpret that encoding correctly. A wrong interpretation can produce readable words mixed with broken characters. It is worth noting which characters fail and whether the problem affects one subtitle file or every track in that language. Increasing the font size will not by itself correct the underlying character interpretation.
Use the player’s documented subtitle encoding control if it provides one, keeping the original value so the comparison is reversible. VideoLAN documents such controls in its subtitle settings; other applications may expose different options or handle encoding automatically. Do not apply this explanation to picture-based subtitles as if they were ordinary editable text. If another track in the same language displays correctly, preserve that difference. A clear report names the affected language, the source item and an example of the broken characters without including a private playlist link.
A useful way to work through it
Separate incorrect letters from a subtitle that is simply early or late.
Compare one documented encoding option and restore the original if it does not help.
Test another available track in the same language before changing global fonts.