Use one page to find unwanted narration
Generate a sample from an ordinary content page and hear both its opening and ending. If a page number or chapter label appears in a strange place, inspect Text to locate it. The voice follows its input, including words you instinctively skip when reading the original layout visually.
A running header differs from the real heading of a new section. Before removing a line, compare it with the PDF and decide whether it merely repeats for orientation or introduces a new idea. Useful cleanup preserves structure instead of flattening the whole book into an uninterrupted stream of sentences.
Reconnect sentences without merging separate ideas
A sentence can break across printed lines or pages. If extraction places the book title inside that sentence, remove the peripheral line and check that the original meaning reconnects. Preserve the author’s punctuation rather than creating a long, difficult spoken sentence by deleting every break and mark indiscriminately.
Avoid applying an overly broad removal rule because several pages look similar. A number may be a page marker, but it could also be a year, a result, or an instruction step. Review the positions you changed instead of assuming that every short line is disposable material outside the main text.
Decide what the listener needs from footnotes
Some footnotes are essential to meaning; others are reference material you do not need to hear during a personal reading session. If keeping one, place it where the listening text remains understandable. If omitting it from your audio copy, retain the source PDF so you can consult the note later.
For scans, clean the recognized words after OCR and before generating audio. Missing Vietnamese marks or joined words may become obvious only in a trial recording. Choose a sample containing the main passage and a note so the test represents the document rather than its simplest, least demanding page.
Keep a finished version after reviewing it
Once the source is tidy, select a Vietnamese voice or Clear English to match the language and regenerate the required passage. Leave the app in the foreground while it works. A completed WAV/M4A export reflects the text at that generation time; it does not automatically change after every later text edit.
Name the finished export so you can distinguish it from an earlier trial containing page numbers or repeated labels. Hear the beginning and ending before relying on a longer session. A little attention to the source preserves continuity more effectively than creating several new recordings without inspecting the words behind them.
TRY IT WITH YOUR OWN WORDS
Your next chapter starts here.
FolioTone works on iPhone and iPad with iOS 18 or later. Voice, OCR and export limits depend on your Free or Pro plan; check the app for current limits.
Get FolioTone on the App StoreCompare Free & Pro