Why scanned PDFs can’t be read aloud, and what to do about it
If every text-to-speech tool you try says your PDF has no text, it’s probably a scan. The file holds pictures of the pages and no actual words, so there’s nothing for a voice to read.
What a scanned PDF is
A text PDF stores characters: this is an “a”, in this font, at this spot on the page. A scanned PDF stores a photo of each page, wrapped up so it opens like a document. To you they look the same. Text to speech reads characters, so a scan gives it nothing to work with.
How to tell which you have
- Try to select a word. Press and hold on your iPhone, or drag across a line on a computer. If nothing highlights, it’s a scan.
- Search for a word you can see. If search finds nothing, there’s no text layer.
- Look at the page. A tilted page, a visible book edge, a grey background or a thumb in the margin give it away.
Some scans already have a hidden text layer added, often called a “searchable PDF”. Those will read aloud, recognition errors and all.
Look for a text edition first
If it’s an old book, a clean digital text may already exist. Project Gutenberg has tens of thousands of proofread public-domain books, and they’ll always read better than text recognised from a scan. Aloud searches Gutenberg and Open Library directly. More on the free catalogue.
How to run OCR
OCR (optical character recognition) turns the page images back into text.
On a Windows PC or Mac, for a whole document
- Adobe Acrobat has a recognise-text tool that adds a text layer to a scanned PDF. It’s in the paid version.
- OCRmyPDF is a free, open-source command-line tool that adds a text layer and keeps the original page images. It runs on your own computer, which matters for private documents. It takes a few minutes to install and it’s the best free option for long books.
- Google Drive converts a PDF to an editable Google Doc when you open it with Google Docs, recognising the text on the way. It’s free and easy, but the file goes to Google.
On an iPhone, for a few pages
Live Text can pick up text in a photo. Open a photo of the page, press and hold on the text, choose Select All, then Copy, and paste it into Notes or Pages. It’s accurate and it runs on the phone, but it’s far too slow for a whole book. For a letter or a handful of pages it’s ideal.
Check the result before you listen
OCR text is only as good as the scan. Read a page or two of the new PDF before you commit a long trip to it, and look for:
- Misread letters, like “rn” read as “m”, or “l” and “1” swapped. The voice reads these exactly as written.
- Merged columns, with lines that jump across the page.
- Hyphens left over from line breaks.
- Old typefaces, especially the long “s” in pre-1800 printing, which often comes out as “f”.
If the first pages are poor, rescanning flatter and brighter usually helps more than any OCR setting.
Then listen
Import the recognised PDF like any other. If OCR didn’t take, Aloud reports the file as a probable scan instead of opening a silent, empty book, so you’ll know straight away.
Aloud, once the PDF has text
After OCR, Aloud reads the PDF in chapters with your place saved, narrated on your iPhone. Your first import is free.
Common questions
What does OCR stand for?
Optical character recognition. It's software that looks at a picture of text, works out the letters and stores them as real text you can select.
Can Aloud read scanned PDFs?
No. If a PDF has no selectable text, Aloud tells you it's probably a scan instead of importing an empty book. Run OCR on it first and import the result.
Is the OCR text perfect?
Rarely. Clean modern print comes out well. Old typefaces, faint photocopies, curved pages and handwriting cause errors, and the voice reads those errors exactly as written.
Is it safe to use an online OCR website?
It means uploading the document to someone else's server. That hardly matters for a public-domain book. For contracts, medical letters or anything private, use an offline tool on your own computer.
Forge
Aloud
Moat