How to Listen to Scanned PDF Books

A scanned PDF is a container. What is inside the container determines whether TTS works. There are two possibilities:

• PDF with a text layer. Behind the page image, invisible to your eyes, is a layer of selectable text that OCR software generated when the scan was created. This text is what TTS engines read. You can test for it by trying to select a word with your cursor. If it highlights and you can copy it, the text layer exists.
• Image-only PDF. The PDF contains page images and nothing else. Each page is a photograph of printed text, not text itself. Clicking on a word does nothing, or selects a rectangular region of the image. No text means nothing for TTS to read.

This distinction is the whole article. Everything downstream depends on which type of PDF you have.

The test takes five seconds and works on any device:

1. Open the PDF in Preview (Mac), the Files app (iPhone), or any PDF viewer.
2. Try to select a word by clicking and dragging over it with your cursor or finger.
3. If the word highlights, the PDF has a text layer. It will work in LoudReader.
4. If nothing highlights and you get a crosshair or a rectangular selection, the PDF is image-only. It will not produce audio in LoudReader.

This is the single most useful thing you can do before importing a scanned PDF into any TTS app, not just LoudReader. Every TTS engine needs actual text to work with.

You need to run OCR on it. The goal is to add a text layer to the existing page images so that a TTS engine has something to read. Here are the options, from simplest to most thorough:

• Apple Live Text (free, built in). Open the image-only PDF in Preview on Mac. Click and drag to select the area with text. If the text is clear enough, Live Text recognizes it and lets you copy it. Paste the copied text into a text editor, save as a new PDF, and import into LoudReader. This works for short documents but is tedious for a 300-page book.
• Adobe Acrobat Pro (paid). Open the PDF, go to Scan and OCR, then Recognize Text. Acrobat processes the entire document and adds a text layer behind each page image. The result works in any TTS app, including LoudReader.
• Tesseract OCR (free, open source). Tesseract is a command-line OCR engine. It requires some setup but produces good results for clear, well-lit scans. The command is something like tesseract scan.pdf output pdf. The output is a new PDF with a text layer.

After OCR, test the new PDF by selecting text. If text is selectable, import into LoudReader and press play. For more on the import workflow, see listening to PDFs on iPhone.

The voice quality is the same natural neural voice LoudReader always uses. But OCR accuracy varies a lot, and the errors become audible:

• Clean scans produce clean audio. A well-lit, high-resolution scan of a modern printed book, run through good OCR software, might have 99% accuracy. One error per hundred words is barely noticeable when listening.
• Poor scans produce noisy audio. A faded, skewed, or low-resolution scan of an old paperback with tight margins produces OCR errors that the TTS voice reads as garbled text. "cl" becomes "d", "rn" becomes "m", and sentences containing these errors sound like a book full of typos. This is not a LoudReader limitation. It is an OCR limitation, and no TTS engine can fix bad input text.

The honest advice: if you are scanning books for TTS listening, invest in a good scan first. Flatbed scanner, high resolution, good lighting, and OCR software with a preview step so you can correct common errors. A clean scan takes more time upfront but makes the difference between a listenable audiobook and a frustrating one.

Most professionally produced PDFs from the last two decades include a text layer because it enables search, copy, and accessibility features. These sources reliably produce text-layer PDFs:

• The Internet Archive. Millions of scanned public domain books with OCR text layers already applied. Download the PDF and test text selection. Most work with LoudReader.
• Google Books. Public domain books on Google Books are available as PDFs with text layers. The OCR quality is generally good.
• Academic databases. JSTOR, PubMed, and university repositories distribute PDFs with text layers. These are designed for text extraction and work reliably with TTS.
• Library scans. Many public and university libraries scan their collections with OCR enabled. The resulting PDFs are searchable and listenable.

For books that are already in the public domain, turning any book into an audiobook covers the full path from file to audio. The same principles apply to scanned PDFs with text layers.