How to scan a textbook
Plan around a textbook's three problems before you start: a thick binding with a deep gutter, diagrams that need to stay legible, and page numbers you'll cite. Then scan, clean and OCR the whole book at once.
Key takeaways
- Decide first whether the book can be cut. A cut textbook scans flat and fast, but loses its resale value.
- Thick bindings curve the lines near the spine. Scan from above or press the page flat, and crop past the gutter shadow later.
- Keep diagram and photo pages in greyscale. Turning them pure black and white wipes out shading and labels.
- Never renumber or reflow a textbook. Your lecturer's page references and the index only work if the pages stay as printed.
- A text layer turns a scan into something you can revise from: search a term, jump to every page it's on.
Why a textbook isn’t just a long book
A novel is one column of words. A textbook is a crowded page: headings, sidebars, diagrams with labels, tables, equations, footnotes and margin notes, often printed on glossy paper in a binding several centimetres thick. Each of those makes scanning harder in its own way.
| Textbook trait | What goes wrong | What to do |
|---|---|---|
| Thick, tight binding | Lines curve and darken near the spine | Scan from above, press gently, crop past the gutter |
| Glossy paper | Glare spots wipe out words | Light from two sides at an angle, never straight on |
| Diagrams and photos | Black-and-white conversion destroys shading | Keep illustrated pages in greyscale |
| Small labels and subscripts | Lost at low resolution | Scan at 300 dpi or more |
| Page references in class | Reflow or renumbering breaks them | Keep the PDF page-for-page |
| 600 to 1,000 pages | Free online tools cap pages | Use a tool that takes the whole book |
First decision: cut it or keep it
Cutting the spine is the fastest route to a perfect scan. Loose pages go through a sheet-feeding scanner flat, with no curve and no gutter. Mail-in services work this way: 1DollarScan charges $1.20 per 100 pages, and its FAQ says the book is cut and recycled unless you pay for a return exception. Once it’s cut, you can’t sell it back.
Keeping it bound means scanning spreads from above. A phone on a stand, or an overhead scanner, both work. People who build their own rigs for this share designs on the DIY Book Scanner forum. Expect curve near the spine on thick books, and plan to straighten and crop afterwards.
Setting up for a bound textbook
- Fix the camera height. Use a stand, a pile of books or a tripod, so every shot is at the same distance. Pages then come out the same size, and one crop fits them all.
- Light from both sides. Two lamps at about 45 degrees cut glare on glossy paper and remove the shadow a single light casts across the spine.
- Lay a dark cloth under the book. The dark border is easy to crop away and stops the desk showing through thin pages.
- Shoot the whole spread. Two pages per shot halves the work. Book apps such as vFlat split the spreads, and so does FixMyScan later.
- Scan the front matter and index. The contents and index are how you’ll navigate the PDF.
- Count as you go. After each chapter, check the last page number you shot against the book. A skipped spread in chapter 4 is miserable to find after chapter 20.
Treat diagram pages differently
The fastest way to wreck a scanned textbook is to push every page to pure black and white. Plain text pages benefit: grey paper goes white, faint print goes solid, and the file shrinks. A page with a graph, a micrograph or a shaded diagram loses everything between black and white, and labels on grey backgrounds vanish.
FixMyScan handles this page by page. Plain text pages become black on white, and pages with pictures stay greyscale, so the figure on page 312 keeps its shading while the text around it is crisp.
Clean and index it in one pass
Once the scan is done, the whole book needs the same treatment: margins and gutter shadow off, pages straightened, spreads split, paper whitened, text layer added. Doing that by hand in a PDF editor, 800 pages at a time, is where most textbook projects stall.
- Upload the scan. A PDF or DjVu file of up to 1,000 pages.
- Draw a single crop for the whole book. You set it over 50 pages stacked on top of each other, so margin notes and page numbers that sit at the edge are visible before you decide.
- Let the clean-up and OCR run. Every page is read and hidden text goes underneath, in 19 languages including English, German, French, Latin, Italian and Spanish.
- Download and test it. Search for a term from the index and check that the hits land on the pages the index gives.
Best suited for: students who want one searchable file per course book, on a laptop, tablet or e-reader.
Revising from a searchable textbook
With a text layer in place, the PDF does things the paper book can’t. Search every page that mentions a term before an exam. Copy a definition into your notes. Highlight on a Kindle Scribe or reMarkable and keep the original page number beside every note, so citations in essays point to the printed edition.
OCR is built to read words. We haven’t tested it on equations or chemical formulas, so search for the words around a formula rather than the formula itself.
What this won’t do
FixMyScan reads printed text, not handwriting, so notes you wrote in the margins stay as images. It doesn’t turn the textbook into an EPUB or rebuild its tables as editable text. And it won’t scan for you: it starts from the PDF your scanner, phone or service gives you.
Questions
What is the best way to scan a textbook?
If you can spare the book, cut the binding and use a sheet-feeding scanner or a mail-in service. If you need to keep it, use an overhead scanner or a phone app on a stand, and press each spread gently flat.
How long does it take to scan a textbook?
It depends on the method and the page count, and we haven't timed it. Scanning spreads rather than single pages halves the number of shots for any bound book.
What dpi should I use for a textbook?
At least 300 dpi. Tesseract, a widely used OCR engine, says it works best from 300 dpi up. Small print, subscripts and figure labels benefit from more.
Should I scan a textbook in colour?
Scan in colour or greyscale, not black and white. You can always simplify plain text pages afterwards, but detail lost at capture doesn't come back.
How do I make a scanned textbook searchable?
Add an OCR text layer. The page image stays as it was, with invisible text under it, so you can search for a term, copy a definition or highlight on your e-reader or computer.
Will the page numbers match my printed textbook?
They will if the tool keeps every page in place. FixMyScan doesn't reorder, drop or renumber pages, so page 214 in the PDF is page 214 in the print.
Can I scan a textbook with a phone?
Yes. Book scanning apps such as vFlat flatten the page curve and split two-page spreads. Use a stand or prop the phone so every shot is taken from the same height.
My textbook is 900 pages. Is that too long?
Not for FixMyScan, which takes books up to 1,000 pages. If yours is longer, split it into two volumes at a chapter break.