Turn a scanned book chapter into notes you can search
Scan each page flat, square and in even light, convert the PDF in parts of up to 10 pages, then check numbers, names, columns and footnotes against the scan. OCR reads printed text well but slips on look-alike characters, so proofing the details is the real work.
General information, not legal advice. Copyright rules differ by country, so check what you may copy.
Why a chapter needs more care than a handout
A book page carries more than a handout does: two columns, footnotes, a running title on every page, hyphenated line ends and page numbers. Each one can leave a mark on the converted text. The aim of this guide is to stop most of them at the scanner and catch the rest in one careful read.
OCR, optical character recognition, reads the letters in a picture of a page. It is good, not perfect. Characters that look alike, such as 1 and l or 0 and O, are the usual slips, so numbers and names are where to look first.
Chapters are also long. A free conversion covers 10 pages per file, so most chapters need splitting into parts. That suits proofing anyway: 10 pages is about as much as you can check in one sitting.
From scanned chapter to note in six steps
- Check what you may copyConvert only material you have the right to use, such as a chapter you are allowed to copy for study. Rules differ by country and by licence, so check yours.
- Scan one page at a timeSplit a two-page spread into two images. Press the page flat, square it up, keep the light even and leave the margins in. On a flatbed scanner, 300 dpi is a sensible setting for printed text.
- Make parts you can proofBuild one PDF per 10 pages or fewer and name them part-1, part-2 and so on. Each part counts as a file against your 3 a day. Split at a section break where you can, so no paragraph is cut between two files.
- Convert and pick your appDrop each part in and choose Notion, Obsidian or another app. Read the Preview tab with the scan open beside it. The tally under it says how many paragraphs were joined and how many repeated lines were removed.
- Proof against the scanCheck, in order: the title against the contents page, then dates and quantities, then names and technical terms, then footnotes, then any table cell by cell. Look at every page end, where paragraphs and footnotes meet.
- Join the partsEach part starts with its own # title when it opens with a heading, so paste them in order and turn the second and later titles into ## headings. Headings that make notes searchable shows the levels to aim for.
Scan problems and what they leave in the note
| Scan problem | What shows up in the note | What to do |
|---|---|---|
| Page curls at the spine | Bent lines and misread words near the gutter | Press it flat, or scan one page at a time. |
| Shadow across the page | Smudged letters turned into the wrong ones | Use even light and keep your hand and phone out of it. |
| Two pages in one image | Columns from both pages run together | Split the spread before you convert. |
| Two columns or side boxes | Reading order may be wrong near a box or quote | Check those pages; reorder by hand if needed. |
| Page skewed or tilted | Lines slope and letters are misread | Square the page up and re-scan. |
| Margins cropped off | A heading or number near the edge goes missing | Keep the edges when you scan. |
| Small print, low resolution | Look-alike characters swapped | Fill the frame with the page, or use 300 dpi on a flatbed. |
Re-scanning a bad page flat and square is usually quicker than correcting its text by hand.
What goes wrong with book chapters
Footnotes land mid-paragraph
Footnote text at the foot of a page is read as ordinary lines, so it can sit in the middle of the text where the page broke. Move it to the end of the section.
Numbers and names are off
A date with an l for a 1, or a surname with a changed letter, reads perfectly well and is still wrong. Check them against the page.
Hyphens end up in the wrong place
A word split across two lines is joined back up, and a short list of prefixes such as self- and cross- keeps its hyphen. A hyphenated name can still go wrong, so look at those. Fix OCR line breaks explains the joining rule.
A figure leaves only its caption
Pictures, diagrams and photos are left out. The caption usually stays as text, so add a screenshot where the figure matters.
The free page limit cuts the chapter
Files over 10 pages are converted up to page 10 on the free allowance. Split the chapter first so nothing is lost.
Margin notes are garbled
Handwriting is not promised. Treat any result from it as a rough draft, and see Handwritten notes to Markdown.
How ocrtomarkdown helps
ocrtomarkdown reads scanned pages with OCR, joins broken lines and split words back into paragraphs, puts the headings in order under one title, keeps lists and simple tables, and removes the page numbers and running titles that repeat. Pictures are left out. Convert on Scanned notes to Markdown. Free: 3 files a day, up to 10 pages each, no account. Files over 10 pages are converted up to page 10 on the free allowance; a page code converts the whole file, up to 150 pages. See pricing.
Scanned book chapters: your questions
Can I convert photos of the pages?
The converter takes PDF files. Most phones turn photos into a PDF in a few taps, with Scan Documents on an iPhone or Google Drive on Android.
How many pages can I convert for free?
3 files a day, up to 10 pages each. Longer chapters need splitting, or a page code for up to 150 pages in one file.
What scan resolution should I use?
On a flatbed scanner, 300 dpi is a sensible setting for printed text. With a phone, fill the frame with the page and keep the light even.
Will two-column pages read in order?
The converter reads columns in order where it can. Check pages with side boxes or pull quotes, where the order is hardest to get right.
What happens to footnotes?
They are read as ordinary lines and can land where the page broke. Check each page end and move them to the end of the section.
Are page numbers and running titles removed?
Yes, when the same line repeats three or more times. A line that appears once or twice is kept, because it might be content.
Is it all right to convert a book chapter?
Only if you have the right to use it. Copyright rules differ by country, and this is general information, not legal advice.
Will it read my notes in the margin?
Handwriting is not promised, so treat any result as a rough draft and keep the scan beside it.