OCR is good - but it's not perfect

by: Garry Pierrepont (18 March 2014)

I thought of my own experiences of proofreading books produced by Optical Character Recognition (OCR) software when I read an article written by a colleague of mine recently: Why Writers Need a Fresh Pair of Eyes.

I have proofread a fair few books now out of print and re-produced as Word documents by OCR software. Books by John Creasey - creator of The Baron, The Toff and Gideon of the Yard, Inspector Roger West - represent many of those books.

OCR software is good, but it's not perfect, and that is why proofreaders are needed to correct errors that creep in.

Here are some examples:

• Hyphens normally removed from line ending are retained in proper nouns.
• Spaces often creep in after hyphens, e.g. well-being becomes well- being.
• Spaces can be inserted before question marks, but is it right ?
• Page numbers can be retained, especially if they look like letters, e.g. 5 or 15. 11 can become ii.
• Splodges on a page can turn into random letters.
• Double quotes can become single quotes if next to, for example, capital T.

You have to be vigilant. But you do get to read a decent book into the bargain!

Views: 1441

Other articles on Proofreading and Technology and e-books