Does anyone know anything about OCR software?

May 5, 2004 9:45 AM

Does anyone know anything about OCR software? [more inside]

I work for a small university press and we are interested in digitizing our backlist. Nothing fancy, just converting them into some sort of digital format. I assume OCR software is what we need, but I know nothing about it. What options are available? How much should we expect to pay? Should we be buying software or sending our books to some company to do it for us? What kind of equipment would we need (scanner, computer, etc.)? Any and all advice would be greatly appreciated.

posted by rorycberger to Technology (9 answers total)

I've found it to be really inaccurate and lossy (and it really depends on how clean and crisp, and in what font the original pages were printed in, too)...a lot of scanners come with OCR software, so you may not need to buy any separately, but people seem to like OmniPage. You'll need proofreaders to fix all the mistakes and dropped words. If you can slice up the book so that the pages are scanned flat, you'll do better. I've found that scanning at a higher resolution, and upping the contrast before running the image of a page through an OCR program helps a bit. I don't know what outside vendors charge, but it may be worth it.

Google answers has a thread on it, too.
posted by amberglow at 10:25 AM on May 5, 2004

there places like this around, but i don't know who's good or how much. It's free to get quotes and more info tho. : >
posted by amberglow at 10:29 AM on May 5, 2004

My Vote: ScanSoft's TextBridge Pro 11 (formerly OmniPage).

Plenty of lower version numbers available cheaper ... and the upgrades don't appeare that significant (except for maybe XP integration). Check ebay.
posted by RavinDave at 10:35 AM on May 5, 2004

We've done this both ways at my press. If you have the interns to scan and proof, it can be done in house. However, it *is* a giant pain in the ass. If you can afford it, I'd go outside; if you're going to do your whole backlist at once, you can probably get a pretty good packaged price.
posted by dame at 12:04 PM on May 5, 2004

dame - Any specific recommendations for outside companies that do this? What can we expect to pay?

We probably will have interns (i'm one now, but graduating in a few weeks), but i don't think we could afford to pay them for the hundreds if not thousands of hours it would take to do the several hundred books we're looking to convert, not to mention that this kind of bitch work doesn't really fit into the press's vision of what interns are supposed to do. In retrospect, I guess my original question probably should have been about outside vendors.
posted by rorycberger at 12:33 PM on May 5, 2004

I should have specified a little more when i said "we can't afford..." We are looking to set aside a chunk of money for this project, but that is separate from the money budgeted to pay interns (which is not a lot of money). I guess one could be used to pay the other, but i'm not sure.
posted by rorycberger at 12:36 PM on May 5, 2004

I'm trying to hunt down our old info (we did this a while ago). I'll get it to you soon.
posted by dame at 3:35 PM on May 5, 2004

Of course as soon as I post that, I find it. We used BookSurge. They charged 25 cents/page, plus $99 for the cover. They're a short-run POD place, though and we used them so we could get small runs of the backlist titles. We also only did about ten books. Hundreds should get you a *much* better deal.

If you aren't printing, you may want to check out compositors; I don't know how their rates run at all, but getting a quote is super easy. Dix! and Westchester Book Corporation have both been recommended to me. The former is a little more expensive because it is a little smaller. However they are apparently super-pleasant. The latter is a big corporation. Fine, but without the extra touch.
posted by dame at 3:50 PM on May 5, 2004

rorycberger - this is my industry, the company I work for does this exact work. Email a phone number to the address in my profile and I'll give you a call and tell you what you could expect to pay and try to help with other suggestions.
posted by vito90 at 9:52 AM on May 6, 2004