Scanning in your own books

[edembowski]

I find page crop to be less than useful. I go though page by page and use the 'erase' function. I haven't really timed it, but it usually takes about 1 song ~ 100 pages :-)

[wayrad]

As far as farming the work out to China, I see some problems:

1) You're talking about what amounts to a scanning operation on a professional scale, not the normal hobbyist scale. I love my Opticbook, but it's not particularly heavy duty, and the availability of support and parts is notoriously bad. Sending one off to China and expecting it to handle that sort of workload sounds like a recipe for disaster.

2) Were you planning to give the guy a scanner and a Finereader license, have him scan like mad, and then spend the rest of your life proofreading? Remember, scanning is the easiest part.

3) How do you plan to proofread? Get the hard copies back? Wait until he's all done and transfer the Finereader license so you can do it onscreen? Or buy *two* Finereader licenses (it's not particularly cheap)? (edited to add: On further thought (and after coffee), I suppose the best way would be to send raw page images and keep the Finereader for yourself. File sizes would be huge, and you would still be doing about 90% of the work, but you would only need one license.)

4) What happens down the road when you find out he missed a page? (yes, it happens. Quite often). Or cut off an edge by mistake?

5) How is he going to send you the files? Over the Internet? Thousands of books? With the hard copies still in his possession, or not? Or bring them across international borders on a laptop or storage device? I'd suggest checking those copyright laws...you may feel you could defend it if challenged, but would you want to risk the trouble and expense of doing so?

[wayrad]

Originally Posted by edembowski

I find page crop to be less than useful. I go though page by page and use the 'erase' function. I haven't really timed it, but it usually takes about 1 song ~ 100 pages :-)

Ah, I tried that once. Drove me out of my skull. Perhaps I should've tried background music.

[edembowski]

Originally Posted by gazza

...
If Google cannot get it right -- and it hasn't -- what chance for mere mortals?

Pretty good, thanks. I have family who works at Google, and I know a little bit about their operations (although nothing more that what's publicly available). They don't proof the work. It gets scanned and spot-checked. They save the images and the text. You can do a much better job yourself.

Finally (how the man does do on) the law of copyright is perfectly clear that if you buy a book you can copy it for your own use. Publishers lie in their teeth to make you believe otherwise but that is in the Berne and Geneva Convention. That does not mean you can copy it and then put it on one of these Bit torrent thingies. But you can do it for your own use. For certain sure.

Absolutely. Many publishers 'discourage' that concept, but it's certainly true. I've proofed books I've scanned, found typos in the original work, and reported it back to the author. The author in question is aware of what I'm doing (and why) and he has voiced no objections. He has thanked me for letting him know about the typos, and has taken time to explain to me a bit about the publishing process.

Originally Posted by wayrad

As far as farming the work out to China, I see some problems:

I don't think farming it out is realistic, but here's my take on the scanning process in general:

1) ...I love my Opticbook, but it's not particularly heavy duty, and the availability of support and parts is notoriously bad. ...

I'm using two Kodak 8.1mp cameras. I scan 2 pages in about 7 seconds, counting page turns.

2) ...then spend the rest of your life proofreading? Remember, scanning is the easiest part.

Absolutely correct. A lot of this depends on what you're after. I find that the more I love a book, the more time I spend proofing it. With Finereader, 95% of the "errors" are words not in it's dictionary. That leaves 5% error, which can be quite a bit. Many times, I'll give a book a quick proof, make the ebook and mark errors as I read the book.

3) How do you plan to proofread? ...

For me, what works best is comparing against the scanned image and having the book next to me in case of a bad scan.

4) What happens down the road when you find out he missed a page? (yes, it happens. Quite often). Or cut off an edge by mistake?

Yeah, that does happen, and it's a pain. I usually deal with this when I first process the images, and before I bring it into Finereader. I've had to re-shoot a page or two and it can be cumbersome getting those pages in the right order but it's very doable.

Originally Posted by wayrad

Ah, I tried that once. Drove me out of my skull. Perhaps I should've tried background music.

I'd go out of my mind without the background music!

- Ed

[wayrad]

Originally Posted by edembowski

Yeah, that does happen, and it's a pain. I usually deal with this when I first process the images, and before I bring it into Finereader. I've had to re-shoot a page or two and it can be cumbersome getting those pages in the right order but it's very doable.

Yeah, it's perfectly fixable when you catch it promptly. But separating the workflow in time and space will make that and other tasks enormously more complicated - all for the sake of offloading the least time-consuming part of the job. As you say, not realistic.

[pilotbob]

Of course you could scan the book to PDF and get a Kindle DX or DR1000/800 and just read the PDF. No proofreading needed in this case since you aren't doing OCR, just making a graphical representation of the pages.

The extra money you spend on the larger format reader will be more than paid for in all the time you would have spent correcting OCR errors.

BOb

[ascherjim]

Originally Posted by igorsk

BTW, ABBYY is having a promo right now: buy FR9 and get a free upgrade to FR10 in a few weeks:


New features in FR10:
Achievements in OCR Accuracy and Performance
ADRT® analyses a multipage document as a single entity
3rd Generation Camera OCR: Reads Phone Camera Photos
Enhanced Usability – New Quick Tasks and Interface Revisions
Saving E-books to HTML Chapters and Flexible HTML
Powerful PDF Compression
Further Improvements in Page Layout Analysis
New Recognition Languages – Korean and Yiddish

I'm intrigued by the new feature in FR 10 concerning E-books and HTML. Not only because it indicates new (or increased) recognition of what many of us will be using scanners (and their software) for.

[HarryT]

Originally Posted by pilotbob

Of course you could scan the book to PDF and get a Kindle DX or DR1000/800 and just read the PDF. No proofreading needed in this case since you aren't doing OCR, just making a graphical representation of the pages.

The extra money you spend on the larger format reader will be more than paid for in all the time you would have spent correcting OCR errors.

BOb

The downside will be HUGE files. A scanned book can easily be 30MB+, compared to perhaps 1MB for the same book in a "textual" format.

[pilotbob]

Originally Posted by HarryT

The downside will be HUGE files. A scanned book can easily be 30MB+, compared to perhaps 1MB for the same book in a "textual" format.

Yes good point. So, an ebook reader that supports PDF and high capacity SD cards if your best bet.

BOb

[pilotbob]

As far as doing OCR on the Mac, have you looked at PDFPen Pro for Mac. It supports PDF editing and OCR. I'm not sure how good the OCR is. But, once you have a text based PDF you can convert it to ePub or something like that.

http://www.smileonmymac.com/PDFpen/

BOb

[eLiNK]

Originally Posted by HarryT

The downside will be HUGE files. A scanned book can easily be 30MB+, compared to perhaps 1MB for the same book in a "textual" format.

It depends on your skills.
Hint: ScanKromsator

[pilotbob]

Originally Posted by eLiNK

It depends on your skills.
Hint: ScanKromsator

I'm not sure what that is. But, my scanner's software lets me save a scan directly as a PDF. I haven't tried a book though, a flatbed scanner to scan a book could take forever.

I have wanted to get a Fujitsu ScanSnap (maybe Harry can get me a discount) and digitize all my documents and put them into evernote. If I get one, I'll try to scan a book with it.

BOb

[=X=]

Originally Posted by HarryT

The downside will be HUGE files. A scanned book can easily be 30MB+, compared to perhaps 1MB for the same book in a "textual" format.

Actually that's not a big deal, with a 4GB card.

The real downside is that the fonts are not as clear on a PDF image than a true type font.

=X=

[DaleDe]

Originally Posted by =X=

Actually that's not a big deal, with a 4GB card.

The real downside is that the fonts are not as clear on a PDF image than a true type font.

=X=

It cannot be searched, and you cannot reflow it or change font sizes. To me those are the bigger limitations.

[wayrad]

Actually, since the O.P. doesn't seem absolutely set on retaining the hard copies, perhaps he should consider a sheet-feed scanner. That would definitely cut scanning effort, and likely cost less than shipping all those books around and paying for someone's time to do the scanning on a flatbed.