Turning Scanned PDFs with facing pages into single pages

[jimteacher]

After scanning in books, I ended up with two facing pages that were often askew, causing OCR errors even with Acrobat 8 Pro. The manual two-file drag and drop takes too long, and other software didn't seem to work easily.

But then, AHA! I googled one of your forum archives, and down at the bottom a suggestion was made which didn't seem to work one of the readers - perhaps because of some Mac problem, but it worked WONDERFULLY and INSTANTLY for me, and should work for anyone with a windows machine.

Here's part of the post:
You can use Adobe Acrobat (not the Reader):

1. Choose Adobe PDF printer
2. Set Page Scaling to "Tile Large Pages", set Tile Scale to 100% and overlap to 0
3. Print the document with Adobe PDF printer as a new .pdf file

When I did this with an old book (with hand notations), it ended up with each of the facing pages on a separate page with a lot of white space around it, easily cropped with alternate page cropping.

After alternate page cropping, I used the OCR recognition, which straightened up each of the pages almost perfectly for the best OCR possible.

I had to clean up a few badly skewed pages - and one that had a fold-out chart, but that only took a few minutes, as opposed to the terribly tedious drag and drop from two files method.

Result: Nice, clean, straight, OCRed text which could be used for reading on any PDF reading device.

(In addition, I found that good new copiers have a SCANNER feature which will scan in books and magazines almost instantly as well, instead of the tedious 1-minute-per-page thing which can take hours to do.)

I hope this helps some folks, as it as been a wonderful find for me!

[owl123]

I'm pretty sure this was my post and I posted it here, on MobileRead. :-)

Thanks anyway. I'm sure many more people will now use this method.

[jimteacher]

Having played with it for a day with other books, I found out why one might have problems with making the "Tile Large Pages" work.

The double page has to be cropped so that both pages are almost exactly in the middle so that it splits right at the page division.

I messed around with page sizes to try to push a page over about 1/2 inch and it didn't work until I went back and cropped the double page spread again to make the division almost exactly in the middle.

Also, it's important to scan the book in very consistently, so that the division is in the same place throughout the book, which isn't that hard to do, as you have about an inch to play with if the book lies flat.

One might have to play with the page size a little too, under properties, to get it nice and tight.

This may sound like a lot of trouble, but once you've done it a couple of times, it just takes a few minutes to set it up on even a 300-page book.

I've taken hours to scan in a book without the big copier and dragging from one Acrobat file to another. This took about 30 minutes total for a whole book.

[ahi]

unpaper (http://unpaper.berlios.de/) is a pretty amazing utility that can do the splitting of pages and more. It is a bit arcane, but easy enough to figure out and very powerful!

[nhimclc]

Would you mind uploading some screen shots of how you did it with Acrobat because I tried and I could not make it work on my Mac. Thanks so much.

[Ragmana]

It took me awhile (several unrelated attempts) to get this to work right, so I'm adding these details "to the record" in case anyone else needs them. I'm using Acrobat 8.0/ OSX 10.4.11; on other setups you might need to do some things differently.

1. Problem: No Adobe PDF Printer (from File > Print > "Printer" dropdown menu)

I (finally) figured out my PDF printer just wasn't installed. Go to Help > Repair Acrobat Installation. Make sure Adobe PDF Printer is checked. Run the repair.

Add the printer and you're set. (Print > Printer > Add Printer > Default Browser > select the PDF printer and install it.)

2. The there is too much whitespace/pages are "too big" for the contents
(i.e., odd pages have a large right margin, even pages a large left margin, and the page contents shift from side to side as a result.)

You could crop the alternative margins, but I found it much easier to play with these settings:

Originally Posted by jimteacher

2. Set Page Scaling to "Tile Large Pages", set Tile Scale to 100% and overlap to 0

Increasing the Tile Scale makes the contents fill the page a little better, so with a little trial-and-error you can get everything just about centered. Adjusting overlap a little an get everything lined up perfectly. The preview thumbnail is not helpful, though; you'll have to check the new PDF after printing to see whether the scale was too large/still too small. (A too-large scale will cause some or all pages to be split in half horizontally.)

3. All OCR-ed text is gone

Still have not figured out how to preserve OCR-ed text for files passed through the PDF Printer. (Can this be done?) Re-OCR-ing the file recompresses the image (especially if some pages need to be de-skewed properly), so there are some more compression artifacts. Just split the pages before running OCR, if you can.