Simplest scanning methods and equipment

[RWood]

Bob, I'll take "Making an eBook from a Binder of 8.5" x 11" for $100"
(Be sure to phrase your response ....)

Cheapo-cheapo production methods would be to use the scanning part of the multifunction printer such as the HP 6210 All in One that I picked up at CompUSA for just under $100 that includes bundled software for both the Mac and PC. Among that software is a fairly good OCR program.

The scanner part has an automatic feed (no duplex) that can be set to as small as a mass market ("it fits in your hip pocket") paperback. I have scanned, rearranged pages, and OCR on the software with very good results. Sure some ls were 1s; but, nobody's perfect all of the time.

[Bob Russell]

Originally Posted by RWood

Bob, I'll take "Making an eBook from a Binder of 8.5" x 11" for $100"
(Be sure to phrase your response ....)

Cheapo-cheapo production methods would be to use the scanning part of the multifunction printer such as the HP 6210 All in One that I picked up at CompUSA for just under $100 that includes bundled software for both the Mac and PC. Among that software is a fairly good OCR program.

The scanner part has an automatic feed (no duplex) that can be set to as small as a mass market ("it fits in your hip pocket") paperback. I have scanned, rearranged pages, and OCR on the software with very good results. Sure some ls were 1s; but, nobody's perfect all of the time.

That's very interesting...I've been perusing those all-in-ones for such a purpose myself. I notice you can even get duplex scanning for about $400. But if you have auto sheet feeders without duplex scanning, isn't it tough to get the book pages in the right order? And don't you get a lot of misfeeds with inexpensive all-in-ones?

I'm hoping that those are not issues -- a decent $100 book scanning solution would be very impressive. Tell us more!

[RWood]

The first book I converted was an old late 1950s paperback with yellowing pages and glue falling into dust for the binding. All bound edges were clean cut to remove traces of the glue and the rough edges. The pages were not smooth surfaced and quite rough as many paperbacks were in those days. There were only about 2 misfeeds for the whole book. As you noted I then had to rearrange the pages manually. This only took about 20 minutes.

Since then I have scanned other books (I wait until I need a rubber band around the book to hold it together) and the higher quality paper does scan and OCR better. Magizines (and any thin clay coated stock) will not feed and must be done on a page-by-page basis.

From examination in the stores it seems to me that there is little difference between the feeders other than the duplexing as you increase in price from one model from a company to another. I suspect that it is cheaper for the companies to offer a standard unit rather than a unique unit for each model.

[Bob Russell]

That's very encouraging.

Now that I think about it, I suppose a clever person could write a DOS batch file or other kind of script to take the default order you get and rename the images in order. Then a simple review of the pages with windows slideshow in file explorer (or equivalent) would pretty quickly catch and fix any anomalies.

Even one misfeed per 100 pages is probably tolerable, although maybe a little bit of a headache. But we're talking a low-budget solution that's accessible to almost anyone.

One more question... what software were you using for scanning and OCR? Was it what comes with the printer, or did you use your own software also?

[RWood]

As noted in my first post this is a cheapo-cheapo method. Therefore all of the software used was provided by HP on the install CD.

Although I have other (they say more advanced) OCR programs like TextBridge, the results do not change that much when using good originals. Higher resolutions and better OCR programs would be critical if the base language was for example Russian. Since English has a fairly simple character set it is easier than many to OCR. If exotic designer fonts are used then the better quality programs are a must. For common everyday typeset books, the supplied software is quite fine.

[Bob Russell]

Okay, I was wrong... I have yet another question!

If you wanted to create a pdf from the images without doing OCR (e.g. a document with lots of formulas and diagrams), can you also do that with the included software?

I think you may have sold a lot of all-in-one HP printers with this information, by the way!

[RWood]

I don't own stock in HP so if sales increase I will not enjoy any additional advantage.

As for output going directly to PDF I am not sure. I had already installed a full copy of Adobe Acrobat 6 Professional on the machine before I hooked up the printer so I am not sure if the capability is native to the HP provided software or available simply because I had already installed Acrobat.

The software uses TIF as the native graphic format for the saved scans so it should be easy for most othe programs to pick up these formats and create a PDF from them.

[RWood]

I went back and reviewed the install CD that came with the HP, there were no PDF creation programs on it that I could find. All of the documentation on the disk was in HTML.

When I set up the scan from the All-In-One using the "Scan To" button, one of the options is "Adobe Acrobat" Since I had a full copy installed before I installed the HP software I feel that this is where it picked up this option from. When used it transfers the scans directly to PDF files and leaves the Acrobat application open on the desktop with the scanned file ready for me. Likewise, if I pick "Scan to Word" it opens Word and leaves an image in Word -- not OCR text, an image. You must go to an HP application and select OCR with a destination of Word to get text. Any word processor will work, I just happen to use Word.

BTW: The All-In-One also comes with a set of Mac drivers and application programs that are said to do the same as the PC drivers and applications.

[Studio717]

Nag, one of the benefits of the Opticbook scanner is that #1 and #3 of your list aren't necessary. After the first Preview scan and boundary setup, the scanner continually scans only the page area you've set up. And the included software (PC only, much to my dismay) has auto odd or even page rotation.

Bob and RWood, the Mac has built in PDF generation under the Print dialog box, so the HP solution could generate PDFs with no additional cost. (I don't know if there's an upper page limit to the PDFs. I haven't run across one, but I use Acrobat for large projects.)

[RWood]

On the PC side I have genreated 3,000+ page documents with no problem (just a lot of time) so I would doubt that there would be a problem on the Mac.

[Nag]

Studio717:

I was suggesting alternatives, which ought to speed-up the scanning process. Not claiming that opticbook cannot do it.

For example, if the scanner-head doesn't illuminate beyond the page area, then there is no need to use white pages. OTOH, if it scans the entire area and then trims the margins, then using white paper is less stressful on the eye and also no need to close the cover of the scanner for every scan.

One can scan odd and even pages continuously and the opticbook will merge them. However, rotating the book for every scan takes time and effort. Scanning all odd pages first and then all even pages makes the scanning faster.

Try this for a book or two.

Best
Nag

Originally Posted by Studio717

Nag, one of the benefits of the Opticbook scanner is that #1 and #3 of your list aren't necessary. After the first Preview scan and boundary setup, the scanner continually scans only the page area you've set up. And the included software (PC only, much to my dismay) has auto odd or even page rotation.

[jackbrown]

RE: Nag's question--

actually when I scan books, I don't use 8 bit grayscale, I do black and white: the files are smaller, and there's no need for anti-aliasing the text; OCR seems to work better without it anyway.