Wednesday, March 19, 2014

An eReader and the Library

I have a first generation Nook, and it's getting a little unreliable. So I looked around and stumbled on a Kobo Mini on sale and decided to give it a go. It's small, lightweight; perfect for sticking in a pocket or purse to have with me wherever I go. I was able to add my Nook books via Adobe Digital Editions, as promised, and most of my small collection of Kindle books don't have DRM, so those went on easily with Calibre. I'm using Calibre for most of my ebook management and I haven't had any problems with it interfacing with the Kobo.

I wondered how it would do with library materials, so I went hunting. As it happens, almost all of our downloads are PDF. Even most of our ebooks are set up to allow PDF downloads of chapters, but not full book downloads in ebook format (ePub, or whatever). And PDFs don't work great on small screens.

The Kobo Mini does have a magnification option for PDFs, but it's a pain shifting the window around to read. Articles in 2 columns aren't bad--a single column is just about right, but full pages are awkward.

I decided to try the conversion features of Calibre and converted a number of PDF downloads to ePubs. For the most part, especially for the publisher-generated PDFs with good metadata and background text, it worked fine. Calibre specifically notes that they convert using the embedded text if available, rather than re-OCRing the image. Some of the line-by-line formatting can be a little weird, with odd line wraps and page numbers stuck in the middle of sentences, but it's certainly readable.

A multiple column PDF article, without publisher metadata, is a disaster to convert. Single lines from each column end up together, and it's not worth trying to pick them apart. It would be less work to scan each column separately and then OCR.

We have several scanners, which have pretty much replaced photocopiers in our library. I gave those a whirl, too. When scanning to PDF, you pretty much get exactly what you'd expect: an image of each page, a little grainy on the basic settings, but readable. Doing a "searchable PDF", which does some basic OCR, did produce something I could convert to ePub without too much trouble, but definitely had OCR issues. Things like 'j' instead of 'i' and 'rn' instead of 'm'. I did a couple of pages scanned to Word, as well, which had the same OCR issues, but I could correct them. That was certainly the most work, but also produced the best ePub, once I corrected all the oddities in Word and saved to RTF for conversion in Calibre. This is essentially the same process that produces Gutenberg books: scan, OCR, correct, convert.

Our public library has an Overdrive collection, and those ebooks that are primarily text based are quite readable (via Adobe Digital Editions). I tried a cookbook, and that didn't do well. It's no better than a PDF, and I have to scroll around the pages in the magnifier.

So, at some point in the past, someone asked me to recommend an eReader. I'd say, if all you want to do is read text, then an inexpensive e-Ink reader, like the Kobo Mini or a basic Kindle, is a nice reader. I really like the e-Ink technology, it's easy on the eyes and the batteries last and last. (My first gen Nook, with its failing battery that wasn't too hot to begin with, still lasts about as long as my new-ish smart phone!) If you want to do anything else, including lots of reading things like PDFs downloaded from your local university library, you are probably better off getting a tablet and loading eReader apps for your books.