Quote:
Originally Posted by El Duderino
First off, I've yet to install KOReader so these are pre-install questions...
KOReader has an OCR function, which is awesome, but how accurate do you find it? I suspect it depends on the image quality? Also, with a Kobo Aura (original), will the OCR process chew through the CPU causing battery drain and sluggish behaviour?
Before installing KOReader, I've been ingesting PDF's into Calibre, but only those with a text layer. I'm looking for as close to an e-pub experience as possible, so my assumption is Reflow works best with text layer as opposed to OCR'd image layer.
Is this assumption anywhere close to accurate?
Thanks!
|
The OCR will definitely affect your battery somewhat (it is based on the free tesseract OCR). The good news is, it is not necessary for the reflow feature at all. Reflow works with the k2pdfopt engine which actually rearranges the picture level of the PDF. So reflow will work any any PDF that you feed into it. The output quality depends largely on the picture quality of the PDF, and it is (like always) advisable to clean up your PDFs a bit before reading. There are various tools to do just that, like “briss” or “ScanTailor” (both help crop out margins and separate double pages, the former is quick and easy, the latter is really thorough and a bit more complicated) and “Tesseract” OCR which should help you with cleaning up your scans.
If you want to see how Koreader reflows your PDFs before installing, give “k2pdfopt” a try.