|
The text layer of DJVU files from Archive.org is usually a raw OCR scan intended for locating text on the image layer.
My experience of it is that it is relatively poor as a source for an e-book without extensive editing by eyeball comparison with the image text.
On the books I have worked on this way I've imported the OCR'd text layer to LibreOffice Writer then edited that on one PC while viewing the DJVU Image on another. Only when happy with the edited text version have I converted it to EPUB for input to Calibre. Doing the sort of text manipulation needed to clean up such poor OCR scan text needs a proper Word Processor.
BobC
|