Clearscan is Adobe Acrobat technique that vectorizes fonts and creates "on-the-fly" fonts of them. Output is then very similar to original but filesize is smaller and printing is faster. Very nice when ocr'ing old books and retain original layout and feeling.
What I have been thinking that I should convert scanned images to B/W-images and the ocr them with clearscan technique. I think this will work. It seems to be easier to delete unwanted backgrounds then, because there isn't any or there only some.
I agree ABBYY is powerful for ocr'ing images to text. Downsize is that it your uses installed fonts. Output is not the same then. If you convert books to ePub os similar that wouldn't issue. I am targeting to get "Xerox copy" of scanned books. That's why ABBYY is not what I am looking for.
By the way, when I use options that you expressed I get nicely looking PDF on monitor but when I transfer it to iPad it looks horrible because of reduce images. There might be something I do wrong, though
Thank you very much for the reply!