badger said:
It's not the lens, it is inside Abbyy.
i'm pretty sure dkretz was pulling our legs when he made his reference to "the lens"... ;+)
The difference depends on whether FR rotates the image by 2 degrees before the recognition starts.
excellent detective work, badger. thanks for that. and, oh my word, who would have ever thought that archive.org would instruct finereader _not_ to rotate a page which is crooked... what an asinine decision... in my original suggested workflow for these projects, i recommended that they do an initial shot of the page, and then analyze the results and rotate the _camera_ ever-slightly as required to get a perfectly straight scan. once again, the ludicrous irony of having a high-res scan that is marred in some way -- e.g., by being crooked -- asserts itself... what a bunch of idiots we have in charge. anyway, thanks again badger... -bowerbird