For several months now, the engine for converting to editable text and images has become quite unreliable: one or two in every hundred pages gets recognized just totally incorrectly, with all letters cut in half, some bits recognized as background: super complex bunches of images and renderables when it should just be one text field on a blank white page.
This will happen on a given scan of a given page no matter what else I try to do to it first, it seems, but it isn't an issue with my scanner because it even happens on old files scanned years ago on completely different equipment. These days I have to scan a document, check which pages Optimize messes up, and then go back and make another physical scan of those pages, optimize them and stick them in by hand. I never used to have to do this extra work in all my years as a user.
Dev team, what gives? Can you change the engine back to an older version or fix this issue? Is this not happening to anyone else?
1
Reply
AdChoices
