thank you thx1138v2

You probably can't tell the difference if all the OCR worked correctly. In general, OCR works the same way as typing - the characters go into a memory buffer and then are saved to disk.

The best way to distinguish the two is that the OCR'd version may have some words that just don't fit the sentence. Here is an example: "Possibly it may still contain a grateful Asmodeus, who will be lot out of his narrow prison cell..." - "lot" should have been "let". The OCR program picked up the "e" as an "o" probably due to weak imprint from the original document. That typically would not be a typing error and since "lot" is a word, a spelling checker wouldn't pick it up either.

Another example from the same document:"Having selected the induction apparatus and type of bulb, tie next important consideration is the vacuum." "tie" should be "the"
On Roentgen Rays by Nikola Tesla

Of course these could have been typing errors so it's not foolproof and, as mentioned, if the original document was very clear the OCR might not have made any mistakes. Or if the person using the OCR thoroughly proof reads the document all the errors might have been corrected.