I ran a new test on 2015-06-02 with my CanoScan LIDE 25 (a simple scanner), OS X 10.10 using the driver for OS X 10.6 and ABBYY FineReader for Mac 12.1.1. The scan was exported to a Microsoft Word document (DOCX) and opened in CafeTran, with two different filters.
The scanned TIFF:
http://cafetran.wikidot.com/local--files/new-ocr-test/Scan-example.TIF.zip
How it looks in FineReader:
How FineReader exported the OCR'd TIFF to DOCX:
http://cafetran.wikidot.com/local--files/new-ocr-test/Scan_example.docx
The blue characters are marked as 'uncertain' by FineReader (this is an option).
The Word document loaded as a normal document in CafeTran:
and
The Word document loaded as an OCR'd document in CafeTran:
and
FineReader Mac vs. FineReader Windows
