I've been doing quite a lot of scanning of textbooks to pdf. Since the scanner produces an image, OCR is necessary to convert it to "searchable text," which is one of the real advantages of having it digital. The ABBYY OCR that came with my scanners does a much better job than previous others, but still leaves quite a few "clinkers." The converter provided with the scanners for conversion from image to pdf is from Nuance, the makers of DNS, and also includes the ability to convert between .jpg, .pdf, Word, and a couple of other formats. Conversion accuracy is heavily dependent on the accuracy of the OCR for scanned documents. Best results probably are obtained when the starting point is "plain text," either Word documents or .txt, with "formatted text" (.rtf) fairly close. In a very few cases I've made the pdf from .jpg, converted .pdf to .txt, corrected the errors, and then made a new .pdf from the corrected text; but that's only justified for "very important documents" that just don't come out right straight from the .jpg scan to .pdf. I haven't messed with text-to-voice or voice-to-text, so only know what I've seen in the comic books (called advertisements - or more often "marvelous special offers" - by the authors). John
|