They changed the names of the libraries. (There are a number of other languages besides German that have used these typefaces, including Finnish, so I’m not sure which language pack to use for those)
With Tesseract 4 deu:
Aktum Dienflags den 3. März 1891. Bor verfammeltem Regierungsrathe.
a few errors, suggesting that it’s been trained on this typeface family, but not reliably.
With Tesseract 4 frk:
Aktum Dienstags den 3. März 1891. Vor versammeltem Regierungsrathe.
Cute. It reminds me of my favorite pub in France: they offered beer from Warstein, which they pronounced „Warfteiner“ because the Fraktur „s“ on the label resembles an „f“.
Hey. I’ve tried to convert an pdf to a searchable pdf by using the DTTG OCR engine. After the OCR process my new pdf+searchable text file had another layer of text. If I do the same on on my Mac in Devonthink 3 the pdf is also searchable but I do not see the convertet text in another layer. That means after the OCR process there are two layers of the text (one aof the image and one from the OCR). Did I made some mistakes in my process?
The text layer is always written in a clear font, so seeing the overlaid text in a black font is very odd. Is this happening on all your OCR’d documents in DTTG?
Yes! I tried 4 different documents, everytime the same issue. I tried the same 4 documents on my Mac Version of Devonthink 3 and everything was fine. No double text layer there…
but to specify: only if I use the OCR process on my iPad. If I run the process on my Mac the documents are shown without the black text layer on my ipad…
)