I checked out the new OCR function with DTTG 3 and encountered an oddity which might be an OS glitch. Running OCR on a two column A4 portrait page resulted in a wrongly stretched out A4 landscape page. The page gathered a searchable text layer but at the cost of vandalized proportions. Another page of the same making came out just fine.
My request beeing: Would it be possible to add more options to DTTG in the future to restrict certain OCR settings (even if this was to be recognised as a simple bug)? When working batches it would not be feasible to check every single page again.
(I would be happy to provide the sample, just let me know.)
The problem occurred on my iPad Pro 11" currently running iPadOS 14.3. Since the document is about a health care issue I would like to extract the relevant pages only. Would that be sufficient?
Appreciated. Here are the four pages, the order is original1—OCR1—original2—OCR2. The first pair went right, the last pair went wrong.Sample_OCR_aspect_ratio_fail.zip (3.7 MB)