The reason Batch OCR is still required is because IDOL needs to get a text rendition from somewhere, and the solution prior to the pagination features was to send the documents through Batch OCR prior to IDOL indexing so the text rendition was present and accounted for as IDOL and OCR were not directly linked.
The client and 32-bit OCR engine performing OCR in the interim do not need to be installed on the server, just any workstation that has an ODBC connection to the OnBase database. The documents can be accessed by the client OCR engine since they will simply be in the database/diskgroups when the batch moves into the Full Page OCR scan queue step.
You can modify the OCR Format, but I would recommend just creating a new OCR Format in the meantime and assign that to the document type(s) being OCR'd for Full Text Indexing. That way you don't have to modify the default Format and lose any settings. And then switch the OCR Format back to the default on the document type(s) when pagination can be turned on.
