OCR'ed PDF-quality not consistent (gray-scale converted to bw-scale)

It seems to me, that OCR’ing a gray-scale PDF (sent to DT3 via ScanSnap) gets converted to bw-scale which results in a file that is much coarser and in less quality:

Here is an example:

(left: “bw”-version after OCR’ing, right: original gray-scale file, both saved in DT 3.9.2; one can also make out the light-gray streak of the original scan, which is gone at the bw-version)

I’ve tried playing with the parameters

defaults read com.devon-technologies.DTOCRHelper
    DPIPDFToImageConvert = 600;
    OutputDPI = 600;

yet to no avail.

Is Abby somehow broken?
How can I reinstall it?

Technical Details:
macOS Monterey 12.6.4
on MacBook Pro 16’’ (2021), Apple M1 Max 10 core, 64 GB memory

Can you turn on the ABBYY OCR logging, to do this:

  • Quit DEVONthink
  • In Finder select the menu Go->Go to folder, copy and paste the the line below and press Go.
    ~/Library/Application Support/DEVONthink 3/Abbyy
  • Copy to this folder the file OCR.plist (223 Bytes)
  • Restart DEVONthink

If you could scan and OCR a document as a grey scale PDF, the select the DEVONthink Help menu and whilst pressing the option key select “Report a Bug”. This will open an email with the log files, if you could also attach both the before and after OCR PDF’s and mark in the email for the attention of Alan.