Running the following command
ocrmypdf --redo-ocr --output-type pdf US10039726_ifi.pdf US10039726_ocr.pdf
Takes significantly longer to complete the scanning phase in the current version ocrmypdf than in version from about 18 months ago. The time for the scanning phase in the difference versions was as follows:
v16.4.2 ~5 seconds
v16.4.3 ~many minutes (cf. #1378)
v16.6.0 ~60 seconds
v17.4.1 ~60 seconds
I have reproduced this on two Windows machines, but performance on RockyLinux was ~5 seconds on the current version, so I suspect the issue is specific to Windows. I attach the PDF I'm using for testing.
US10039726.pdf
Running the following command
Takes significantly longer to complete the scanning phase in the current version ocrmypdf than in version from about 18 months ago. The time for the scanning phase in the difference versions was as follows:
v16.4.2 ~5 seconds
v16.4.3 ~many minutes (cf. #1378)
v16.6.0 ~60 seconds
v17.4.1 ~60 seconds
I have reproduced this on two Windows machines, but performance on RockyLinux was ~5 seconds on the current version, so I suspect the issue is specific to Windows. I attach the PDF I'm using for testing.
US10039726.pdf