OCR problems - conversion progress not appearing

In the Non-OCR smart group, I can select a file and select OCR / Convert to searchable PDF. Sometimes it works. At other times, the OCR progress information in the bottom left doesn’t appear. It’s therefore difficult to tell whether it is doing OCR or not. This is a particular problem when I try and select a number of documents in bulk to OCR. Any advice here?

Anything logged in the Log panel? And did you check the Activity panel too?

Yes! When I open the activity panel, I see the OCR happening. But why doesn’t it appear in the usual bottom left hand corner?

The Activity pane should appear in main windows and only if the Activity window isn’t open.

I have noticed this behavior too, not sure if it’s a bug or if I was doing something incorrectly.

When I apply OCR to PDFs one by one manually, occasionally (I’d estimate 5% of the time) the conversion progress would not appear in the bottom left corner. If I then (i.e. the Activity window is closed beforehand) use ⌘⌥A to open the Activity window, I would see the progress has started already.

Even if this is a bug, I have not found a way to reliably reproduce it, so I chose not to report it. :slight_smile:

Yes, this description is similar to what I’m getting. It seems to happen more than 5% if you are doing a bulk OCR rather than one by one.

Another related problem is that the activity window disappears whenever it isn’t the active window - unlike the Devonthink window, which remains.

Also, I am getting this error a lot when trying to do OCR:

Screenshot 2021-12-11 at 22.36.18

I don’t have any license pages remaining??

You may want to check out this thread:

Are you running the trial still?

Not running the trial any more. I purchased DT Pro some days ago.

I followed the instructions in the thread posted above, but I am still getting the same problem:

Screenshot 2021-12-12 at 12.10.37

It seems to work for a certain number of pages, then stops working and gives me that error. It says “license pages remaining”, so do we need to pay for a kind of license? It does seem to happen after I’ve done a good number of pages.

I searched here for “ocr license” and found OCR "license pages remaining: 0" - #3 by banshee which may help.

Thanks, but that is the same link that @xurc posted above, and it didn’t work for me.

Could the conversion progress not appearing be independent of the PDF functions?

I use things like the new check file integrity script all the time. Also, verify and optimize databases, plus a couple of scripts of my own that act on all open databases.

The progress bar doesn’t always appear. The best indicator that a script is finished is to see if the script menu entries are grayed out or not.

Did you read the full thread I posted? The solution is indeed a temporary fix, you have to delete ~/Library/Application Support/DEVONthink/Abbyy every time the error occurs.

No. It’s a bug in ABBYY’s library and they’re already aware of it (source).

I did read the thread which you kindly posted; indeed I had commented on that thread also. It wasn’t clear to me from the thread however that the solution was just a temporary fix. I’m surprised that I’m unable to let the OCR run overnight due to this problem, and have to delete and reinstall the OCR component every day. Considering the cost of the software I had expected the OCR to work without issues.

It is indeed very frustrating. Although in DEVONtechnologies’ defense, there’s little they could do to alleviate the frustration since the bug is in ABBYY’s software library.

Thanks @xurc . Maybe consideration should be given to a different OCR API if ABBYY isn’t up to the job.

Sure but that’s also like considering rewiring part of your house. Is such a project doable? Sure. Easily and quickly done? Nope.

No worries, I am troubled by this issue too, after all. :smiley:

As for switching to another OCR solution, besides Jim’s point on the huge development undertaking, I would wager many customers would be unhappy about it too, since its OCR output quality is one of the best.