OCR produces blank pages

[thomas e komsta]

Looked through the topics but didn't see anything related.

I open a pdf and select Document | OCR Pages. It goes through the motions, opens a new document and all page(s) are blank. Tried multiple docs. Even created a plain text doc from Open Office and saved as a pdf. Same result.

Can you point me to where 19 other people have already asked and you replied?

Windows 7 Home Premium SP1 - 6.1.7601
PDF-XChange Editor V3 Build 300.0 (Download today)

[Stefan - PDF-XChange]

Hello Thomas,

Which build of the Editor are you using? Is it the very first beta or the current build 300?

I have a copy of build 301 here on my machine - and OCR is working correctly in it. This build should be available publicly either later today or on Monday.

Best,
Stefan

[thomas e komsta]

PDF-XChange Editor V3 Build 300.0 (Download today)

PDFXVE3.exe File Version: 3.0.300.0

No hurry. I'll hang out for 301 and see what happens. Address it then. Since I saw no other one posting about it, just assumed it was something with setup/configure on my end.


Thanks for the quick reply though.

[Stefan - PDF-XChange]

[Walter-Tracker Supp]

Can you tell me the language you were using and the settings you had selected for OCR? (Language, output type, accuracy, etc)? I suspect this is some issue related to the settings choice that should be resolved in the release version which handles defaults, error checking, etc, a lot more robustly.

Can you check the preference for OCR in the Edit->Preferences->OCR tab?

Also, FYI, the official release version has a new OCR run dialog where you can select (and review) settings on the fly whenever you select OCR.

[Willy Van Nuffel]

I suppose that Thomas will send you his answer too, but I just like to already let you know that OCR creates blank pages when you just "Create New Searchable PDF". I did a few tests with English and French, with Low, Medium and High Accuracy... (by use of PDF-Editor pre-release build 301).
It is only the Dpi setting that makes a difference:
- At 300dpi the result is a blank page;
- At 600dpi it keeps the original scan, but there is no text layer.

Something that I also did remark (maybe due to preceding test versions) is that the "Output type"- field was empty by default in the "Run OCR"-dialogbox. At Edit > Preferences it was also empty. After setting a default choise, this problem was solved.
When the "Output type"-field was empty, then the OCR-process kept hanging at "Rasterizing". I had to Close the PDF-Editor (via right click on the PDF-Editor button in the Windows taskbar and Close window) and restart it.

[Walter-Tracker Supp]

Thanks; we will investigate the cause of this (probably related to the default settings that ship with the editor). Meanwhile, we have put in a warning that will inform you if you need to check your OCR preferences.

Your feedback certainly helps.

[thomas e komsta]

Walter/Sefan:

Updated to 301.0. No more blank pages. Other issues with the way output looks after trying to edit but will address that in another thread if necessary. I have not had much use with editing so it could be user error.

Will have to look through existing posts to see if it's been brought up yet.

Thanks again.

[Walter-Tracker Supp]

Sometimes the "invisibility" of the OCR text layer is lost if you edit it; this is a known issue and will be addressed in the future.