OCRing Speeds

Different PDF editors OCR PDF files at different speeds. Today at work I had a chance to use FoxIt Phantom for the first time. It includes at 'Quick OCR' option. I tried it and it was fast and relatively accurate, but not so impressive compared with its competitors. The FoxIt Quick OCR tool took 1 minute and 10 seconds to create a searchable version of 39 page chart. While the resulting quality wasn't bad, Nuance Power PDF Advanced took only 36 seconds for the same PDF, and actually rendered noticeably superior OCR text.

Adobe Acrobat -took 2 minutes 9 seconds on the 'Searchable Image' setting, and 2 minutes, 21 seconds using the superior ClearScan setting.

Not a very scientific study, but it appears as though Nuance has a clear edge.

Acrobat Java Script to Find Pages Marked for Redaction

'try67; has posted a javascript here, which you can use to count the number of pages in a PDF file which are marked for redaction. As the discussion on Adobe forum shows some people have not had much luck getting the script to work. The problem can be overcome by inserting the javascript into an Adobe action. Follow these steps:

1. In the Tools console, go to Action Wizard and select 'Create New Action'.

2. Choose More Tools . . . Execute Javascript and click it into the 'Action steps to follow' area on the right. Uncheck the 'Prompt User' box, and click in 'Specify Settings'.

3. Enter this script in the Javascript Editor:

this.syncAnnotScan(); var counter = 0; for (var p=0; p<this.numPages; p++) { var annots = this.getAnnots({nPage: p}); if (annots==null || annots.length==0) continue; for (var i in annots) { if (annots[i].type=="Redact") { counter++; break; } } } app.alert(counter + " pages contain Redaction annotations in this file.",3);

4. Name and save the action.

5. When you click on the new action in the Action Wizard menu, you will be prompted to select multiple PDF files.

6. Click start and Acrobat will start to open the files and indicate the number of pages marked for redactions in each PDF.

It would be helpful if the javascript was modified to generate a textual list for each file rather than showing the counts in successive dialog boxes.

PDF Vole

If you don't have a license for Adobe Acrobat you can review the internal structure of a PDF with the free utility PDF Vole, available for download at http://www.softsea.com/review/PDF-Vole.html . After extracting the contents of the zip file, click on the Java file named, 'pdfvole_20090208.jar' to run the program. You can open one file at a time with PDF Vole.

Unfortunately PDF Vole does not give you the ability to search through the contents of the internal structure.

If you're looking for highlights or redactions try checking the objects towards the bottom of the Object Reference Table menu. Look for one with a type listed as '/Annot'.