PDFA-1b with OCR doesnt works

218 Views
Last Post 10 November 2022

TobiasD posted this 04 November 2022

Hello all,

I have installed the latest version of DNv5 for a customer.

This requires as output profile, PDFA + OCR.

I created a workflow where I put both in the output profile. Unfortunately it does not work, I checked the output PDF with the program "verapdf".
And this check shows that the PDF does not comply with the PDFA1b standard.

Then I made a second test, with 2 workflows, in the first workflow the output profile is only PDF-OCR, then in the 2nd workflow is PDFA(without ocr) as output profile. When testing via "verapdf" the result is OK, but the file is no longer searchable.

Info: It is a simple license with the "Omnipage Engine" without any additional extensions.

Thank you very much for your ideas.

ocr verapdf pdfa

Attached Files

2 Comments

Order By: Standard | Newest | Votes

luca.scarpati posted this 04 November 2022 - Last edited 04 November 2022

Hi Tobias,

absolutely no bug.

The check for PDf/A compliance is strongly suggested to be done with the official tool form Adobe and everything needs to be checked and confirmed against the official Adobe one. It is widely known that extra tools may have different "profiles" or different "rules" or even bugs.

For example VeraPDF (which is very popular in Germany) always suffered of bugs reporting false errors, see here for example: https://tex.stackexchange.com/questions/426946/pdf-a-1b-adobe-preflight-verification-passes-and-verapdf-fails.

And you can find in the history many more so might be possible also yours are false positive.

In fact if you analyze for example the second report where it says the N parameter of the profile isn't existing or not matching that's actually not true, because if you check with Adobe the N parameter is correctly existing and set to 1 as PDF1.4 specifications are reporting, so at least this one is false positive .. and for sure the other error too.

Please rely on the official tool only (Adobe, the creator of the PDF standards), as external software's often have bugs even if the government decided to use them. How do they deal if the validation software has a bug just like the example I've linked you above? eheheh