1

I am about to submit my PhD thesis but I cannot convert the PDF output generated by Overleaf (using the premium version) into the proper PDF-A standard. The document is more than 400 pages long, and I don’t even reach the point where I get to a error log or something (using Adobe Acrobat, the subscription version).

I have also tried various online converting software. None of them work. Do you know what could be the issue? Is my PDF output corrupt? What kind of information would you need, as the experts, need to potentially help solving the issue?

Submitting the thesis in PDF-A format is a requirement. Hence, I really need this to be sorted out - in the best case without much effort since re-formatting the whole document/Latex code might take a lot of time.

Thank you very much for your help in advance.

Kind regards, Max

MaxPy
  • 49
  • Generating PDF-A still is hard. Please tell us, what you tried already, it's hard to guess where the problem lies, without any code oder logs. Also, have you searched the pdf-a tag? – DG' Aug 03 '22 at 21:45
  • I think you should start with pdfx. I only tested it on one of my documents and after loading following packages with their exact options: \usepackage[rgb]{xcolor} \selectcolormodel{natural} \usepackage{ninecolors} \selectcolormodel{rgb} \usepackage{tabularray} \usepackage{luatex85} % Required before pdfx \usepackage[a-3u]{pdfx} % PDF compiant with PDF/Au Level 3, and then tested with verapdf, the test passed. I don't claim this is a solution but might help you to start searching. BTW I compile documents using lualtex engine. – Celdor Aug 03 '22 at 21:52
  • If nothing else is specified and PDF/A conformanc is the only requirement, try printing to file with adobe acrobat and select PDF/A-2b as output: https://helpx.adobe.com/acrobat/using/pdf-x-pdf-a-pdf.html – DG' Aug 03 '22 at 21:53
  • https://tex.stackexchange.com/a/536060/107497 claims to be able to do this. – Teepeemm Aug 03 '22 at 21:54
  • 3
    Create a new project and create a small document there using only your current preamble and a bit text and try if acrobat can convert that. If yes expand the document slowly with text from your thesis. – Ulrike Fischer Aug 04 '22 at 06:20
  • 1
    Using your version of Acrobat, copy 10 pages of your thesis into a new file. Try converting that into pdf-a. If it succeeded, you could kind of bisect the whole PDF. – Keks Dose Aug 04 '22 at 07:30
  • Hi all. Thanks for your suggestions. I tried printing the PDF with Adobe Printing and followed the link you provided. Unfortunately, also this is not working!? It iterates endlessly embedding the fonts and reducing the file size, until after ca. 1 hour it stops and gives an error that the conversion did not succeed. This even happens when trying to simply save the normal PDF into another normal PDF (no PDF-A at this point). I am really worried now and also do not understand the problem as I do not receive any log files. I tried it with Adobe's Preflight and it also yields errors. – MaxPy Aug 04 '22 at 10:45
  • As I have just read your comment, Keks Dose, I will bisect the whole PDF file and try if that works. BTW, I also used other software, e.g. Foxit Phantom, and they all yield problems. Either, they cannot convert at all or the produced/"converted" PDF file cannot be opened at all... I think this is very strange, since I could share the generated output PDF easily in the past (and people could also open it). The Overleaf PDF output is 121 MB large. Maybe that's another reason why there might be problems? Figures are 99% in PDF, with a few in PNG. I used a Helvetica-like (adapted) font. – MaxPy Aug 04 '22 at 10:47
  • If your figures are pdfs, you load them with \includegraphics? If so, why don't you convert them to pdf-a before including them into the thesis? As far as I remember, some Acrobat version have batch processing. – Keks Dose Aug 04 '22 at 19:10
  • @Keks Dose: That is actually a good idea. Might take a while with more than 100 figures, though. – MaxPy Aug 05 '22 at 02:08

1 Answers1

2

Problem re-solved!

Solution: I managed to run Adobe's Preflight tool again, which is integrated in Adobe Acrobat by analysing just a few pages. It stopped at a figure (PDF) that was created in a specific software, which did not fulfill the font embeddment standards. After converting all PDF outputs from this software into JPEG format, the PDF/A conversion worked properly. Make sure in future, to particularly check the figures.

MaxPy
  • 49
  • While it's great that you solved your problem, the answer section is supposed to be for answers only. It'd be nice if you would re-word this answer without the questions (which should have been put in the original question), as a clear answer to the problem described in the question. Thank you. – Miyase Aug 05 '22 at 06:36
  • 1
    Sure, just edited it. – MaxPy Aug 05 '22 at 10:09