21

My university requires submitted PDF files to be in PDF/A format.

I tried to find a converter, but they are all very expensive and/or complicated.

How do I convert my existing PDF file into a PDF/A?

Thank you very much!

  • 1
    If it's only word-processing documents with simple formatting, you could open them in OpenOffice and save them as PDF (OpenOffice has a native PDF export but I don't know if it's PDF/A compliant). Other than that, ask your school if they can provide any tools, seeing as it's their requirement to do this. – FrustratedWithFormsDesigner Sep 15 '10 at 15:25
  • 3
    @FrustratedWithFormsDesigner: LibreOffice lets you export text documents as PDF/A, as explained here. – ThatGuy Oct 17 '15 at 16:35
  • @ThatGuy -100 read the question: he was asking to convert PDF!!! – Tomas Aug 1 '18 at 0:52
  • 1
    @Tomas: ¯\_(ツ)_/¯ ...I wasn't answering OP's question. I was clearly responding to FrustratedWithFormsDesigner. – ThatGuy Aug 11 '18 at 18:41
5

For saving existing Word documents to PDF/A all you need is a recent version of Microsoft Word. For creating from other products that print you can use a free PDF/A creator like the one at www.freepdfcreator.org

If you need to validate that a PDF/A file is compliant, you can use our free service at www.validatepdfa.com

Converting existing PDF files to PDF/A in a lossless way is a bit more tricky and not always possible. Acrobat 9 and 10 can do this. Our business, Solid Documents, also sells a product that does this (and other common archiving functions) for $99: Solid PDF Tools

| improve this answer | |
  • 3
    Unfortunately Free PDF Creator often caused the programs I tried to print from to hang or crash, and briefly flashed some popup about "special offers." When I contacted Solid Documents support, I was advised that "Free version is an intermediate release, and we cannot consider its issues." – Mark Berry Mar 22 '13 at 23:05
  • Kurt Pfeifle's answer below is way better. – mb21 Jan 23 '18 at 12:38
16

PDF/A is an international ISO standard for archiving PDFs. The standard requests strict compliance to its set of rules (like: "embed all fonts", "don't use transparencies", "don't use JavaScript", "no encryption",...).

There are a lot of PDFs out there which claim to be PDF/A, but fail a real smoke test. That claim is just a tag in the file's metadata. That tag can make f.e. Acrobat Reader display a special hint when rendering it.

A check for real compliancy requires some rather expansive commercial "preflight" software. Currently I'm not aware of any Free utility to do that job. See also here for some test results: Isartor testsuite.

You can use Ghostscript to (try to) convert PDF to PDF/A. How to do this is documented here (Update: for newer versions here).

But note: this document was updated only very recently [*]. Previous versions of Ghostscript's Ps2pdf.htm did mis-lead users to run a command that created PDFs claiming to be PDF/A but which failed real smoke tests.

How to convert PDF to PDF/A with Ghostscript:

Here is a commandline:

gswin32c ^
   -dPDFA ^
   -dNOOUTERSAVE ^
   -dUseCIEColor ^
   -sProcessColorModel=DeviceCMYK ^
   -sDEVICE=pdfwrite ^
   -o output_pdfa.pdf ^
   -dPDFACompatibilityPolicy=1 ^
    PDFA_def.ps ^
    input.pdf

[*] Note: The problem lays with the parameter PDFA_def.ps. This is a file you need to edit to suite your needs. Ghostscript ships with a sample of it in its /lib subdirectory. This sample will not work as-is without you editing it. How to edit is inside the sample's comments.

| improve this answer | |
  • So is it safe to say that PDF/A isn't a file format per se, but really just a set of compliance requirements for creating a PDF file using existing technology? – JohnFx Apr 11 '12 at 15:22
  • 3
    @JohnFx: in a way, yes. Every PDF/A-1b (or PDF/A-1a) is also a valid PDF in the general sense. But not every PDF successfully validates as a PDF/A*. The PDF/A-* specifications do 2 things: (1) limit the set of features specified in the general PDF-1.4 specification to a subset of allowed ones; (2) require to strictly follow a few rules which in the general PDF spec are "do or don't do -- as you like" (like embedding all fonts used). – Kurt Pfeifle Apr 11 '12 at 16:06
9

I used the following command to convert PDF to PDF-A:

gs -dPDFA -dBATCH -dNOPAUSE -dNOOUTERSAVE -dUseCIEColor -sProcessColorModel=DeviceCMYK -sDEVICE=pdfwrite -sPDFACompatibilityPolicy=1 -sOutputFile=MyOutPutPDF-A.pdf PDFOriginal.ps

If you have a PDF file, first convert it to PS to work with the above code. I tried several times with the PDF file to be directly converted to PDF-A. However, it didn't work.

Here is a way to convert your PDF to PS file:

pdftops PDFOriginal.pdf PDFOriginal.ps
| improve this answer | |
  • Unfortunatelly, these PDFs are not accepted by my university :/ – Tomas Aug 1 '18 at 1:26
  • This was accepted by the US National Science Foundation's system for publication archiving. Thanks! – CephBirk May 5 at 18:23
4

Microsoft Office 2007's 'Save as PDF' tool saves in PDF/A format.

A PDF/A document is just a PDF document that uses a specific subset of PDF that is designed to ensure it is 'self-contained'. Ie it is not permitted to be reliant on information from external sources (e.g. font programs and hyperlinks). From wikipedia:

Other key elements to PDF/A compatibility include:

* Audio and video content are forbidden.
* JavaScript and executable file launches are forbidden.
* All fonts must be embedded and also must be legally embeddable for
  unlimited, universal rendering. This also applies to the so-called     
  PostScript standard fonts such as Times or Helvetica.
* Colorspaces specified in a device-independent manner.
* Encryption is disallowed.
* Use of standards-based metadata is mandated.

Edit:

Since there aren't really any tools to test if a PDF is PDF/A, it's a safe bet that just like you, your university also has no way to test that the document you send them is PDF/A.

It's likely that the only reason they specifically request it is so they can be sure that all the content will be "there" when they open it. They just expressed this requirement rather cryptically (And badly) as being that it had to be PDF/A. So a simple way to test if the PDF meet their true requirement of self-containment is to transfer the PDF and view it from another (preferably offline) computer and ensure that everything appears as it should.

| improve this answer | |
  • 2
    ah, is there a way to determine if an existing PDF is PDF or PDF/A? – ben Sep 15 '10 at 16:37
  • 1
    @ben: I don't know of any tools that determine if a PDF really is PDF/A. PDF/A is an international ISO standard for archiving PDFs. The standard requests strict compliance to its set set of rules ("embed all fonts", "don't use transparencies"). There are a lot of PDFs out there which claim to be PDF/A, but fail real smoke tests. That claim is just a tag in the file's metadata. That tag can make f.e. Acrobat Reader display a special hint. A check for real compliancy requires some rather expansive commercial software. I'm not aware of any Free utility to do that job. – Kurt Pfeifle Sep 15 '10 at 18:29
  • Thanks for the info Pipitas - it then seems likely that even the uni are unable to check for certain the the doc is PDF/A, and so it's not a hard requirement - see my edit. – imoatama Sep 15 '10 at 23:58
  • @imoatama: did you see my other answer? It explains how to convert PDF to PDF/A. If you use -dPDFACompatibilityPolicy=0, your PDF will claim to be PDF/A even if Ghostscript's conversion process noticed things that don't comply and which it could not repair. You may have created such type of 'fake' PDF/As as I mentioned elsewhere... – Kurt Pfeifle Sep 16 '10 at 15:35
  • The MS Word 'save as PDF' also claims the PDF is PDF/A. No need to use Ghostscript. – imoatama Sep 17 '10 at 1:56
2

In macOSX without using pdftops which I wasn't able to install (as @soham.m17 proposed) you can do:

pdf2ps oldPdf.pdf psVersionOfOldps.ps

to convert your pdf to ps format and then:

gs -dPDFA -dBATCH -dNOPAUSE -dNOOUTERSAVE -dUseCIEColor -sProcessColorModel=DeviceCMYK -sDEVICE=pdfwrite -sPDFACompatibilityPolicy=1 -sOutputFile=MyOutPutPDF-A.pdf psVersionOfOldps.ps

to convert in pdf/A.

| improve this answer | |
2

The only way that worked for me on Windows 10 was a slight modification of soham's answer (which for me was working but with errors):

  • converting the PDF to ps (with pdf2ps source.pdf temp.ps)
  • converting back the ps to PDF/A using Ghostscript 9.22 with

gswin64c -dPDFA -dBATCH -dNOPAUSE -dPDFSETTINGS=/prepress -dNOOUTERSAVE -sColorConversionStrategy=UseDeviceIndependentColor -sProcessColorModel=DeviceRGB -sDEVICE=pdfwrite -sPDFACompatibilityPolicy=1 -sOutputFile=destination.pdf temp.ps

(You might need to add before gswin64c the complete path, which for me was C:\Program Files\gs\gs9.22\bin\ -- the same for pdf2ps, in this case the path for me was C:\Program Files\gs\gs9.22\lib\)

If the resulting file is too big you can omit the -dPDFSETTINGS=/prepress option. Without this option the resulting file is smaller (even smaller than the original in my case), but degradation in image quality can be noted.

The resulting PDF seems to be conform to PDF/A even without the -dNOOUTERSAVE option, although less fonts are embedded (maybe someone can explain the difference).

When I say "conforms to PDF/A" I mean not only that the PDF/A "banner" shows up on Acrobat Reader but that the file passes the online validation provided by https://www.pdf-online.com/osa/validate.aspx

I don't know how much such online validation service is reliable, however it managed to detect non-conforming files that were showing the PDF/A "banner" on Acrobat Reader (which therefore does not seem reliable in this respect).

| improve this answer | |
0

Simple answer:

You can install Libre Office , it's free

Open the PDF and save as PDF... [check the PDF/A option]

...you're done!

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.