Preflighting your own PDF, free

Check the file you are sending, not the file you made

The design document is your intention. The PDF is the artefact. Every check in this lesson runs against the PDF, because that is what the press receives.

You do not need Acrobat Pro. Two free packages cover nearly everything: poppler-utils (pdfinfo, pdffonts, pdfimages, pdftoppm) and Ghostscript. Both install on Windows, macOS and Linux, and both run on a laptop that could not open Illustrator.

The seven checks

1. Page size and boxes.

pdfinfo -box job.pdf

Confirm TrimBox equals the agreed finished size and BleedBox extends past it. Previous lesson.

2. Fonts.

pdffonts job.pdf

Every row must show yes under emb. A no means the font is referenced but not embedded, and the press will substitute. A six-letter prefix like ABCDEF+ means a subset, which is normal and good. A font you did not choose appearing in the list means something substituted silently.

3. Image resolution.

pdfimages -list job.pdf

Read the x-ppi and y-ppi columns. Anything well below your target is a problem; anything enormous in pixel dimensions next to a small placed size means a clipped image was not cropped on export.

4. Colour and ink coverage.

gs -q -o /dev/null -sDEVICE=inkcov job.pdf

You get four numbers per page: C, M, Y, K coverage. A page you believe is mono showing cyan, magenta and yellow means your “black” is a four-colour build. A page showing coverage far above the press’s ink limit needs attention.

5. Separations.

gs -o sep-%02d.tif -sDEVICE=tiffsep job.pdf

One TIFF per plate, spot plates named. If you expected two plates and get five, you have accidental spots or RGB that separated. Next lesson covers reading these.

6. Overprint. Ghostscript can simulate it:

gs -o overprint.png -sDEVICE=png16m -dSimulateOverprint=true -r150 job.pdf

Compare against a normal render. Anything that disappears was set to overprint and should not have been. This catches the vanishing white object.

7. Structure.

qpdf --check job.pdf

Reports damage, broken cross-reference tables and recoverable errors before the RIP finds them.

A script you can keep

#!/bin/sh
f="$1"
echo "== BOXES ==";  pdfinfo -box "$f"
echo "== FONTS ==";  pdffonts "$f"
echo "== IMAGES =="; pdfimages -list "$f"
echo "== INK ==";    gs -q -o /dev/null -sDEVICE=inkcov "$f"
echo "== CHECK ==";  qpdf --check "$f" 2>&1 | head -5

Save it as preflight.sh, make it executable, and run it on every file before you send one. Thirty seconds, and it has caught every failure in this module at least once.

The GUI options

• Scribus has a built-in Preflight Verifier that runs before export and reports missing images, low resolution, overset text, non-embedded fonts and transparency in a flavour that forbids it. It is good and it is free.
• Acrobat Pro has the most complete preflight available, with editable profiles and fixups.
• PDF24, Ghostscript front-ends and online validators exist. Be careful about uploading a client’s unreleased artwork to a free web service; read what the service says it does with your file, and default to local tools.

What preflight cannot tell you

It cannot tell you the phone number is wrong, the client’s name is misspelled, the photograph is of the wrong building, or the price is out of date. Those are the errors that actually get reprinted, and the only defence is a human reading the proof.

Print the PDF at 100%, read every word out loud, and have somebody who did not design it read it too. Both of those catch things no software will.

Rendering a page to look at

Sometimes you want to see what the file contains rather than what a viewer chooses to show you.

pdftoppm -r 300 -png -f 4 -l 4 job.pdf page

That writes page 4 at 300 ppi as a PNG, rendered by poppler rather than by Acrobat. Open it in GIMP and zoom in on the small text, the hairlines and the edges of any shadow. Rendering artefacts that were hidden by a viewer’s anti-aliasing become visible, and you are looking at something much closer to what the RIP will produce.

For a long document, render every page at a low resolution and flick through the images. Missing content, a page in the wrong orientation and an accidentally blank page all show up in a contact sheet faster than in a scroll through a viewer.