← All posts

Is flattening the same as printing to PDF?

You have a filled-in form, or a marked-up draft, and it needs to go out as a record rather than as something anybody can re-type. You asked how. One person said flatten it. Another said just print it to PDF, which they have been doing for years and it works fine.

Both produce a file where clicking on an answer does nothing. So the question is reasonable: are they the same thing with two names, and if not, which one leaves you worse off?

They are not the same thing, and searching does not tell you, because nobody has written the comparison down. The top result is a video. The nearest article compares flattening to compressing, which is a different pair entirely.

One is surgery on a file, the other is a rebuild from what the page looked like

Flattening opens your document, changes the specific objects it was told to change, and writes the rest back. Everything it is not aimed at — the text, the bookmarks, the page geometry, the attachments, the structure tree — is carried across untouched, because it was never involved.

Printing to PDF does not open your document in that sense at all. It draws every page through the operating system's printing machinery, catches the marks that come out and builds a new document around them. Nothing is deleted, because nothing was copied in the first place. What arrives is whatever survived the trip.

That difference is the whole answer, and it explains every consequence below without you having to memorise a list.

Why printing loses things it never touched

The printing system speaks the language of marks on paper. Put this glyph at this coordinate in this font. Stroke a line from here to here. Paint this image into this rectangle.

That vocabulary has no word for "this rectangle links to page 40". It has no word for "this is a level-two heading", or "this is a text field named client_name whose value is Ms Okonkwo", or "this sticky note belongs to Priya and was written on Tuesday". Those things are not properties of ink. A laser printer would have no use for them, so the printing path was never built to carry them, and a virtual printer at the other end can only write down what it was sent.

So a printed copy routinely arrives without:

The text usually survives, and that is not the reassurance it sounds like

You will read everywhere that printing to PDF turns your pages into pictures. That is mostly wrong, and the truth is more awkward than either version.

Most virtual printers — Microsoft Print to PDF, macOS's Save as PDF, Chrome's — send real glyphs with embedded fonts, so the result is genuinely text. But there are three possible outcomes and they look identical on screen:

  1. Real, searchable text. What you usually get.
  2. Text that copies out as gibberish. The page carries glyph indices into a subsetted font, and the mapping back to actual characters — the ToUnicode table — is optional. Some drivers do not write one. The page reads perfectly and searching for "invoice" finds nothing, because as far as the file is concerned those shapes are numbers 43, 81, 22 and not letters. This is a favourite trap of print-driver output and it is why "I can see the word but Ctrl+F cannot" happens on files that were never scanned. It is a different cause from the usual one, which is a scan, and it has no repair short of OCR.
  3. A picture of the page. Acrobat has a Print as image checkbox under Advanced, some drivers fall back to rasterising when a page uses transparency, and printing from a phone often does the same. Then nothing is text and the file is several times the size.

Flattening cannot produce any of those, because it does not re-draw the text. The glyphs and their character mapping are the same bytes they were before.

The page quietly changes size as well

A physical printer cannot print to the edge of a sheet, so the printing dialogue defaults to fitting your page inside the printable area. A 595 × 842 point A4 page comes back at about 96% of itself, centred, with new white margins. Do it twice and it shrinks twice. If your default paper is Letter, an A4 document comes back 612 × 792, which is not A4 and will be noticed by whoever has to bind it.

Our PDF inspector reads page sizes, the form field count, whether the file is tagged and whether there is a real text layer, in your browser without uploading anything. Run it on the original and on the printed copy and the differences are on the screen in front of you.

What flattening does instead

It finds each form field's appearance — the drawing instructions the reader was using to show your answer — writes those into the page's own content stream at exactly the rectangle they occupied, and then deletes the field. The full mechanism is worth reading if you are locking a form, because it is what makes the result hold up in software neither of you has heard of.

The part that matters here is what it does not do. It does not re-draw your text, does not resize the page, does not rebuild the outline, does not touch the tags, and leaves links clickable. Comments are a separate tick box, off by default, so a draft still in review keeps its notes live. Run Flatten PDF and the notes tell you how many fields became page content, how many comments were stamped down and how many links were left alone — which is a much better position to check your work from than looking at a printed copy and hoping.

When printing to PDF is the right answer anyway

It is not folk nonsense. There are cases where it is the only thing that works:

If you want that re-render without a printer dialogue, the free command-line version is better behaved than any driver, because it is deterministic and you can see what it did:

pdftocairo -pdf in.pdf out.pdf

pdftocairo ships with poppler-utils. It draws each page exactly as a viewer would, keeps the text as text with a proper character map, and drops the interactive layer — the same trade as printing, minus the paper-size surprises and the gibberish. Ghostscript's gs -sDEVICE=pdfwrite does the same job with more control and more ways to change the appearance of transparency, so check the result.

Three things neither of them is

Neither removes the marks. If the comments should be gone rather than frozen, that is Remove annotations from PDF, which strips notes, highlights, links or form fields without redrawing the page. Be careful with the form fields option: deleting a field deletes the answer that was typed into it.

Neither is a compressor. Flattening saves a little by dropping annotation objects. Printing usually makes the file bigger, because fonts get re-embedded and images get re-encoded at whatever resolution the driver chose. If the printed copy has to go by email, Compress PDF is the tool, and it works on images rather than on text — so a printed text document may not shrink at all.

Neither flattens transparency or layers. Designers use "flatten" for a third, unrelated thing: resolving overlapping transparent objects into opaque ones for a press. Our flatten does not do that, and does not touch optional-content groups either. If you genuinely want pages turned into pictures, do it deliberately with PDF to JPG at a resolution you choose rather than whatever a driver decided, and understand that you are throwing the text away.

The short version

Flattening removes the mechanism. Printing rebuilds the document from its appearance and loses everything that was not ink. If the file is one page and disposable, the difference does not matter. If it has a contents page, working links, a title that means something, or anyone using a screen reader at the other end, it matters a lot — and only one of the two knows your document was ever more than marks on paper.

Tools this post uses: