Finding what changed between two versions of a PDF
A contract comes back from the other side as a fresh PDF, a supplier sends a revised price list, or a draft has been through another round of edits with no tracked changes. Reading the two side by side is slow and easy to get wrong. Comparing them shows the words that were added and removed on each page, and a picture comparison catches what the words alone would miss.
Updated
Which tool you need
A revised contract or draft where the wording matters
The text comparison. Each page's words are matched against the same page in the other file, with removed words struck through and added ones underlined.
A logo, signature, stamp or chart that may have changed
Compare that page as pictures. Both versions are drawn at the same size and every part that looks different is marked in red.
Two scans of the same document
Scans have no words to compare, so the picture comparison is the one that works. Pages scanned at a slightly different angle or position can show as changed even when nothing on them is.
A page was added or taken out near the start
Pages are paired by number, so everything after it would look changed. Delete PDF Pages on a copy of the longer file, so the rest line up, and compare again.
Both texts wanted in your own diff or review program
PDF to Text writes each file's words to a .txt file that any program can compare.
What counts as a change in the words
Each page's text is read in reading order and split into words and punctuation marks, so a changed figure, a new clause or a comma moved from one place to another all show up. Spaces and line breaks are not compared, which means text that simply reflowed onto different lines does not count as changed.
Bold, fonts, colors and sizes are not part of the text comparison at all. When a page has been rewritten so heavily that matching it word by word is not worth the time, the page falls back to showing its earlier text and its later text in full, one after the other.
What the picture comparison sees
The chosen page of each file is drawn 1,000 pixels wide and the two are compared pixel by pixel. Tiny differences from smoothing at the edges of letters are ignored, so identical pages read as identical; everything else that differs is painted red over a faded copy of the changed file's page, with the share of the page that changed.
It draws one pair of pages at a time, when you ask for it, which keeps long documents manageable. A page that only one file has is compared against a blank page, so all of it shows as new.
Which file goes first
The first file is treated as the original and the second as the changed version, so words only in the second are additions and words only in the first are removals. Choosing them the wrong way round turns every addition into a removal; the arrows swap the two without choosing them again.
The full comparison can be downloaded as a text report, useful for keeping a record of what changed between two versions.
What to know before you start
- Pages are matched by their number, not by their content.
- Changes to formatting appear only in the picture comparison.
- Two files at a time, and one page at a time as pictures.
- Neither file can be password-protected; open it with Unlock PDF first.