File size
Why a PDF gets large, and what shrinks it
A PDF that is too large to email usually owes its size to a few causes rather than to page count alone.
Images are the main weight
High-resolution photographs and scans account for most of the bulk. Lowering their resolution is the change that saves the most space.
Other contributors
- Fonts embedded in full instead of in part.
- Pages saved repeatedly, leaving unused data behind.
- Attachments and layers carried inside the file.
The trade-off
Shrinking a file reduces quality somewhere. Check a few pages afterwards, particularly small print, and keep the original until the result has been confirmed.
Approvals
Getting a document approved and signed
Documents that pass through several people raise different questions from documents that one person edits.
Decide who changes what
Reviewers who only comment need different access from people who edit. Permission settings let a document be read without being altered.
Signatures differ
- A typed or drawn mark placed on a page is the simplest form.
- Digital signatures tie the signer to the exact content, so later changes can be detected.
- Requirements vary by country and by type of document, so it is worth checking what is expected.
Track the versions
When changes arrive from several people, comparing two versions shows exactly what moved and avoids rereading the whole file.
Scans
Scanned pages and text recognition
A scanned page is a picture of text, which is why it cannot be searched or edited until it has been recognised.
What recognition does
Optical character recognition reads the image and adds a text layer, making words selectable and searchable while the page looks the same.
Getting reliable results
- Straight, well-lit scans at a reasonable resolution read most accurately.
- Unusual fonts, handwriting and tinted backgrounds lower accuracy.
- Names, numbers and dates are worth checking by eye after recognition.
Order of steps
Run recognition before compressing, because heavy compression can blur the characters the process depends on.