Why is my PDF so big? Find what takes the space

Look inside before you fix anything

Guessing wastes time. A file that is big because of scans needs a different fix from one that is big because of fonts, and a compressor cannot help with weight that is not pictures. So the first step is a look inside.

1. Drop the PDF into the inspector above (or open /inspect).
2. Read the section Why it is this size. It shows how many images there are, how many bytes they take and what share of the file that is, the image formats, the largest image with its page and pixel size, and Everything else: the bytes that are not pictures.
3. Read the line above the tables. When images are more than 60 percent of the file, the inspector says so in words: compressing them is where the size goes.
4. Check Scanned pages (image, no text) in the Pages section. A number above 0 means pages that are photographs of paper, not text.
5. Check Attachments in the Security section, and the Fonts list.

Two real readouts as a reference. Our 1-page scanned lease (scan-lease.pdf, 34 KB) shows 1 JPEG image, 97 percent of the file, and 1 scanned page. Our 500-page text file (117 KB) shows 0 images and 1 font: every byte is text and page structure.

Images are the usual weight

In most oversized PDFs, more than 9 of every 10 bytes are pictures. They come in 3 kinds, and each has a size signature you can spot in the inspector.

Scans. A scanner stores every page as a photograph of the paper. An A4 page at 300 dpi in colour is 2480 by 3508 pixels; saved as a JPEG it is about 1.8 MB, so our 10-page test scan weighs 17.6 MB. The signature: 1 image per page, all the same pixel size, and a scanned-page count equal to the page count. Why scans grow this big, with the math, is in Why scanned PDFs are huge.

Phone photos. A 12 megapixel phone photo is 4032 by 3024 pixels and about 3 MB as a JPEG. Turn 5 of them into a PDF and you have 14.7 MB before anything else. The signature: JPEG images (shown as DCTDecode) of 3,000 to 4,000 pixels across.

Design exports. A brochure, a CV from a design tool or a slide deck often carries every photo at its original size, even when it shows on the page as a thumbnail. Some exports also store photos losslessly (shown as FlateDecode), which is far bigger than JPEG: our test file has one 1200 by 900 photo stored that way, and that one picture is about 3 MB.

The fix for all 3 is the same: re-encode the pictures at a resolution that fits how the file is read. Our 10-page scan goes from 17.6 MB to 1.7 MB on the Email preset of Compress PDF, and the text-heavy parts of a file are not touched at all.

Fonts: usually cheap, sometimes not

A PDF can carry its fonts so the text looks the same on every device. Most writers embed a subset: only the letters the document uses, a few KB to a few dozen KB per font. That is cheap.

Two patterns are not. A font embedded in full, with every glyph for every language it supports, can add hundreds of KB, and a font family with 4 or 6 weights multiplies that. And some tools embed the same font again for every page or every merged document, so a 20-file merge can carry 20 copies.

The inspector lists every font by name, embedded or not, and on how many pages it appears. It does not give each font’s bytes. If Everything else is large and the font list is long, fonts are the likely weight. Our compressor does not subset or remove fonts, and it says so; the honest fix is to export again from the source with font subsetting on (most office apps and design tools have it).

Hidden weight you cannot see on the page

Some bytes belong to nothing a reader ever sees:

• Old versions. Many editors save changes by appending them to the end of the file. The old objects stay inside, unreachable. Compress PDF sweeps them out and reports “Removed N objects nothing pointed at”.
• Duplicate pictures. A logo or letterhead stored once per page instead of once per file. The inspector counts identical pictures once, so the extra copies show up under Everything else. The compressor merges them and reports how many went.
• Page thumbnails. Small preview images some older writers embed in every page. No current reader uses them; the compressor removes them.
• Attachments. A PDF can carry other files inside it, such as a spreadsheet or the original image files. The inspector counts them and lists their names. The compressor keeps them, because they are content, and no tool on this site removes them yet.

Sometimes big is correct

This is the part most “shrink your PDF” pages skip. A big file is not always a problem.

A 300 dpi archive scan of a contract holds the detail you may need to zoom into years from now. A print master for a brochure must keep its full-resolution photos, or the printer gets soft images. Shrinking those files is a loss, not a fix.

The habit that works: keep the big original, and compress a copy for the job in front of you (the email, the portal, the chat). Our compressor always writes a new file and never overwrites yours.

And when compression cannot help, because the file is text, fonts or vector drawings, the tool says so and hands the original back unchanged. PDF will not compress covers what to do then: send fewer pages, split the file, or export it again from the source.

When you need a deeper breakdown

Our inspector answers the question most people have: is it the pictures or not. It does not break Everything else down further into fonts, drawings and structure. If you need that level of detail, Adobe Acrobat Pro has a feature called Audit space usage in its PDF Optimizer, which lists the bytes taken by fonts, images, bookmarks, forms, comments and more, each as a share of the file (Adobe’s help page, checked on 11 October 2026). It is part of the paid Acrobat Pro.

For everything else, the path is short: inspect, then compress if the pictures are the weight. Both run on your device; use each once with a connection, then turn Wi-Fi off, and they still work.

Compress PDFPresets or a size target, shows what changed. Free, runs on your device.Open

The whole toolkit for getting a file down to size, with all the levers in order, is in Make a PDF smaller.