How to Reduce PDF File Size Without Losing Quality

Why PDFs get big in the first place

A PDF is best thought of as a container. Inside that container you can have text, vector graphics, embedded raster images, fonts, form fields, attachments, page-level metadata and revision history. The total file size is just the sum of those parts plus a small overhead for the structure that ties them together. That sounds obvious, but it is the key to compressing intelligently — instead of running every PDF through the same generic "compress" button, you ask "what is heavy in this particular file?" and only attack that part.

For most files, the answer is one of three things. First, raster images that were saved at print resolution (300 DPI) when the PDF will only ever be read on screens. Second, scanned pages, where the entire page is a single big image even though it visually looks like text. Third, fonts that were embedded in their entirety instead of subset to just the glyphs the document actually uses. Tackling those three covers nearly every real-world case.

Image compression — the single biggest win

If your PDF was exported from Word, Pages, Keynote, Google Docs or LaTeX with images dragged in from a phone, the images inside almost certainly weigh far more than they need to. A photo from a modern phone is around 4000 × 3000 pixels and 5–8 MB compressed. The author drops that into a document, sizes it down visually to 4 cm wide, but the underlying file is still the original 4000-pixel JPEG. Multiply by 12 such images in a report and you have a 70 MB PDF that visually looks identical to a 4 MB version.

The fix is to downsample images to the resolution that actually matches their on-page size. A simple rule: 150 DPI is excellent for any document that will be read on screens; 200 DPI is a good universal target if the document might also be printed; 300 DPI is for high-end print and almost never necessary for everyday work. Most desktop PDF editors expose this as "Reduce file size" or "Optimize" — when they do, they are mostly doing image downsampling under the hood.

The other half of image savings is the codec. JPEG works well for photos. PNG works well for screenshots and diagrams with flat colours. A PDF that mistakenly stored your photos as PNG can easily be three or four times bigger than it needs to be — and a PDF that stored a logo or diagram as JPEG will look blocky for no reason. Modern optimizers detect this and re-encode appropriately.

Font subsetting and embedded fonts

Fonts are the second sneaky source of size. When you embed a font in a PDF, you can either embed the full font file (every character it contains, including all the ones you did not use) or a subset containing only the glyphs that appear in the document. A complete OpenType font with full Unicode coverage can easily be 10–30 MB. Subset to the 200 letters and punctuation marks an English-language report actually uses, and it drops to a few dozen kilobytes.

Most PDF generators subset fonts by default these days, but older tools and some specialist exporters still embed full fonts. If your PDF is unexpectedly big and you cannot see why, look at its file properties — Adobe Reader and most other tools list embedded fonts. If you see the same font appearing five times because each chapter was concatenated from a different source, you have a candidate for re-saving the file.

Scanned documents — a different problem

Scanned PDFs are a special case because they are not really PDFs of text — they are PDFs of pictures. Each page is one big raster image. That has two consequences: the file is huge (one A4 page at 300 DPI in colour is around 5 MB), and you cannot search the text or copy from it. There are three sensible compression strategies for scans, and they involve different trade-offs.

Removing leftover bloat

Even after images and fonts are addressed, PDFs can carry a surprising amount of dead weight: unused objects from earlier revisions, embedded thumbnails, JavaScript actions, unused form fields, and metadata streams from the source application. None of this is visible to the reader, but it all adds bytes. Most "Optimize PDF" features include a "remove unused objects" pass, sometimes labelled "linearize" or "save as optimized". If your file size still looks high after addressing images and fonts, run that pass and the file will often shrink another 5–15%.

One specific source of bloat to know about: incremental saves. PDFs support saving changes by appending to the end of the file rather than rewriting it, which is fast but accumulates over time. A document that has been edited and saved 30 times by Acrobat may be three times its theoretical minimum size. A "save as" instead of a plain "save" usually cleans this up by writing the file from scratch.

Hitting common email size limits

Most of the time when someone asks "how do I make this PDF smaller" the real question is "how do I get it under my email provider's attachment limit". Useful numbers to know:

• Gmail: 25 MB per outgoing message, 50 MB incoming. Above 25 MB Gmail offers a Drive link automatically.
• Outlook.com / Microsoft 365: 33 MB attachment cap by default, sometimes lowered to 20 MB by corporate admins.
• Yahoo Mail: 25 MB.
• iCloud Mail: 20 MB.
• Corporate gateways: often 10 MB. If you do not know the recipient's limit, 10 MB is a safe target.

If a single PDF is too big even after compression, splitting it into multiple files and sending them across two messages is almost always a better experience than uploading to a sharing service. Our PDF splitter can extract specific page ranges into separate files in one step.

What not to compress

Some PDFs should not be reprocessed even if they are large. Two important categories:

• Documents with digital signatures. A PDF signature is a hash over the file's bytes. The moment you re-save the file, the hash changes and the signature is invalidated, even if the content looks identical. If a contract has been digitally signed, send the original.
• Legal or audit-trail documents. Some regulated workflows require the exact bytes of the file to be preserved. Re-saving for size will produce a "different" file even though it reads the same. Check your organisation's policy first.

Quick summary

For everyday cases, the entire compression playbook fits on one hand:

1. Identify what is heavy: photos, fonts, or a scan.
2. If photos: downsample to 150–200 DPI and re-encode as JPEG.
3. If fonts: subset, do not embed full files.
4. If a scan: lower the resolution, drop colour if possible, optionally OCR.
5. Save as optimized to drop leftover bloat.
6. If still too big for email, split the file into ranges instead of further degrading quality.

That is essentially what every "compress PDF" tool on the internet is doing under the hood. Knowing which step actually applies to your file means you can avoid the ones that would lose quality unnecessarily.