How to Reduce PDF File Size Without Ruining It

A PDF that will not fit through an upload portal, or that bounces back from an email server, is one of the most common document problems there is. The fix is compression — but compression is not a single dial, and knowing what actually consumes the bytes tells you how much room there is to save.

What actually makes a PDF large

Ranked by how much they typically contribute:

1. Images. This is almost always the answer. A 12-page color scan stored at 600 DPI can reach 80MB; the same pages at 150 DPI grayscale are often under 5MB. Embedded photos, full-resolution screenshots saved as PNG, and scans stored as uncompressed bitmaps dominate everything else.
2. Embedded fonts. A document that embeds whole font families instead of only the glyphs it uses carries dead weight. This is severe for Chinese, Japanese, and Korean fonts, where a single full font can add 10–20MB.
3. Uncompressed content streams. Older exports and some generators leave text and vector instructions uncompressed. Modern writers compress these by default.
4. Duplicated resources. A logo embedded once per page instead of once per document multiplies quietly.
5. Metadata and thumbnails. Usually kilobytes, occasionally megabytes if preview images were stored at high resolution.

The practical consequence: file size tells you what to expect from compression. A 30MB scan has enormous room to shrink. A 900KB, 40-page text-only report has almost none — and no compressor can change that, because there is nothing redundant to remove.

The three levels, and when each is right

Open PDF Compression (single files up to 100MB) and you get three levels. They differ in how aggressively images are downsampled and re-encoded.

Low compression — preserve highest quality

Keeps images essentially intact and recompresses the streams. Expect modest savings, often in the range of 5–20%.

Use it when the document will be printed, archived, or reviewed closely, and you only need to shave off some weight. It is also the safe choice when you do not know how the file will be used.

Medium compression — balance size and quality

Downsamples images toward screen-appropriate resolution and re-encodes them with lossy compression. On image-heavy documents, savings of roughly 40–75% are typical.

This is the right default for the everyday cases: email attachments, form uploads, sharing with colleagues, posting to a site. Text and vector graphics stay sharp; only the embedded images lose detail, and at normal viewing sizes that is usually invisible.

High compression — smallest file size

Aggressive downsampling. Scans can shrink by 85% or more. The trade-off becomes visible: photographs go soft, thin lines and light gray text can blur or disappear, and small print may become hard to read.

Use it for portals with a hard size cap where text legibility still matters more than image fidelity. Then check the file — do not assume.

Reading the result honestly

The tool reports Before, After, Compression Ratio, and Space Saved once it finishes. Use those numbers as a diagnosis, not just a score:

• Medium produced a 60% saving? The file is image-heavy, exactly as expected. You now know where the weight was.
• Medium produced a 4% saving? The file is text- and vector-heavy. Switching to High will not help much either, because there is little image data left to discard. Look at the alternatives below instead.

Whatever you do, open the compressed file and check the pages that matter — the signature block, the table with small print, the photo on page 9 — before you delete the original. Compression is lossy at Medium and High, and it is not reversible.

When not to compress

• Legal, court, and archival copies where fidelity is part of the requirement.
• Print production files that depend on 300 DPI images and embedded color profiles.
• Digitally signed PDFs. Recompression modifies the document and invalidates the signature.
• Files you still need to edit precisely, since re-encoded images will not come back.

Alternatives that often beat compressing

Compression is post-hoc damage control. If you control the source, these get better results:

• Re-export from the original. If the PDF came from Word, Google Docs, or a design tool, exporting again with a "minimum size" or "reduce file size" option often produces a smaller file with text still lossless, and no lossy re-encoding of anything.
• Rescan at a saner resolution. 150–200 DPI in grayscale is more than enough for reading and printing a document. Scanning at 600 DPI color and compressing afterwards always looks worse than scanning properly once.
• Split instead of compress. If you only need to send part of the document, send that part — see how to split a PDF. A quarter of the pages is a bigger saving than any compression level.
• Remove what you do not need. Blank pages, inserted cover sheets, and duplicate appendix pages all cost bytes.

Privacy

Compression runs entirely in your browser — the file is never uploaded, which matters if the document is confidential or covered by a data agreement. The same caution applies to any online tool you use: check whether it processes locally or uploads, and for sensitive documents prefer local processing.

Related guides

• How to merge and split PDF files — merging usually grows a file; this is how to cut it back
• How to password-protect a PDF — encrypt after you have settled on the final size
• PDF tools — merge, split, compress, encrypt, decrypt, rotate