How to Reduce a PDF File Size (Without Ruining It)

A PDF is big for one of a few concrete reasons, and knowing which one applies tells you which fix will work. Recompressing the wrong thing saves nothing and damages the document for no benefit.

What makes a PDF large

Three things account for most of the weight. First, embedded images: a photograph saved inside a PDF is stored as a bitmap, and a single high-resolution photo can be larger than all the text in a hundred-page report. Second, fonts: a PDF may embed the complete glyph set of a font even if the document uses a dozen characters. Third, uncompressed streams: page content and images are often stored with no or weak compression, which is the cheapest thing to fix.

Which reductions are lossless

Some reductions discard nothing that affects what a reader sees. Compressing uncompressed streams is lossless — the pixels and text are identical, only the encoding changes. Subsetting a font keeps only the glyphs the document actually uses; if the PDF still contains every glyph of a large font you never used, this alone can be a big saving with no visible change. Stripping metadata — author, producer, timestamps, unused objects — is also lossless in every sense that matters to a reader.

A structural rewrite can recover space left behind by earlier edits. When a document is edited, deleted objects are often left in the file until it is rebuilt, so a clean rewrite can shrink it without touching content.

Which reductions are lossy

Downsampling images is the big one and the one that actually changes pixels. Reducing a photo from 4000 px to 1200 px wide saves a lot of bytes and is invisible at screen size but visible when printed. Recompressing an image to a lower JPEG quality discards detail permanently — you cannot recover it from the smaller file. The distinction matters: downsample and recompress are irreversible, and once applied there is no undo except to go back to the original.

Downsampling also has a floor. Push image resolution below roughly 150 dpi and printed text in the picture starts to look soft, even if the screen looks fine.

The right order to reduce size

Start with the lossless steps: strip metadata, subset fonts, compress streams. Only if the file is still too large should you downsample images. This order preserves quality for as long as possible and only spends image fidelity when you actually need to. When a document is text-only, the lossless steps are usually enough; when it is a scanned batch, image handling is the whole job.

What not to do

Do not convert pages to images to shrink a file. That destroys selectable text and searchable content, and it usually makes the file larger, not smaller. Do not re-save a good PDF repeatedly at lower and lower quality; each pass throws away more detail and the damage compounds. And do not compress before checking whether the size problem is images at all — a text-heavy report that is already using compressed streams and subset fonts has little left to give.

Open the PDF tools — free and private

← All PDF guides