← All guides

Why Some PDFs Barely Compress

Run a 40MB PDF through a compressor and get a 38MB file back, and it's easy to assume the tool didn't work. Usually it did work. There just wasn't much in the file that a compressor is able to touch.

What a PDF compressor actually compresses

A PDF file is a container holding several very different kinds of content: vector drawing instructions, embedded fonts, plain text, and, usually the biggest contributor to file size, embedded raster images, most commonly JPEGs from scanned pages or inserted photos. "Compressing a PDF" almost always means finding those embedded JPEG images and re-encoding them at a lower quality setting, which is where the size reduction actually comes from.

Why a PDF with little to no images barely shrinks

Vector graphics, embedded fonts, and plain text are already compact by nature. A page of text takes up a few kilobytes regardless of how many words are on it, because it's stored as instructions and character codes, not pixels. A compressor re-encoding JPEGs has nothing to do to a document built entirely from text and vectors, since there's no image data to shrink. A 40MB "PDF" made of a hundred text pages was never going to compress much no matter what tool you use. The size is coming from somewhere compression doesn't reach, or the size estimate itself was off.

What actually determines the win

• Scanned documents: usually compress well. Each page is one large JPEG, and re-encoding it at lower quality has a real, visible effect on file size.
• Photo-heavy reports: compress well for the same reason. The embedded photos are the bulk of the size.
• Text-only documents: barely change, because there's no image data to touch.
• Already-compressed images: if the source JPEGs were already saved at a low quality, re-encoding them again has little left to gain. You're compressing an already-compressed file.

This site's Compress PDF tool follows exactly this approach: it finds embedded JPEG images and re-encodes them at a lower quality, leaving text, fonts, and vector content untouched, which is why the result you get tracks directly with how image-heavy your original file was.

Related

Why is my PDF so big? covers the other side of this, what actually causes a large file size in the first place.