Make a file smaller

In a PDF it is almost never the text. A page of words is a few kilobytes — glyph codes, positions, and a font subset shared across the document — while a single photograph on that page can be two megabytes on its own. So the PDF compressor re-encodes the photographs and leaves the text, the fonts and the vector drawings exactly as they were, which is why the result is still selectable, searchable and readable aloud. In a photograph the weight is the pixels; in an MP3 it is the bitrate. Different material, same question: is there anything here actually worth taking?

A JPEG and an MP3 have both been through a compressor already, and re-encoding either throws away a second helping of what survived the first pass. That is the difference between compressing a PDF, where the text comes through untouched, and compressing a photograph or a recording, where every pass costs something you cannot get back. It is also why the PDF compressor refuses to rasterise: turning every page into a picture is how most “compress PDF” tools get their impressive numbers, and it destroys selectable text, screen-reader access and print quality to get them. Worth doing when a file has to fit somewhere; not worth doing to a file you are keeping. Either way, keep the original.

If a PDF holds no images, or its images are already efficiently stored, or the saving comes to under one per cent, or under ten kilobytes, or the rebuilt file is larger than the original, you are offered nothing and told which of those happened. The image compressor hands your original back when re-encoding would only have made it worse. The MP3 compressor refuses to raise a bitrate at all, because encoding 96 kbps audio at 192 kbps produces a file twice the size containing exactly the same music. A true number can still be a lie about the outcome: “2% smaller — 2 KB became 2 KB” is arithmetic, not a result.