Does Compressing a PDF Blur the Text?

Compressing a PDF blurs text only when the text is part of a picture. A PDF made by exporting from Word, Google Docs or a browser stores letters as font data, and a compressor that works on images never touches them, so those pages stay sharp and the text stays selectable. A PDF made by scanning paper stores each page as one big photo, and that photo is exactly what gets squeezed. So the real question is not whether compression hurts quality, it is whether your pages are text or pictures. You can find out in ten seconds by trying to drag-select a word in your PDF reader. If the word highlights, it is real text. If the whole page highlights as a block, it is a picture.

The Compress PDF tool does two things and nothing else. First, it finds embedded JPEG images (the format scanners and phone cameras produce) and re-saves each one at a lower JPEG quality, and scales it down if it is larger than a size cap. Second, it repacks the file's internal structure so it saves a little more space. It does not rasterise pages, so typed text, links and bookmarks are left as they were. It also skips CMYK JPEGs, because a browser cannot decode them reliably and a wrong guess would shift colours, and it keeps a re-saved image only if the new one is at least 10 percent smaller than the old one. A PDF with no JPEG images in it therefore shrinks very little, which is an honest result, not a failure.

The cap is on the longest side of each image, in pixels: 1200 for Low, 1600 for Medium and 2000 for High, with JPEG quality 40, 60 and 75 percent. Spread over a full US Letter page, which is 11 inches tall, that works out to roughly 109, 145 and 182 pixels per inch. A scanner at its usual 300 dpi makes a Letter page about 3300 pixels tall, so every level throws away most of the resolution. Small print on Low is the first thing to go soft, and a 109 ppi page is readable on screen but looks rough when printed. For a scanned document you mean to print, start with High, and only go lower if the file is still too big to send.

Open the compressed file, zoom to 100 percent, and read the smallest text on the busiest page, such as a footnote, a table or a stamp. If you can read it without squinting, the file is fine. If you cannot, go back to the original and compress it again at the next quality level up. Always start again from the original, never from the compressed copy, because every JPEG re-save adds a little loss on top of the last one. If the tool reports that it found no compressible embedded photos, the file was text or vector only, so nothing was squeezed and nothing was lost.

If the PDF had a real text layer before, it still has one afterwards, because the compressor only edits images. A scan with an invisible text layer from OCR keeps that layer too. The case that surprises people is the opposite one: a plain scan has no text layer at all, so Ctrl+F finds nothing before or after compression, and the compressor does not add one. Run it through OCR PDF first if you need to search or copy from it, then compress. Doing it in that order matters, since OCR reads small letters more accurately from the original image than from a heavily squeezed copy.

Sometimes the PDF was already soft before you compressed it, for example a photo of a page taken at an angle, or a scan done at 100 dpi. Compression cannot sharpen that and will not make it worse by much. A second cause is a viewer that shows a low-resolution preview while a large page loads, which looks like blur until it finishes drawing. If a file is still too large after the highest quality level, the images are probably not JPEGs, and the guide on PDFs that stay huge after compression covers the fixes.