Where a PDF's bytes actually are
A PDF can contain two very different kinds of content. Text and vector graphics are described as instructions — a font reference and a glyph path — and they cost almost nothing no matter how many pages you add. Images are stored as pixel data, often embedded at far higher resolution than the page will ever display, and they dominate the file. A 200-page text report can be 800 KB while a 20-page photo booklet can be 40 MB.
Scanned documents are the extreme case: every page is an image of a piece of paper, so a hundred-page scan is a hundred photographs. That is why the practical advice for scans is almost always about image resolution and format rather than about the PDF container itself.
A practical reduction sequence
- Establish the baseline: open the file and note the page count and byte size, so you can tell whether a change actually helped.
- If the PDF is a build from images you still hold, go back to the source images. Compress those with Compress Image and rebuild with Image to PDF. This almost always beats trying to shrink the finished PDF.
- If you only have the PDF, convert the pages that carry images to raster with PDF to Images at a moderate resolution, compress those images, and rebuild the document from them. Do this only for pages that are genuinely images — text pages lose their selectable text this way.
- Remove what you do not need. Split PDF or Extract PDF Pages lets you drop the appendix, the blank page or the duplicate scan before the file is sent anywhere.
- Re-check: open the result, read a paragraph at 100% zoom, and confirm the byte size is inside the target.
Text quality is where shortcuts hurt most. Never rasterise a text-only document to make it smaller — you will trade a near-zero cost for an enormous one: unsearchable, unselectable, blurry text.
How much detail is enough
Resolution for a page image is governed by how large it will be printed, described in DPI. The usual landmarks: 96–120 DPI is adequate for on-screen reading, 150 DPI is the floor for a legible photocopy-style print, and 300 DPI is the normal quality target for print. Scanning at 600 DPI because “higher is better” quadruples the pixels versus 300 with no visible gain on a standard page.
A4 is 210 × 297 mm, which is about 8.27 × 11.69 inches. At 300 DPI that is roughly 2481 × 3507 pixels — about 8.7 megapixels per page, and comfortably enough for print. At 150 DPI the same page is about 1241 × 1754, roughly a quarter of the data, still perfectly readable on paper for ordinary documents. Doubling DPI quadruples the pixel count, which is exactly why scan resolution is the main lever for scanned-PDF size.
What not to do
- Do not re-compress a PDF repeatedly. Each pass through an image pipeline adds generation loss.
- Do not strip a PDF of its fonts or structure expecting a large saving — they are a tiny fraction of a typical file.
- Do not lower quality so far that signatures, stamps or fine print become ambiguous. A document you cannot read is not a smaller document; it is a failed one.
- Do not assume an online “compressor” that asks you to upload is equivalent — the tools here run in your browser precisely so the content never leaves the device.
Common questions
Why is my PDF so much bigger than the Word file it came from?
Because the PDF almost certainly embedded fonts at full size and turned any images into pixel data. The Word file references fonts installed on your machine; the PDF must carry them.
Does merging PDFs make them larger?
Roughly by the sum of the parts, minus duplicated shared objects. If the combined file is too large, the right fix is reducing the biggest contributor first, not the merge step.
Is the text still selectable after compression?
It depends entirely on the method. Anything that keeps the original PDF objects preserves selectable text; anything that rasterises pages into images does not. Choose deliberately.