PDF Compression, Honestly: What "Compress PDF" Really Does
"Compress PDF" isn't one operation — it's at least two different ones, with different trade-offs. Here's what's actually happening to your file, and why it sometimes doesn't shrink at all.
· 5 min read
What makes a PDF big
A PDF's file size is mostly a function of what's embedded inside it. Plain text takes up almost no space — a page of body copy is a tiny amount of data. The weight comes from three things: embedded images (photos, screenshots, logos), embedded fonts (a font file bundled into the document so it renders the same everywhere), and scanned pages, which are really just a special case of embedded images — a scanner doesn't produce text, it produces one full-page photograph per page.
A text-only contract is usually small, because it's mostly instructions for drawing letters. The same document scanned as high-resolution page images can be dramatically larger, because every page is stored as a full photo rather than as text — same page count, completely different weight, because one of them is text and the other is a stack of images wearing a PDF wrapper.
Two ways to compress a PDF
Because the weight lives in the images, both real compression approaches work on images — they just do it at different points, with a different consequence for the text.
The first approach recompresses the images that are already embedded in the PDF — downsampling their resolution and re-encoding them at a lower quality — while leaving everything else in the document untouched. Text stays text, exactly as it was, because the tool never touches it. Only the pictures get smaller.
The second approach rasterizes the whole page: it renders every page — text, images, everything — onto a canvas as one flat picture, then re-encodes that picture at a chosen quality. This is much more effective on scanned documents, because a scan is already one big image per page and there's no separate text layer to preserve in the first place. It's a blunter tool, but on the right input it does far more work.
Why compression can remove selectable text
If a PDF you compressed used to let you select, copy, or search its text, and now it doesn't, the tool almost certainly used the rasterize approach. Once a page has been flattened into a single image, every letter on it is a pattern of pixels, not a character — there's no text layer left to select, because there's no text left, only a picture of text.
That's not a bug, it's the mechanism doing its job — the same flattening that makes a dense scanned document shrink dramatically is what erases the text layer on any page it touches. It's a real trade-off, not a side effect to be surprised by: you're choosing smaller file size in exchange for losing selectable, searchable text on that page.
When a PDF won't shrink (or gets bigger)
Compression only has room to work where there's redundancy to remove. A PDF that's already small, or one that's mostly text with few or no images, doesn't have much for a compressor to squeeze — re-encoding it can leave the size about the same, and sometimes slightly larger, since the re-encoding process has its own overhead. That's expected behavior, not a broken tool.
LK Forge's Compress PDF tool uses the rasterize approach, with a quality control you can adjust — it's built for large, image-heavy or scanned PDFs, where it can do real work. It also runs entirely in the browser with no upload, no OCR, and no server round trip; it just renders and re-encodes the pages you give it.
The simplest safeguard is to compare: check the output file size against the original before you replace it. If compression didn't help — or made the file bigger — there's nothing wrong with just keeping the original.
FAQ
Why didn’t my PDF get smaller?
Compression shrinks the parts of a PDF that have room to shrink — mainly embedded images. If your PDF is already small, or is mostly text with few or no images, there isn’t much redundant image data to squeeze out. Running it through a compressor can leave the file roughly the same size, or even slightly larger, because re-encoding adds its own overhead. In that case the right move is to keep the original file.
Does compressing a PDF reduce quality?
It can. Compression works by re-encoding images at a lower quality setting, which is a trade-off, not a free lunch — you’re exchanging some visual fidelity for a smaller file. At moderate settings the difference is usually hard to notice; pushed too far, images can look soft or show compression artifacts. Text rendered from a PDF’s font data (not from an image) is unaffected either way — it stays sharp because it isn’t re-encoded as a picture.
Why is my text no longer selectable after compressing?
This happens when a compressor rasterizes the page — turning the whole page, text included, into a single image — rather than only recompressing the images already embedded in it. A rasterized page looks like the original, but every character is now a pixel, not text, so you can no longer select, copy, or search it. It’s the trade-off that makes this approach effective on scanned documents in the first place: a scan is already just a full-page image with no real text layer to preserve.
How do I compress a PDF for free?
LK Forge’s Compress PDF tool does it in the browser, for free, with no upload — pick a file, adjust the quality control, and download the result. It uses the rasterize approach, so it’s a strong fit for large scanned or image-heavy PDFs; for an already-small or text-only file, check the output size against the original and keep whichever is actually smaller.
Try It on Your File
Free, no upload, no account — adjust the quality and compare the result against your original.