Screenshotting a PDF page gives you a re-rendered, re-compressed copy of an image that was already sitting inside the file at full resolution. Here's what's actually happening, and how to get the original.
The thing nobody tells you about PDFs
A PDF is not a picture of a document. It's a container. When someone places a photo into a page, the original image file — often several megabytes, often thousands of pixels wide — gets stored inside that container more or less intact. What you see on screen is that stored image, rendered down to fit your display.
This matters because of what happens next. When you need that photo back, the common advice is to screenshot the page or use "export as image". Both of those capture the rendered version: the image after it's been scaled to page size, rasterised at screen resolution, and compressed a second time on the way out.
The short version: a 4000px product photo inside a PDF becomes an 800px screenshot with visible compression artefacts. The original never left the file — you just took a photograph of it instead of opening it.
Why this bites hardest on catalogues and brochures
The situation where this hurts most is also the most common one: a supplier sends a product catalogue as a PDF, and you need the product images for a store listing. Screenshot them and every image arrives at roughly web-thumbnail quality — too small for a product page, far too small for a zoom view, and already compressed once so any further processing makes it worse.
The same applies to brochures you need to repurpose, spec sheets with technical diagrams, and scanned archives where each page is one large image captured at 300 DPI or better.
What extraction actually does
Extraction reads the PDF's internal structure, finds the embedded image objects, and writes them out as files. No re-rendering. No second compression pass. The bytes that come out are the bytes that went in.
Practically, that means:
- A photo embedded at 4000×3000 comes out at 4000×3000, not at whatever size it appeared on the page.
- Transparency is preserved where the source had it, instead of being flattened onto a white rectangle.
- A scanned page comes out at scan resolution, which is usually far higher than screen resolution.
The parts that get messy
Real PDFs are less tidy than the theory. A few things worth knowing before you start:
Images split into tiles
Some export tools slice a single large image into a grid of smaller ones. Extracted naively, you get twelve strips instead of one photo. Any tool worth using should recognise and reassemble those.
Repeated assets
The logo in the header appears on all 60 pages. So do the rules, the background textures and the page furniture. Extract everything and you'll get several hundred files, most of which you don't want. A minimum-size filter and duplicate detection cut that down to the images you were actually after.
Vector graphics aren't images
Charts and logos are often vector drawings rather than embedded pictures. There's no image file in there to extract — it has to be converted, and that is a render. Worth knowing so you're not hunting for a file that was never there.
Doing it in bulk
One PDF is a small job. A folder of 200 supplier catalogues is a different problem, and it's the normal one. At that scale you want the extraction to run across the whole directory, keep output organised by source file and page, apply the same filters everywhere, and not require you to sit there clicking.
You also want it running locally. Client catalogues, contracts and unreleased product data shouldn't be uploaded to a free web tool to save five minutes — that's a confidentiality problem dressed up as a convenience.
A quick checklist
- Extract, don't screenshot — the original is already in the file.
- Filter by minimum size so you don't drown in logos and page furniture.
- Watch for tiled images that need reassembling.
- Accept that vector graphics need conversion, not extraction.
- Keep sensitive documents off web upload tools.