You need the photo from a PDF, not a screenshot of the page it's on. Maybe it's a product shot from a brochure, a picture from an old newsletter, or a diagram someone sent you without the original file.
There are two different jobs hiding in that request. Extract images pulls out the image files stored inside the PDF, pixel for pixel. PDF to JPG takes a picture of each whole page. Which one you want depends on what you're after, and the difference comes down to how PDFs store pictures.
Pictures in a PDF are files inside the file
When a PDF contains a photo, the photo is stored as its own object with its own pixel dimensions. The page then says "draw that image here, at this size". The image keeps all of its pixels no matter how small it's drawn. A 4000 by 3000 pixel photo shown as a two-inch thumbnail still carries 4000 by 3000 pixels inside the file, unless whoever made the PDF shrank it.
Extract images goes through each page, finds those image objects and saves them. It doesn't look at the page as you see it at all. An image used on many pages, like a logo in every header, is saved once and named after the first page it appears on.
Most photos inside PDFs are stored as JPEG. Extract images saves those as .jpg files byte for byte, with no re-compression, so you get exactly the file that was put in. Everything else (images stored without loss, CMYK photos from print workflows, images with a transparency mask) is decoded and saved as a lossless PNG. A logo with a transparent background comes out as a PNG with the transparency kept.
Extract images or PDF to JPG
Use Extract images when you want a photo or picture that's in the PDF. You get its original resolution, which is often far more than what you'd get by rendering the page.
Use PDF to JPG when you want what the page looks like: text, layout, everything. It renders each page at the resolution you choose, from "Small (75 dpi)" to "Print (300 dpi)". The guide on converting PDF pages to JPG covers those settings.
The catch with PDF to JPG for photos is that the page resolution limits everything on it. A photo printed 3 inches wide on a page rendered at 150 dpi comes out about 450 pixels across, however many pixels it had in the file.
The catch with Extract images is that it only finds raster images. Charts, diagrams, logos and maps are often drawn as vector shapes, which are lines and curves, not pixels. They aren't images in the PDF's sense, so Extract images won't find them. For those, render the page with PDF to JPG at "Print (300 dpi)" and crop the picture out in any image editor.
One case where Extract images wins outright: scanned documents. Each page of a scan is one big image, so extracting gives you the scan exactly as the scanner saved it, without a second round of rendering and compression.
Extracting images step by step
- Open Extract images and add the PDF. If it has a password, you'll be asked for it.
- To scan only part of the document, type pages into the "Pages" box, such as 1-3, 8. Leave it empty for all pages.
- Leave "Skip tiny images" ticked unless you need icons. It leaves out anything smaller than 20 pixels on either side, which is mostly bullets, rules and bits of decoration.
- Click "Extract images". Each image found is shown with its page number.
- Click the x on any image you don't want ("Leave out of the download"), or "Save" to download just one. Undo and redo work if you remove the wrong one.
- Click "Download 6 image(s)" (the number matches what's left). Several images arrive as one ZIP file.
Files are named after the PDF, the page and the order on that page, like brochure-p3-1.jpg and brochure-p3-2.png. Everything happens in your browser, so the PDF isn't uploaded.
Why an image doesn't look like it did on the page
The extracted file is what's stored, not what's shown. Those can differ more than you'd expect:
- The page may show only part of an image, cropped by a clipping path. You get the whole image, including the parts that were hidden.
- The page may draw an image rotated or flipped. You get it the way it's stored, which can be sideways.
- Some PDF makers cut one picture into strips or tiles. You get the pieces, not the picture.
- Text laid over a photo on the page isn't part of the image, so it won't be there.
- Colours in a CMYK image can look a little different on screen, because they had to be converted to RGB for the PNG.
If the result says "No embedded images found", the PDF has no raster images on those pages. Either it's all text and vector graphics, or the images are stored in a way the tool doesn't look for. A few PDF makers put small images straight into a page's drawing instructions instead of storing them as separate objects, and Extract images doesn't pick those up. Very large non-JPEG images, over about 40 megapixels, are scaled down to stay within what the browser can handle.
Original quality means the quality in the PDF
Extract images never adds detail that isn't there. If the PDF went through a compressor, whether Compress PDF or the "reduce file size" option in some app, the photos inside were already resampled and you get those smaller versions. The original 12-megapixel photo isn't hiding in the file somewhere. If you need it, ask whoever made the PDF for the source image.
And the usual caveat: a photo in someone else's brochure still belongs to them. Extracting it is easy. Using it is a separate question.
If you're pulling images out to put them into a new document, JPG to PDF takes the JPG and PNG files straight back in. Check the sizes first, though. A folder of extracted images often includes a few large backgrounds you didn't notice on the page.
Also useful: PDF to JPG · JPG to PDF
