← All posts

Why is my JPG-to-PDF file so much bigger than the photos?

Twelve photos. About 400 KB each, so five megabytes of pictures. You put them through a converter and got back a PDF of 30 megabytes, containing nothing but those twelve photos.

Nothing was added. There is no text in it, no fonts to speak of, no second copy of anything. And yet the document weighs six times what went into it, and running it through the same converter again with different settings does not change the number.

The photos did not grow. They were unpacked — and once you know which of four things unpacked them, you know whether there is anything to fix.

A JPEG is a recipe, not a picture

This is the whole mechanism, and every page that ranks for your question leaves it out.

A photograph as your camera stores it is not a grid of pixels. It is a compressed description of one, and the compression is brutal: JPEG throws away detail your eye does not miss and then encodes what is left. A 400 KB JPEG is typically holding around two megapixels. Two megapixels as actual pixels — three bytes each, red, green and blue — is six megabytes. That is what your photo really is. The 400 KB is the recipe for rebuilding it.

A PDF can store that picture either way, and both are perfectly legal:

The second one is where your 30 MB came from, because zip is nearly useless on a photograph. Zip works by spotting repetition, and in a photograph of a garden or a face, no two neighbouring pixels are the same. It claws back perhaps a quarter. So six megabytes of pixels goes into the PDF at somewhere between two and four megabytes, and your 400 KB photo has arrived weighing five to ten times what it did on disk. Twelve of them is thirty megabytes. That is your file, exactly.

Nobody chose this to annoy you. It is the lazy path through an image library: decode the picture because that is what a decoder does, then draw it. Print-to-PDF drivers, "export as PDF" in photo viewers and word processors, and a lot of quick scripts all take it.

The other three ways the number gets away from you

The converter saved the sheet, not the photo. Anything that goes through a print driver renders an A4 page at 300 or 600 dpi and stores that as the image. A4 at 300 dpi is 8.7 megapixels of page, most of it white margin — and white margin stored as zipped pixels is not free either. The tell is that every page is exactly A4 and the file size barely varies with how big your photos were.

The photos were HEIC. Everything shot on a recent iPhone is HEIC, which is roughly half the size of a JPEG of the same picture at the same quality. Almost nothing that makes PDFs reads HEIC, so something in the chain converted it first, usually to a maximum-quality JPEG — two to four times the HEIC before the PDF had even started. Ours does not accept HEIC either. Set the iPhone camera to "Most Compatible", or share a copy out of Photos, which hands over a JPEG.

The 400 KB you compared against was not the file. If that number came off the phone and iCloud "Optimise iPhone Storage" is on, the copy living on the device is a reduced one. Sharing or exporting fetches the full-size original from iCloud, which is several megabytes. The converter never saw the small version.

Two minutes of diagnosis

  1. Divide. PDF size in bytes, over the number of photos. Compare it with what one photo weighs on disk — not in a gallery app. Roughly equal means the conversion was honest and your photos are simply that big. Five to fifteen times means the pixels were unpacked.
  2. Look at the pages. Our PDF inspector reads the file in your browser without uploading it. If every page is exactly A4 or Letter, a print driver made this. If the pages are odd sizes matching the shapes of your photos, the photos were placed rather than printed. The Producer field usually names the program that did it, which is the fastest way to work out what to stop using.
  3. Ignore "no text layer." A document made of photographs has none, and that is correct, not a fault. It only matters if the photographs are of documents you expected to be searchable — a different problem entirely.

What our converter does, and why it offers you no size setting

JPG to PDF takes the other path. A file that is already a JPEG is written into the document byte for byte — never decoded, never re-encoded. The result is the sum of the photos you gave it, plus a few kilobytes of structure per page, and the notes it hands back tell you how many JPEGs went in untouched.

So the file it produces is as small as your photographs and no smaller. There is deliberately no quality slider on it, because a converter that quietly re-encodes your photographs is losing detail at a moment when you did not ask it to.

Which also means: the page settings do not change the file size. A4 or Letter or fit-the-picture, zero margin or fifteen millimetres, portrait or landscape — those decide geometry, not bytes. The same photographs come out to within a rounding error of the same number every time. Do not spend twenty minutes re-running the conversion.

Two cases where our tool does unpack a picture, for honesty's sake. Anything that is not already a JPEG is decoded and re-stored losslessly, which is right for screenshots and logos but means a PNG of a photograph stays as large as PNGs of photographs are — several times the JPEG would be. And a CMYK JPEG, the kind that comes out of print-oriented exports, cannot be handed over whole, so it gets decoded too. If one page is inexplicably heavy, that is usually why.

Making it smaller afterwards

The size reduction is a separate, deliberate step, and this is where you decide what to give up.

  1. Run the PDF through Compress PDF at the default: balanced, 150 dpi.
  2. Read the notes rather than only the new number. They say how many pictures were re-saved and what they weighed before and after.
  3. Open the heaviest page at full size before you keep the result.

The reason this works so well on a PDF of photographs is arithmetic. Compress measures how large each picture is actually drawn and turns that into a resolution. On a fitted page the longest edge is capped at A4's, so a 4000-pixel photo is being drawn across about 11.7 inches — 340 dpi. On an A4 page with a margin it is closer to 500. A screen shows you 90 and an office printer uses 300, so at a 150 dpi ceiling three quarters of those pixels were never going to be visible. Expect a large drop, and expect it to be invisible on screen.

If it comes back saying the pictures were already at or below the target, believe it: your photos were modest to begin with and the document is honest. That is the same answer as a PDF that will not compress, and it is not a failure of the button.

Better still: shrink the photos before you convert

Order of operations matters more than people expect. Converting first and compressing afterwards means your photograph is decoded from a JPEG and re-encoded as a JPEG — two generations of lossy encoding for one result. Resizing the originals and then converting is one. If the photos matter, do it in that order, and these are all free:

Keep the originals

Everything above except img2pdf throws detail away, and none of it is reversible. Keep the photographs on disk until you are certain the PDF is good enough for whatever it is for — and if you later pull the pictures back out of the compressed document and they look soft, that is the compression you asked for, though it is not the only cause.

The one thing no tool can do is make a PDF of photographs smaller than the photographs. It is only ever a question of which pixels you are willing to lose.

Tools this post uses: