Most people learn these three formats by accident — whatever the Save dialog offered, whatever the phone produced, whatever the website demanded — and end up treating them as interchangeable names for the same thing. They are not remotely the same thing, and the differences explain a surprising share of everyday document frustration: the screenshot that went fuzzy, the logo with a grey box behind it, the upload form that would not take the file, the 14 MB email attachment.
The single most useful thing to understand is that this is not a three-way comparison at all. JPG and PNG are image formats, and they differ in one specific way. PDF is a document container, and it belongs to a different category entirely.
The short answer
- JPG — photographs, and anything else where the content is continuous tone and small file size matters more than pixel-perfect accuracy.
- PNG — screenshots, logos, diagrams, anything with sharp edges or flat colour, and anything that needs a transparent background.
- PDF — anything that is a document: multiple pages, text meant to be read, something to be printed, signed, archived or submitted to somebody.
If you take nothing else away: the moment a thing is a document rather than a picture, the answer is PDF, and it does not matter how good the JPG looks on your screen.
What JPG actually does
JPG — also written JPEG, they are the same format — uses lossy compression. It analyses the image, decides which detail human vision is least likely to notice, and permanently discards it. The result is a file perhaps a tenth the size of the raw pixel data, at a cost you usually cannot see.
For photographs this trade is excellent. Photographic content is continuous tone — gradual shifts of colour and brightness, few hard edges — which is precisely what the algorithm was designed for. A well-saved JPG of a landscape is visually indistinguishable from the original at a fraction of the size, which is why every camera and phone on earth produces them.
For anything with hard edges the same algorithm fails visibly. Text, line drawings, screenshots and logos are full of abrupt transitions from one flat colour to another, which is the case JPG handles worst. The compression produces faint smears and blocky halos around the edges of letters — artifacts — which is why a screenshot saved as JPG looks slightly dirty next to the original.
JPG also has no transparency. Every JPG is a solid rectangle. Save a logo with a transparent background as JPG and the transparency becomes white — which is the origin of the white box that appears behind logos placed on a coloured page.
What PNG actually does
PNG uses lossless compression. It finds patterns in the pixel data and stores them more efficiently, but reconstructs the image exactly — every pixel comes back identical to the original. Nothing is discarded, ever, no matter how many times the file is opened and re-saved.
That makes PNG the correct choice for exactly the content JPG mangles. A screenshot saved as PNG is pixel-accurate: the text is as crisp in the file as it was on the screen. Diagrams, charts, logos, icons and anything with large areas of flat colour compress extremely well in PNG, often better than in JPG, because repeated identical pixels are what its algorithm is good at.
PNG also supports a genuine alpha channel — per-pixel transparency, including partial transparency for soft edges. This is why logos and icons are distributed as PNG: the same file drops onto a white page, a dark page or a photograph without a rectangular background following it around.
The cost is size, on the one kind of content PNG is bad at. Photographs have almost no repeated pixels, so there is little for lossless compression to exploit; a photograph saved as PNG is routinely five to ten times larger than the equivalent JPG at no visible benefit. Sending holiday photos as PNGs is the mirror image of saving a screenshot as JPG — the wrong format for the content.
Why PDF is not in the same category
JPG and PNG store a grid of pixels. That is all they store. A PDF is a container that describes a document: pages with defined dimensions, text as actual text with embedded fonts, vector graphics that stay sharp at any zoom, and images where images are needed — held together in one file with metadata, bookmarks and a page order.
This is why the comparison keeps confusing people. Asking whether to use JPG or PDF for a photograph is a reasonable question with a clear answer (JPG). Asking whether to use JPG or PDF for a signed contract is not really a format question at all — one is a document and the other is a snapshot of a document.
What the container buys you:
- Real pages. A PDF page has fixed dimensions — A4, Letter — so it prints predictably. An image is a rectangle of arbitrary pixel size that every application scales according to its own guesswork.
- Many pages in one file. A twelve-page report is one PDF. As images it is twelve files, twelve chances to lose one, and a problem for any form that accepts a single attachment.
- Selectable, searchable text. Text in a PDF can be selected, searched and copied, and read aloud by a screen reader. Text in an image is a picture of text and none of that works.
- Security and signatures. Passwords, encryption, permissions and digital signatures are part of the specification. Images support none of it.
- Archival guarantees. PDF/A is a formal standard for long-term preservation, which is why records systems and compliance regimes ask for it.
The subtlety worth knowing: a PDF can contain nothing but a photograph of a page, in which case it inherits the image's limitations — no selectable text, nothing searchable. That is what a scan is before OCR. The container is right; the contents still need a text layer, which is what OCR PDF adds.
Need a document rather than a picture?
Turn JPG, PNG, HEIC, WebP or TIFF images into a proper PDF with real pages, in the order you choose. Free, no sign-up, files deleted after 2 hours.
Open Image to PDF →Side by side
| JPG | PNG | ||
|---|---|---|---|
| Type | Lossy raster image | Lossless raster image | Document container |
| Quality on re-saving | Degrades every save | Never degrades | Unchanged (text stays text) |
| Photographs | Excellent | Works, but 5–10× larger | Fine for delivery, not for editing |
| Text and screenshots | Visible artifacts around edges | Pixel-perfect | Best — text stays real text |
| Transparency | None | Yes, including partial | Yes |
| Multiple pages | No | No | Yes |
| Searchable text | No | No | Yes (with OCR if scanned) |
| Password protection | No | No | Yes |
| Typical use | Photos, camera output | Screenshots, logos, diagrams | Forms, contracts, reports, anything submitted |
Choosing by what you are saving
Working backwards from the content is faster than reasoning about compression:
| What you have | Use | Why |
|---|---|---|
| A photograph | JPG | Continuous tone is exactly what lossy compression handles well |
| A screenshot | PNG | Sharp text edges are exactly what JPG handles badly |
| A logo or icon | PNG | Needs transparency and hard edges |
| A chart or diagram | PNG | Flat colour and thin lines; JPG blurs both |
| A signed form | It is a document — fixed pages, printable, submittable | |
| A multi-page anything | Images cannot hold pages | |
| Something to be printed | Defined page size means it prints as intended | |
| Something to be archived | PDF (ideally PDF/A) | The only one of the three with an archival standard |
| Something being uploaded to a portal | It is what almost every form asks for |
HEIC, WebP and TIFF
Three other formats turn up often enough to be worth a line each.
HEIC
The format iPhones save photos in by default. It is genuinely better than JPG — comparable quality at roughly half the size — but support outside Apple's ecosystem is still patchy, which is why an emailed HEIC regularly arrives as a file the recipient's computer refuses to open. If you need to send photographs to someone whose setup you do not know, convert to JPG first, or change the capture setting under Settings → Camera → Formats → Most Compatible.
WebP
Google's web image format, supporting both lossy and lossless modes plus transparency, typically 25–35% smaller than the equivalent JPG or PNG. Every current browser handles it. Desktop software and upload forms often do not, which is why images saved from a website frequently arrive in a format nothing local will open.
TIFF
The old professional scanning and print format — lossless, high fidelity, supports multiple pages, and very large. Still standard in printing, medical imaging and some archival systems. If you have been handed a multi-page TIFF and need something ordinary people can open, convert it to PDF.
Converting between them
Conversion is straightforward in every direction, but two rules govern whether the result is any good.
Converting up is safe; converting down is not. PNG to JPG discards detail permanently. JPG back to PNG does not restore it — you get a lossless copy of an already-degraded image, at a larger file size. Once detail is gone it is gone, so keep the original in the better format and export copies from it rather than converting a converted file.
Images to PDF is the common direction, and the one worth doing deliberately. Image to PDF accepts JPG, PNG, HEIC, WebP and TIFF, takes several files in one pass, and lets you set the page order, page size and orientation before converting. Those settings are the difference between a usable document and a small picture marooned in white space — the details are covered in how to convert an image to PDF.
Going the other way, PDF to Image renders each page as a picture, which is what you want for a thumbnail or a slide. Be clear about what you lose: the text stops being text, the file stops being searchable, and one page becomes one image. It is the right move for embedding a page into a presentation and the wrong move for anything that still needs to function as a document.
Two practical follow-ups. If the images made your PDF unreasonably large, Compress PDF re-encodes them without disturbing the page structure. And if your PDF is made of images that need to be readable as text — a scan, or photographs of pages — OCR PDF adds the invisible text layer that makes the document searchable.
Frequently asked questions
PDF, without qualification. A JPG of a document is a photograph of it — fixed to one page, with no selectable text, no defined print size, no password protection and no signature support. A PDF is a document: it has real pages that print predictably, it can hold many pages in one file, its text can be searched and copied, and it is what virtually every upload form and institution actually asks for. Use JPG for photographs, PDF for anything that is a document.
JPG is lossy — it permanently discards detail to make the file smaller, and discards a little more every time the file is re-saved. PNG is lossless — it reproduces every pixel exactly, forever, and supports transparency. The practical consequence is that JPG is excellent for photographs and poor for text and sharp edges, while PNG is excellent for screenshots, logos and diagrams and unnecessarily large for photographs.
It was almost certainly saved as JPG. Screenshots are full of hard edges between flat colours — exactly the content JPG's compression handles worst — so it produces faint smears and blocky halos around the letters. Save screenshots as PNG instead, which is lossless and reproduces the text exactly as it appeared on screen. If the screenshot has already been through several JPG saves the degradation is permanent and cannot be reversed.
Not by itself. The image is embedded in the PDF at its original resolution and nothing is re-compressed. Quality loss can happen at two other points: if you compress the PDF afterwards to reduce its file size, and if the conversion scales the image to fit a page smaller than its natural dimensions. Choosing a page size and orientation that match your image's proportions avoids the second problem entirely.
No. Once JPG compression has discarded detail it is genuinely gone, and converting to a lossless format afterwards only preserves the damage at a larger file size. The rule is to keep your original in the higher-quality format and export copies from it, rather than converting an already-converted file. If all you have is the JPG, that is the best version that exists.
Send JPG unless you know the recipient uses Apple devices. HEIC is technically superior — similar quality at about half the size — but support outside Apple's ecosystem is still inconsistent, and an emailed HEIC frequently arrives as something the recipient's computer will not open. You can change the capture format under Settings → Camera → Formats → Most Compatible, or convert the images when you send them.
Turn your images into a proper document
JPG, PNG, HEIC, WebP and TIFF into a single PDF with real pages, the right size and the order you choose. Free, no sign-up, files deleted after 2 hours.
Open Image to PDF →