DPI, PPI and resolution explained
DPI is one of the most misunderstood numbers in imaging, and the misunderstanding costs people money at print shops. It is a setting about physical size, and it says nothing at all about quality.
Last reviewed
The one sentence version
An image file contains a grid of pixels and a note saying how densely to print them. DPI is that note. Changing it changes how big the print comes out and nothing else — not the pixels, not the detail, not the sharpness on screen.
Why that is so hard to believe
Because the word “resolution” is used for two different things, and people reasonably assume that a higher number means a better picture.
Pixel dimensions — 4000 × 3000 — are how much information the file holds. That is fixed when the photograph is taken and it is the number that determines detail.
DPI is a division instruction. 3000 pixels at 300 DPI is a ten-inch print; the same 3000 pixels at 150 DPI is a twenty-inch print. Same file, same information, different physical size — and the twenty-inch print looks softer, not because anything changed but because the same detail is spread over four times the area.
So DPI does affect how sharp a print looks. It does it by changing the size, not by changing the image.
DPI or PPI?
Strictly, PPI — pixels per inch — describes an image, and DPI — dots per inch — describes a printer, which lays down several tiny ink dots to reproduce one pixel. A good inkjet prints at 1440 DPI and is still reproducing a 300 PPI image.
In practice every camera, every editor, every form and every help page uses DPI for both, and the field in a JPEG’s header is literally called a density. Correcting somebody who has just rejected your file is not a winning strategy, so this site uses DPI too. The distinction matters when you are talking to a printer about their machine, and almost never otherwise.
The 72 DPI myth
“Web images should be 72 DPI” is advice that has been wrong for about thirty years and is still repeated constantly.
It comes from the original Macintosh, whose screen happened to display 72 pixels per inch, which made a point of type on screen the same size as a point of type on paper. That was a genuine convenience in 1984. It has not been true of any mainstream display since, and no browser has ever read the DPI field.
On the web an image is measured in pixels and displayed in pixels. A file marked 300 DPI and an identical file marked 72 DPI render identically, byte for byte the same picture. Setting it to 72 “for the web” does nothing whatsoever — it does not reduce the file size, it does not speed anything up, it changes two bytes in the header.
What a sharp print actually requires
Multiply the print size in inches by the DPI you want. That is the pixel count you need, and no header edit substitutes for it.
| At 300 DPI | At 150 DPI | |
|---|---|---|
| 6 × 4 inch photo | 1800 × 1200 | 900 × 600 |
| 7 × 5 inch photo | 2100 × 1500 | 1050 × 750 |
| 10 × 8 inch photo | 3000 × 2400 | 1500 × 1200 |
| A4 page | 3508 × 2480 | 1754 × 1240 |
A modern phone photograph is comfortably enough for anything up to A4. A screenshot, or an image saved from a website, usually is not — 800 pixels is a two-and-a-half inch print at 300 DPI, and enlarging it does not help.
Why 300, and when less is fine
300 DPI is roughly the point at which a person holding a photograph at normal reading distance stops being able to resolve individual dots. It is a convention rather than a law of physics, and it assumes something held in the hand.
Things viewed further away need far less. A poster on a wall is fine at 150 DPI; a large banner read from across a room is fine at 50. Print shops that ask for 300 regardless are applying a safe default, and for anything large it is worth asking rather than assuming — the difference between 150 and 300 is a factor of four in pixel count.
The upscaling trap
Give some DPI converters a small image and a request for 300 DPI, and they will enlarge the picture so the pixel count matches the physical size. The header then says 300 and the image is blurrier than it was, because the added pixels were averaged from the ones already there.
That is not a fix. It hides the shortfall inside a larger file and destroys the sharpness the original genuinely had. FileBelow’s DPI tool writes the header value and separately tells you the pixel count your intended print size requires — which, when there are not enough, is the real answer.
So when does setting DPI actually help?
When something is literally reading the field. Some journals, some government submission portals and some printing services validate the header and reject anything that does not say 300. In that case writing the number is the entire fix, it costs nothing, and your image did not need to change.
It also helps when handing a file to a designer or a print shop, as a statement of intent: “print this at ten inches wide”. That is what the field is for. It is a note about intended size, and it has never been a measure of quality.
Need to actually do this?
It writes the field and tells you how many pixels your print size genuinely needs. Free, no sign-up, and the image is processed on your own device.
Set an image’s DPIOr browse every explanation on this site, and the requirements published by named destinations.