ImageTools

DPI, PPI and print size, explained properly

DPI is the most misunderstood number in imaging. Changing it does not change your image. It is a note in the file header telling a printer how large to print each pixel.

The thing almost everyone gets wrong

An image has a fixed number of pixels — say 3000 wide. That is all the detail it will ever have. The DPI value stored in the file does not add or remove a single one of them; it only says how densely to lay them down on paper.

In all three cases the file is identical. Nothing was resampled. When a print shop asks for "300 DPI", what they mean is "enough pixels that it works out to 300 per inch at the size I am printing" — which is a statement about pixel count, not about the tag.

On a screen, DPI is ignored entirely. A web browser cares only about pixels.

The arithmetic

Print size in inches = pixels ÷ DPI, and therefore pixels needed = inches × DPI.

Print sizeAt 300 DPIAt 150 DPI
4×6 inch1200×1800600×900
5×7 inch1500×2100750×1050
8×10 inch2400×30001200×1500
A42480×35081240×1754
A33508×49611754×2480
16×20 inch poster4800×60002400×3000

How much do you actually need?

300 DPI is the traditional standard for things held at reading distance: photo prints, magazines, brochures, business cards. It is genuinely the point at which most people stop seeing pixels at about 30 cm.

But viewing distance changes everything. A billboard is printed at 10–20 DPI and looks perfect from across the street. Sensible working figures:

A 12-megapixel phone photo is about 4000×3000 pixels, which is a genuine 13×10 inch print at 300 DPI, or a very respectable 26×20 at 150. Most people have far more resolution than they think.

DPI vs PPI

Strictly, PPI (pixels per inch) describes an image or a screen, and DPI (dots per inch) describes a printer laying down ink dots — and a printer uses several dots to make one pixel, so a 1200 DPI printer is not producing 1200 PPI of image detail. In practice the two words are used interchangeably everywhere including by print shops, and the tag inside a JPEG or PNG is universally called DPI. Nothing goes wrong if you treat them as the same thing.

When you do need to set the tag

Some print services and layout software read the embedded value and use it to place the image at a default size, or reject the file outright if it says 72. Setting it to 300 satisfies them without touching a pixel. That is exactly what the DPI converter does: it rewrites the density field in the file header and copies the image data through untouched.

Common values: 72, 96, 150, 300, 600.

If you genuinely do not have enough pixels

Then you need more pixels, not a higher number in the header. Resizing up interpolates: it invents the new pixels by averaging neighbours, and the result is soft. An AI upscaler does better, because it predicts plausible detail rather than smearing what is there — but it is still inventing, and it cannot recover something that was never recorded.

The real answer is to shoot or scan at a higher resolution in the first place. If you are digitising prints or film, scan at 600 DPI or more; storage is cheap and rescanning is not.