DPI, PPI, and Print Size: What the Numbers Actually Mean

Three numbers, one equation

Everything about this topic follows from a single relationship: pixels equals inches multiplied by DPI. Three quantities, one equation. Fix any two and the third is determined.

An image that is 3000 pixels wide can be printed at 10 inches wide at 300 DPI, or at 30 inches wide at 100 DPI, or at 6 inches wide at 500 DPI. The file is identical in all three cases. Nothing about the image changes — only the decision about how large to print it, which is made at printing time.

The DPI value stored in a file's metadata is simply a recorded preference about which of those choices was intended. It is a sticky note attached to the image saying 'the person who made this had a 10-inch print in mind'. A printer can honour that note or ignore it, and you can override it at any time without altering a single pixel.

Strictly speaking, DPI and PPI are different

PPI means pixels per inch and describes a digital image or a display. DPI means dots per inch and describes a printing device — how many ink dots the hardware places in an inch of paper. In everyday use the two terms are swapped freely, and for the arithmetic above it rarely matters.

Where the distinction becomes real is at the printer. Many printing processes cannot vary the intensity of an individual dot, so they simulate shades by clustering dots in patterns. Such a printer may need several ink dots to represent one image pixel convincingly, which is why a device advertising 1440 DPI is not asking you to supply a 1440 PPI image. Its dot count and your pixel count are measuring different things.

The practical consequence: ignore the printer's advertised DPI when preparing files. What you need is the image resolution the print shop asks for, which is a PPI figure even when they call it DPI.

Where the 300 DPI rule comes from

The familiar 300 figure is a convention for material held at reading distance — roughly 25 to 30 centimetres. At that distance, a person with normal vision stops being able to resolve individual dots somewhere around 300 per inch, so printing finer detail adds cost without adding anything anyone can see.

It is a rule of thumb, not a physical constant, and it is tied to that viewing distance. This is why it applies to books, magazines, brochures, and photographic prints, and why it does not apply to a poster on a wall or a banner across a building.

Viewing distance changes the requirement dramatically. A poster viewed from two metres is comfortable at 150 PPI. A large-format banner viewed from ten metres may be printed at 30 PPI or less and look perfectly sharp in place. Printing a billboard at 300 PPI would require a file of absurd size to represent detail no viewer could ever be close enough to see.

So the honest version of the rule is: ask the printer what resolution they want for this specific job at this specific size. They know their process and the viewing distance. A number from a general article is a fallback for when nobody is available to ask.

Working out what you actually need

Start from the physical size and multiply. For a 10 by 8 inch print at 300 PPI: 10 × 300 = 3000 and 8 × 300 = 2400, so you need 3000×2400 pixels. For an A4 page, which is 8.27 by 11.69 inches, at 300 PPI you need roughly 2480×3508 pixels. For a 24 by 36 inch poster at 150 PPI, 3600×5400 pixels.

Then compare against what you have. If your image meets or exceeds those numbers, you are finished — a file with more pixels than needed can simply be printed at a higher effective resolution, which harms nothing.

If it falls short, you have a genuine problem and no software fix. You can print smaller, accept a lower resolution if the viewing distance permits, or obtain a better source. Enlarging the file does not help, for reasons worth being precise about.

Why upscaling does not solve a shortfall

Enlarging an image adds pixels by interpolating between the ones that exist — calculating plausible intermediate values from neighbours. The new pixels contain no information about the scene that was not already present. The image becomes larger and correspondingly softer, because the same amount of real detail is now spread across more pixels.

Changing only the DPI tag, without resampling, is even less useful: nothing at all happens to the image. Some people do this because a submission system checks the metadata field, and it may satisfy that check, but the printed result is identical to what it would have been anyway.

Modern upscaling tools driven by machine learning produce visibly better results than simple interpolation, and for some purposes they are genuinely useful. They are still inventing plausible detail rather than recovering real detail. For a decorative enlargement that is often acceptable. For reproduction of a document, artwork, or anything where accuracy is the point, it is not — the invented detail is fiction, and it can be confidently wrong.

Screens do not work this way at all

On a screen, an image's DPI metadata is ignored entirely. A browser lays out an image according to its pixel dimensions and whatever CSS governs it. Two files with identical pixels and different DPI tags render identically, byte for byte in the output.

What does matter on screen is device pixel ratio. High-density displays pack more physical pixels into the same area, so an image intended to occupy 400 layout pixels needs roughly 800 image pixels to look sharp on a 2× display. This is the screen equivalent of the print resolution question, and it is expressed through markup rather than through a metadata tag.

This is also why 'save for web at 72 DPI' is advice from a much older era. The number came from the nominal resolution of early displays and has no bearing on anything today. Exporting at 72, 300, or any other value produces the same file for web purposes, provided the pixel dimensions are the same.

Common situations and what to do

A print shop says your file is too low resolution. Compute what they need with the multiplication above and compare honestly. If you are short, print smaller or find a better source; do not upscale and resubmit, because they will usually notice and it will look soft in any case.

A form demands 300 DPI. Check whether it is reading the metadata field or checking pixel dimensions against a physical size. If the latter, supply an image with enough pixels. If the former, setting the tag will pass — but be aware that you have satisfied a check rather than solved a problem.

You are scanning something for print. Set the scanner's resolution from the output size you need, not from the original's size. Scanning a 4×6 photograph that will be printed at 8×12 means doubling the linear dimensions, so you need twice the PPI at scan time — 600 rather than 300.

You are choosing a camera or a phone for print work. Divide the sensor's pixel dimensions by your intended print size in inches to get the PPI you would achieve. Most current phone cameras comfortably exceed 300 PPI for prints up to A4, which is larger than most people ever print.