Print-Ready Images: The DPI Myth, Pixel Math, and How to Prepare a File

A digital image has a fixed pixel count; that is the physical fact. DPI (dots per inch — what the printer puts down) and its input-side cousin PPI (pixels per inch — what the file proposes) describe how densely those pixels should be placed on paper. A 2400×1600 image carrying a '300 DPI' tag is not a different file than the same pixels tagged 72; the tag is a hint some software reads to compute a default placement size and other software ignores completely.

Which is why the common complaint — 'my photo is only 72 DPI, is it bad?' — is a category error. The pixel count is what you have to print with; the DPI field is a suggestion about size. The correct question is always: how many inches wide, and how many pixels do I have across those inches? Divide one by the other and you know whether you have 300, 150, 80 or 40 — and whether the print will be sharp, acceptable, soft or a pixel-grid.

The target depends on viewing distance, and the table everyone actually needs is short. Wallet 2×3: 600×900 px at 300 DPI. 4×6: 1200×1800. 5×7: 1500×2100. 8×10: 2400×3000. A4 full-bleed (297×210 mm): ~3500×2480. Letter 11×8.5: 2550×3300. Poster 24×36: at 150 DPI, 3600×5400; at 100, 2400×3600 — viewed from more than a metre away, the 100 version is indistinguishable to most eyes.

The sanity check worth memorising: a modern phone's 12 MP (4000×3000) main camera covers every print up to 8×10 at 300 and a 24×36 poster at 100 comfortably. If your file is bigger than the table demands, the printer downsamples it — which costs nothing. If it is smaller, upscaling is a cosmetic choice, not data recovery — the physics covered in the upscaling guide apply.

Start from the largest clean original — never from a print you scanned at a cheap setting, never from a chat-downloaded copy; generation loss is exactly the problem the pixel arithmetic cannot fix. Crop to the print's aspect ratio before sizing, because the printer's auto-fit will crop something you wanted. The reference line is 300 pixels per inch at the final print size for anything read at arm's length (photos, flyers, business cards, gallery prints); posters and signage read from distance and want only 100-150.

Colour reality: the printer works in CMYK and cannot reproduce the glowing saturated range your screen shows, and it has less shadow separation than a backlit display. The prep consequences: trust a proof print over the screen for deep shadows, keep saturation a notch below what the monitor suggests, and never judge print contrast from a screen at night. The shop that tells you your RGB file 'prints fine unmanaged' has a colour technician making it print fine.

Format: deliver a high-quality JPEG (the print pipeline's native assumption) or a lossless PNG/TIFF when the source is graphics; supply at the printer's requested DPI if they ask, but know the pixels are the same either way. Flatten any transparency to the intended background — a transparent logo printed on a dark flyer does not auto-magic its way onto the ink.

Three honest options, in ascending compromise. Print smaller: a 1200×1800 photo prints beautifully at 4×6 and 'only' 150 at 8×10, and 150 at a wall distance genuinely looks fine to eyes older than twenty-five. Modest upscaling: a 1.5-2× AI pass produces a printable file that survives a few feet of viewing distance — the upscaling guide's limits apply, and faces are the first casualty.

And the option people forget: change the print. A file that cannot carry a sharp 16×20 makes an excellent 8×10 framed, or a wall grid of 5×7s. Nobody who did not see the pixels will ever know that the 8×10 was the print the file was always big enough for.