Change a PNG's DPI

writes a real pHYs chunk · shows print size · 72 / 150 / 300 / 600 presets

A printer or journal asks for "300 dpi" and your PNG reports 72. This writes the resolution tag that software reads — a proper pHYs chunk with a correct CRC — and tells you what physical size your pixels then represent. Pixels are not changed or invented.

Read this before you file the ticket

Changing DPI does not add detail. It is metadata — a note in the file saying "these pixels are intended to be this many per inch". A 600×600 pixel image tagged 300 dpi is a two-inch square; tag the identical file 72 dpi and it's an 8.3-inch square. Same pixels, same sharpness, different intended physical size.

So if a printer says your image is "too low resolution for 300 dpi at A4", re-tagging won't fix it. A4 at 300 dpi needs roughly 2480×3508 actual pixels. You need a bigger source, or you accept a smaller print. Upscale can interpolate to those dimensions and will look softer than a true high-resolution original — sometimes that's an acceptable trade, sometimes it isn't, but at least it's an honest one.

What gets written

PNG stores resolution as pixels per metre in a pHYs chunk, so your dpi is converted (dpi ÷ 0.0254) and written with the unit flag set to metres. Any existing pHYs is replaced, and the new chunk is placed immediately after IHDR where the specification requires it — before IDAT. The CRC is recomputed. Because PNG has no separate horizontal and vertical setting here, both axes get the same value, which is what every normal image wants.

Drop a JPG or WebP and it's rendered to PNG first, then tagged — the output is always a PNG. JPEG resolution lives in JFIF and EXIF headers instead, which is a different mechanism.

FAQ