Someone asks for a photo "at 300 DPI" and you have no idea whether the one on your desktop qualifies. Or you resize an image and the file barely shrinks. Or a print shop rejects a picture that looks perfectly sharp on your monitor. All three confusions come from the same place: mixing up three things that are related but not the same.
The three numbers
- Pixel dimensions — how many pixels wide and tall, like 3000 × 2000. This is the only measure of how much detail the image actually contains.
- DPI (or PPI) — a number written into the file saying how densely those pixels should be packed when printed on paper. It is an instruction for printers.
- File size — how many megabytes the file occupies, which depends on the pixel dimensions, the format, and the compression setting.
DPI does nothing on a screen
This is the single most useful thing to understand, so it is worth being blunt about it.
So when a web designer asks for "72 DPI images for the web", the number is a leftover habit from the print world. What they actually need is sensible pixel dimensions. You can change an image's DPI tag from 72 to 300 and back all day without altering a single pixel.
Where DPI genuinely matters
Print. Paper has no pixels, so a printer needs to know how big to make each one. DPI is that instruction, and the arithmetic is simple:
A 3000 × 2000 photo at 300 DPI prints at 10 × 6.7 inches.
The same file at 150 DPI prints at 20 × 13.3 inches — twice as big, half as sharp.
300 DPI is the traditional standard for quality printing because it is roughly the point at which a person holding a page stops being able to see individual dots. Large posters viewed from a distance are routinely printed at 150 DPI or lower, because nobody presses their nose against a billboard.
The practical consequence: when a printer asks for 300 DPI, they are really asking whether you have enough pixels for the physical size they intend to print. Multiply the inches by 300 and see whether your image has that many pixels. If it does not, changing the DPI number in the metadata will not help — you need more pixels, or a smaller print.
What actually drives file size
File size is governed mostly by total pixel count, and pixel count grows with the square of the dimensions. This catches people out constantly:
| Dimensions | Total pixels | Relative size |
|---|---|---|
| 4000 × 3000 | 12.0 million | 100% |
| 2000 × 1500 | 3.0 million | ~25% |
| 1000 × 750 | 0.75 million | ~6% |
Halving both width and height leaves a quarter of the pixels, not half. This is why resizing is by far the most effective way to shrink an image, and why it usually beats turning the compression up.
The mistake almost everyone makes
Uploading a photo straight off a camera or phone to a website. A modern phone shoots something like 4000 × 3000. Displayed in a blog post 800 pixels wide, roughly 96% of that data is downloaded, decoded, and then thrown away by the browser before anything appears on screen. The visitor waits for megabytes they will never see.
Resize to the size it will actually be displayed at — or twice that, if you want it to stay sharp on high-density phone and laptop screens — and the file typically drops by ninety per cent or more with no visible difference at all.
Sensible target dimensions
| Use | Suggested width |
|---|---|
| Thumbnail | 300–400 px |
| Image inside a blog post | 1200–1600 px |
| Full-width hero banner | 1920–2400 px |
| Email attachment | 1200–2000 px |
| Print at 300 DPI | inches × 300 |
Enlarging: what is and isn't possible
Making an image smaller is easy — you are discarding information, and the result looks fine. Making one bigger means inventing information that was never captured, which is a fundamentally harder problem.
Simple enlargement stretches existing pixels and produces a soft, blurry result: a 500 px image blown up to 2000 px has no more real detail than it started with. Modern upscaling does better by reconstructing plausible edges and texture, and the results can be genuinely impressive — but it is still an educated guess, not recovery. Detail that was never recorded cannot be restored.
Aspect ratio, and why images stretch
Aspect ratio is the proportion of width to height. If you change one dimension without changing the other proportionally, the image distorts — faces go wide or tall in a way that is immediately obvious even when people cannot say why.
Keep the proportions locked when you resize. If you need a different shape — a square crop for a profile photo, say — crop the image rather than squashing it. Crop removes part of the picture; resize changes its size. Confusing the two is what produces stretched images.
Common questions
How do I change an image to 300 DPI?
Ask first what physical size it needs to print at. Multiply those inches by 300 to get the pixels required. If your image already has at least that many, you are fine — the DPI tag itself is just metadata. If it does not, you need a higher-resolution original.
Why didn't my file get smaller when I lowered the DPI?
Because DPI is only a metadata label. Changing it alters no pixels and therefore no file size. To shrink the file, reduce the pixel dimensions or increase the compression.
Is DPI the same as PPI?
In everyday use, yes, and the terms are used interchangeably. Strictly, PPI (pixels per inch) describes an image or display and DPI (dots per inch) describes a printer's ink dots. The distinction rarely matters outside professional printing.