Image Resolution Explained: Pixels, Dimensions, PPI and DPI
Four terms get mixed up all the time: pixel dimensions, megapixels, PPI and DPI. Only one of them decides how much detail an image actually has.
Pixel dimensions: the number that matters
An image is a grid of pixels. A photo that is 4,000 × 3,000 pixels has 4,000 columns and 3,000 rows, and that grid is all the detail it will ever contain. On screens, pixel dimensions are the only thing that counts: a 1,200-pixel-wide image fills a 1,200-pixel-wide space, whatever its DPI says.
Megapixels
Megapixels are simply width × height ÷ 1,000,000. A 4,000 × 3,000 photo is 12 megapixels. The apples photo in our compression example is 2,576 × 1,717, about 4.4 megapixels, while a modern phone camera captures 12–50 megapixels. Higher megapixels allow bigger prints and tighter crops, but they also mean bigger files and more memory when you edit, which is why very large images can fail on some phones.
PPI and DPI: only for printing
PPI (pixels per inch) says how many pixels are placed in each printed inch. DPI (dots per inch) is technically about printer ink dots, but people use it to mean the same thing. Changing the DPI number in a file doesn’t add or remove a single pixel; it only changes the suggested print size.
The formula is: print size in inches = pixels ÷ PPI. A 3,000 × 2,400 image prints at 10 × 8 inches at 300 PPI, or 20 × 16 inches at 150 PPI. The pixels are identical; they’re just spread over a larger area.
How many pixels do you need?
- Website images: 1,200–2,000 pixels wide is plenty for most layouts. Larger images just slow pages down.
- Social media: each platform has its own size, such as 1,080 × 1,350 for an Instagram portrait post.
- Prints: about 300 PPI for photos you’ll look at closely. A 6 × 4 inch print needs about 1,800 × 1,200 pixels.
- Application forms: the form states the pixels, such as 600 × 600 for a US visa photo.
Can you add resolution?
Resizing an image to more pixels with ordinary methods spreads the existing detail more thinly, so it looks soft. An AI upscaler can predict plausible new detail and gives sharper results, but it can’t recover information that was never captured. To go smaller, use Resize Image, which reduces pixel dimensions and usually file size too.