Image Size vs Dimensions: What’s the Difference?

“Reduce the image size” can mean two completely different things, and the wrong reading wastes an afternoon: a photo shrunk to a quarter of its pixels is still rejected by a form that wanted fewer kilobytes, and a photo compressed to 200 KB is still too big for a template that wanted 600 pixels. The two numbers are measured in different units, changed by different tools, and they can move in opposite directions.

Three numbers that get called “size”

The rest of this article is about the first two, because they are the ones forms, templates and upload boxes confuse. The third is covered in DPI vs PPI.

Dimensions: how much picture there is

An image is a grid of pixels, and its dimensions are the number of columns and rows in that grid, written width × height: 4032 × 3024, 1920 × 1080, 600 × 600. Multiply the two and you have the total number of pixels, which cameras quote in megapixels: 4032 × 3024 = 12,192,768, about 12.2 MP. The dimensions decide how much detail the picture can hold and how large it can be shown or printed before it looks soft. They say nothing about how many bytes it takes to store.

“Resolution” is the awkward word here. Camera makers and phone settings use it for the pixel dimensions (a “12 MP resolution”). Print shops and scanners use it for pixels per inch, which is a density, not a count. When a form says “minimum resolution 600 × 600”, it means dimensions; when it says “300 DPI”, it almost always means “enough pixels for the print size”, and the DPI guide works that number out.

File size: how many bytes it takes to store

File size is the length of the file on disk, in bytes, shown as KB or MB. It is what an upload limit of “2 MB”, an email attachment cap or a “keep it under 100 KB” rule is talking about. Five things decide it:

• The pixel count. More pixels means more to store, all else being equal. This is the only link between the two numbers, and it is why resizing also shrinks the file.
• The format. JPG and lossy WebP discard detail the eye barely notices and store the rest compactly. PNG stores every pixel exactly, which for a photo costs many times more bytes. JPG vs PNG vs WebP compares them.
• The quality setting. For JPG and WebP, a lower quality throws away more detail and saves more bytes, at the same pixel dimensions.
• What is in the picture. Flat colour, sky and plain walls compress to almost nothing; foliage, fabric, grain and noise barely compress at all. Two photos with identical dimensions and settings can differ in size several times over.
• Metadata. Camera settings, dates and location travel inside the file, but they are small: the test photo carries about 15 KB of it, under half a percent of the file.

One more trap: a kilobyte is 1,000 bytes to some programs and 1,024 to others, so the same file can read 48.8 KB in one place and 50 KB in another. When a limit is tight, work in bytes; the 50 KB guide shows how.

Why a bigger image can be a smaller file

Because the two numbers have different causes, they do not have to agree. We ran the 12-megapixel test photo through the image resizer in Chrome and saved the result in different ways. Every file in the table is the same picture.

Three rows tell the story. The PNG has a sixth of the original’s pixels and is a fifth larger, because lossless storage of a photo costs more than the camera’s own JPG compression saved. The two JPGs at 1600 × 1200 are the same picture at the same size and differ by 40% on quality alone. And the last row has exactly the original’s dimensions at a little over half its bytes, yet it is the worst file in the table: it was made by enlarging the 2016 × 1512 sample copy, so it holds the detail of 3 megapixels spread across 12. Dimensions tell you how many pixels there are, not how much each one is worth.

Which tool changes which number

• Resizing changes the dimensions, and the file size follows: fewer pixels, fewer bytes. It is the right tool when a requirement is in pixels, and a strong first move when it is in KB, because a 4000-pixel photo rarely needs to stay that wide. Shrinking keeps every detail that could be seen at the new size; only enlarging costs quality, because it has to invent pixels.
• Compressing changes the file size and leaves the dimensions alone. The image compressor keeps 4032 × 3024 at 4032 × 3024 and stores it in fewer bytes. It is the only answer when the pixels must stay as they are.
• Cropping changes the dimensions by removing part of the picture, and the file shrinks in proportion to what was cut.
• Converting changes the format, so it changes the file size, often dramatically, without touching a single pixel.

Reading a requirement

Most rules name their unit, if you know what to look for:

• “Maximum 2 MB”, “under 100 KB”, “up to 5 MB per file”: file size. Compress, and resize if the image is far larger than it will ever be displayed.
• “At least 600 × 750 pixels”, “1080 × 1080”, “minimum 5 MP”: dimensions. Resize or crop; no amount of compression changes these.
• “300 DPI”: usually dimensions in disguise. Multiply the print size in inches by 300 to get the pixels you need.
• “Image size” with no unit: look at the examples the form gives. “Image size: 2 MB” is bytes; “image size: 800 px” is dimensions. A form that gives both means both, and you meet the pixel rule first, then the byte rule.

How to see both numbers

The image dimensions checker reads the width and height from the decoded picture and shows the file size next to them, for up to 20 files at once. Without it: on Windows, right-click the file, choose Properties and open Details, where Dimensions and Size sit a few rows apart; on a Mac, select the file and press Command-I, and Get Info lists the size at the top and the dimensions under More Info. In both, the number with “KB” or “MB” after it is the file size, and the pair with “×” between them is the dimensions.