What each one measures
Dimensions are how many pixels wide and tall the image is — 4000 × 3000, 600 × 600, 1920 × 1080. This is the size of the grid.
File size is how many bytes the file occupies on disk — 240 KB, 3.7 MB. This is how much space it took to describe that grid.
The relationship between them is loose, because compression sits in between. Two images with identical dimensions can differ in file size by a factor of fifty. A 600 × 600 photograph saved as a high-quality PNG might be 900 KB; the same 600 × 600 image as a moderate JPEG might be 40 KB. Same grid, wildly different files.
Why a small picture can be a big file
File size depends on three things, and dimensions are only one:
- How many pixels there are. Doubling both width and height quadruples the pixel count, so this dominates when the difference is large.
- How complicated the image is. A photograph of foliage — thousands of tiny distinct details — compresses far worse than a photograph of a clear sky. Compression works by finding predictable patterns, and detail is the absence of prediction.
- Which format and quality setting was used. Lossless formats store everything; lossy formats throw away what you are least likely to notice. This routinely makes a ten-fold difference.
This is why "just make it smaller" is ambiguous advice. Smaller in what sense?
Working out which limit you failed
Read the form's wording carefully, because it usually says, just not prominently:
- "Maximum 2 MB", "under 500 KB", "file must not exceed…" — a byte limit. Compress the image; you may not need to change its dimensions at all.
- "Maximum 600 × 600", "images must be at least 1200 px wide", "35 × 45 mm" — a dimension limit. Resizing is the only thing that helps; compressing further will not change the pixel count.
- Both stated together — common on visa and ID photo portals, and the reason those are the most frustrating uploads. Resize first, then compress to the byte limit, in that order.
If the form gives no error detail at all, check dimensions first. A byte limit failure is usually reported clearly; a dimension failure is more often silent or vague.
Which lever to pull first
When only the byte count is a problem, reduce quality before you reduce dimensions. At normal viewing sizes a quality reduction is much harder to notice than a resolution reduction, and dropping a JPEG from maximum quality to around 70% commonly halves the file with no visible change.
That ordering is exactly what the compress-to-a-size tool does: it searches for the highest quality that fits your budget and only scales the image down if no quality setting can get there.
When the dimension limit is the problem, resizing is unavoidable — and it will also reduce the file size as a side effect, often dramatically. Halving both width and height removes 75% of the pixels.
The resolution myth
You may see instructions asking for "300 DPI" images. DPI — dots per inch — is a number stored in the file's metadata that says how large the image should be *printed*. It has no effect on the pixels or the file size, and changing it does not make an image sharper.
A 600 × 600 image is 600 × 600 whether it is tagged 72 DPI or 300 DPI. What changes is the physical size it would print at: 600 pixels at 300 DPI is a two-inch square; at 72 DPI it is an eight-inch square. If a print shop asks for 300 DPI at a given physical size, what they actually need is enough pixels — multiply the inches by 300 and supply that many.
Practical numbers
- A current phone camera produces roughly 4000 × 3000 pixels, 3–8 MB per photo.
- A full-screen laptop screenshot is around 2560 × 1600, and as PNG can easily exceed 2 MB.
- An image displayed at 800 pixels wide on a web page gains nothing from being 4000 wide — it is downscaled in the browser, having already cost the visitor the full download.
- Visa and ID photos are usually specified in millimetres with a pixel minimum; both must be satisfied.