Dimensions, file size and the difference between them
These get conflated constantly and they are not the same thing. Dimensions are how many pixels wide and tall an image is. File size is how many bytes it occupies on disk. They are related — fewer pixels generally means fewer bytes — but you can change either without changing the other much, and forms sometimes specify both independently.
When a form asks for an image of 200 by 230 pixels, it means dimensions and it usually means exactly: some systems reject anything that is not the stated size rather than scaling it for you. When a form asks for an image under 50 KB, it means file size, and the dimensions are your business. Read the requirement carefully, because satisfying the wrong one is a common reason for rejection.
Reducing dimensions is the most powerful lever available on file size, because pixel count scales with the square of the dimensions. Halving both width and height leaves you with a quarter of the pixels, which typically produces a file well under a third the size before you touch quality at all.
Aspect ratio, and what happens when you break it
Aspect ratio is the proportion between width and height. Change both by the same factor and the image is simply smaller. Change them by different factors and the image is stretched — faces widen, circles become ovals, and text leans. It is instantly obvious to anyone looking, and on a document or an identity photograph it looks like a manipulation rather than a mistake.
This is why locking the aspect ratio is the default. Set the width and the height follows automatically, and the image stays honest. Unlock it only when a system demands exact dimensions and you have already decided that a small distortion is acceptable — which, on anything showing a face or a signature, it usually is not.
When a form demands exact dimensions with a different ratio than your image, the correct answer is to crop rather than to stretch. Crop to the required proportions first, keeping the subject where it needs to be, then resize the cropped image to the exact pixel dimensions. That gives you the size the form wants and an image that still looks like the thing it depicts.
Enlarging, and why it disappoints
Making an image larger cannot add detail that was never captured. Scaling a 400-pixel-wide photograph to 1600 gives you an image with sixteen times the pixels and exactly the same information, spread more thinly. The result is soft, and pushing sharpening at it produces halos rather than detail.
Small enlargements — up to perhaps 150 percent — are usually acceptable and often unnoticed. Beyond that the softness becomes obvious, particularly on text and edges. If you need a genuinely larger image, going back to the source and re-exporting or re-scanning at the size you need will always beat enlarging afterwards.
Reducing, by contrast, is nearly free in quality terms and frequently improves apparent sharpness, because detail is being averaged into fewer pixels. If you have a choice, capture large and reduce, rather than capture small and enlarge.
Common size requirements and what they are for
A few dimensions come up constantly because institutions standardised on them. Passport-style photographs are frequently specified at 200 by 230 pixels for online forms, or 35 by 45 millimetres for print. Signatures are often wanted at around 140 by 60 pixels and under 20 KB. Profile photographs on most platforms are square and displayed small, so anything above about 400 pixels is wasted.
When a form gives both a pixel size and a file size limit, satisfy the dimensions first and the file size second, because resizing changes the file size but not the reverse. When it gives a range rather than an exact figure, aim for the middle of it — some systems check strictly at the boundaries and reject values that are technically within them.