An aspect ratio is the shape of a rectangle with the size taken out of it: the width divided by the height. 16:9 means 16 units across for every 9 units down, so 1920 × 1080, 1280 × 720 and 640 × 360 are all the same shape and all 16:9. The ratio says nothing about how big the picture is or how sharp it will look. It only decides whether it fits the hole you are putting it in, or whether something gets cropped, squashed or surrounded by black bars.
How do I read 16:9 and 2.39:1?
There are two conventions and they describe the same thing.
Two whole numbers — 16:9, 4:3, 3:2, 1:1 — is how screens, photos and social platforms write it. The pair is the ratio reduced as far as it will go, so 1920:1080 becomes 16:9 once you divide both sides by 120.
A decimal against 1 — 1.85:1, 2.39:1, 1.43:1 — is how cinema writes it, because film ratios rarely reduce to anything tidy. 2.39:1 means the frame is 2.39 times wider than it is tall. You can convert either way with one division: 16 ÷ 9 is 1.778, so 16:9 and 1.78:1 are the same shape written twice.
The order never changes. Width first, height second. A ratio written 9:16 is the portrait version, not a typo.
How do I work out the aspect ratio of an image?
Divide the width by the height, then reduce the fraction. For 1600 × 1200: the division gives 1.333, and the greatest common divisor of 1600 and 1200 is 400, which reduces the pair to 4:3.
The more common job is the other direction — you know the shape and one dimension, and you need the other. Cross-multiply: new height = new width × original height ÷ original width. A 1920 × 1080 source at 1280 wide gives 1280 × 1080 ÷ 1920, which is 720.
That arithmetic is easy and easy to fumble at four in the afternoon, which is the entire reason the aspect ratio calculator exists: type three numbers and the fourth appears, along with the reduced ratio and a note when the result does not land on a whole pixel. It works on numbers only — no file goes into it and nothing comes out resized, so what you get is the pair of dimensions to hand to whatever does the resizing. Doing it by hand is fine. Doing it by hand eleven times for a set of banner sizes is how a 1 px seam ends up in production.
Which aspect ratios are actually in use?
- 16:9 (1.778) — video, laptop and TV screens, and the default for anything that moves. It is a deliberate compromise rather than anything found in nature: Kerns Powers put it to an SMPTE working group in 1984, having cut equal-area rectangles for every ratio then in use and stacked them on a common centre. 1.78:1 was both the smallest shape that contained all of them and the largest shape all of them contained. It also lands near the geometric mean of 4:3 and 2.35:1, which is 1.77.
- 3:2 (1.5) — the shape of a 35 mm negative, and still what most mirrorless and DSLR cameras produce. A 6000 × 4000 file is 3:2.
- 4:3 (1.333) — older televisions and monitors, most tablets, and many phone sensors at their full resolution.
- 1:1 — square. Avatars, album art, and the safest shape when you have no idea how a feed will crop it.
- 4:5 (0.8) and 9:16 (0.5625) — portrait and vertical video. Anything below 1 is taller than it is wide.
- 2.39:1 — widescreen cinema, the reason a film on a 16:9 television has bars top and bottom.
Two traps in that list. Phone screens are not 16:9 any more — most are around 19.5:9 or 20:9, taller than the video they play, which is why full-screen video on a phone either has side bars or loses its edges. And 21:9 is a marketing label rather than a measurement: monitors sold as 21:9 are usually 2560 × 1080, which reduces to 64:27, or 3440 × 1440, which is 43:18. Calculate against a literal 21:9 and you will be out by enough to see.
Why does the height come out as 562.5?
Because pixels are whole and ratios are not. Scale 16:9 to 1000 px wide and the exact height is 562.5 px. Round it to 563 and the real ratio is now 1.7762 instead of 1.7778 — a shift of under a tenth of a percent, invisible in isolation.
It stops being invisible in two situations. The first is when the rounded image sits inside a container sized to the true ratio: the difference shows up as a one-pixel line of background down one edge, and it will look like a rendering bug rather than arithmetic. The second is video. Nearly all common video encoding uses 4:2:0 chroma subsampling, where the colour planes are half the width and half the height of the brightness plane, so both dimensions have to be even. An odd height gets rejected or silently nudged.
The fix is to pick widths that divide cleanly. For 16:9 use multiples of 16, for 4:3 use multiples of 4, and the rounding never happens: 1280, 1600 and 1920 all give whole heights at 16:9. All of this assumes square pixels, which anything modern has. DVD-era footage stored at 720 × 480 and displayed at 16:9 does not, so for that material the stored frame and the displayed frame are two different shapes and the arithmetic here describes only the stored one.
Can I change an aspect ratio without ruining the image?
Not for free. Moving a picture from one shape to another has exactly three outcomes, and you are choosing between them whether you notice or not.
Crop. Cut pixels off two sides until the remainder is the shape you want. You lose content, and the content you lose is at the edges, which is where heads and hands live. Going from 3:2 to 16:9 means taking about 16% off the height.
Pad. Add bars to two sides. Nothing is lost, but the subject now occupies less of the frame, and on a feed that crops thumbnails you can end up with bars inside bars.
Stretch. Force the pixels into the new shape. This is the one nobody chooses on purpose and everybody ships at least once. It is instantly visible on faces, circles and type.
Crop is usually the right answer, and it is a decision about composition rather than arithmetic — which is why it wants a preview rather than a formula. Dragging a fixed-ratio box over the picture in a browser-based cropper shows you what you are giving up before you commit to it. It keeps 16:9, 4:3, 3:2, 1:1, 4:5 and 9:16 as locked presets, so the shape stays exact while you decide where the box sits.
Aspect ratio is not resolution, and it is not DPI
1280 × 720 and 3840 × 2160 are both 16:9. One is 0.92 megapixels and the other is 8.3, and the ratio cannot tell them apart, because shape and sharpness are separate questions. If what you are actually deciding is how many pixels to ship, the right image size for the web is the question underneath that one.
DPI is a third, unrelated thing. It is a number stored in the file that tells a printer how large to make the print. It has no effect on a screen at all. Changing an image from 72 to 300 DPI without touching the pixel dimensions changes nothing a browser can see.
How do I hold an aspect ratio in CSS?
Modern CSS has a property for this: aspect-ratio: 16 / 9 on an element makes the browser derive the height from whatever width the layout gives it. It is supported in every current browser, and it replaces the old trick of a wrapper with padding-top: 56.25% — a percentage padding resolves against the container's width, and 9 ÷ 16 is 56.25%. You will still meet that pattern in older code. There is no reason to write new code that way.
One caveat: aspect-ratio is a suggestion, not a law. If the content inside the box is taller than the ratio allows and nothing constrains it, the box grows and the ratio is ignored. Pair it with object-fit: cover on images, or overflow: hidden, if the shape has to hold no matter what goes in.
Why is A4 paper a different shape from every screen?
A4 paper is 1:1.414, which is 1:√2. That number is the only ratio where cutting the sheet in half across its long side produces two sheets of the same shape. That is the whole design of the ISO 216 paper series: A4 halves into A5, A5 into A6, and nothing ever needs rescaling. No screen format has that property, which is why a document laid out for A4 never fits a monitor without a decision.
When you need the fourth number rather than the theory, the aspect ratio calculator takes the three you have, reduces the ratio, and says how far a rounding moved the shape when the width you picked does not divide cleanly. Its table of common widths marks the ones that come out odd, which is the check worth doing before a size goes near a video encoder.
Knowing the shape is the easy half of the job. Deciding which part of the picture to lose is the other half, and cropping to an exact aspect ratio takes it from here: which side to cut, where the rectangle sits, and the numbers for the shapes you get asked for most.
Frequently asked questions
What does 16:9 mean?
It means the picture is 16 units wide for every 9 units tall, which works out to 1.778 times wider than it is high. It describes shape only, so 640 by 360 and 3840 by 2160 are both 16:9. It is the standard shape for video, televisions and most laptop screens.
How do I find the aspect ratio of an image?
Divide the width in pixels by the height, then reduce the two numbers by their greatest common divisor. A 1600 by 1200 image divides to 1.333 and reduces to 4:3. If the pair shares no common factor, such as 1921 by 1080, the only honest form is the decimal.
Is 4:3 or 16:9 better?
Neither is better in general; they suit different things. 16:9 matches how screens and video are built, so it fills a display with no bars. 4:3 is taller and captures more of a vertical subject, which is why many camera sensors and tablets still use it.
Can I change the aspect ratio of a photo without cropping?
Only by adding bars to two sides, which keeps every pixel but shrinks the subject within the frame. The alternative, stretching the image to fit, distorts everything and is obvious on faces and circles. Cropping is usually the better trade because you control what gets cut.
Is aspect ratio the same as resolution?
No. Resolution is how many pixels there are; aspect ratio is the shape those pixels form. 1280 by 720 and 3840 by 2160 share the ratio 16:9 but differ by nine times in pixel count, so one is sharp on a large screen and the other is not.
Why does my aspect ratio calculation give a decimal instead of a whole pixel?
Because the width you picked is not divisible into the ratio. A 16:9 image at 1000 pixels wide needs a height of 562.5, which has to be rounded, shifting the true shape slightly. Choosing a width that is a multiple of 16 avoids it entirely for 16:9.
Last updated September 19, 2026