Aspect Ratio Calculator

What an aspect ratio is, and what it is not

An aspect ratio is the width of a frame divided by its height, written as two numbers with a colon between them. A 1920 by 1080 video gives 1920 / 1080 = 1.7778, which is the ratio 16:9. That pair of numbers describes shape only. It says nothing about size, sharpness or file weight.

This is where most people get stuck. A 640 by 360 clip and a 3840 by 2160 clip are both 16:9, so they have the same shape and different amounts of detail. Resolution is the pixel count. The ratio is the shape those pixels are arranged in. You can change one without touching the other, and that is exactly what this calculator separates: it reports the ratio, the decimal form and the diagonal of the frame you typed, then shows the height that keeps the same shape at a new width.

The colon form is the friendly version. 16:9 means 16 units across for every 9 units down. The decimal form is what you need for math, because you cannot divide by a colon. Once you have 1.7778, the height for any width is that width divided by 1.7778, and the width for any height is that height multiplied by 1.7778.

Ratios also describe things that are taller than they are wide. A phone video is 9:16, which is 0.5625 as a decimal. An Instagram portrait post is 4:5, or 0.8. The math works the same way, and the answer is smaller than 1 instead of larger.

16:9 versus 4:3: which one to pick

16:9 is the default for anything that plays on a modern screen: YouTube, a TV, a laptop, a phone held sideways. It came from the shift to high definition and it is now the shape of almost every camera's video mode.

4:3 is the older television shape, and it is still the shape of most tablets, many point and shoot cameras and a lot of document scans. It is also the shape of an 8 by 6 inch print, which makes it handy for photos you plan to print rather than post. The practical difference is height. Take a 1920 pixel wide frame. At 16:9 it is 1080 tall. At 4:3 it is 1440 tall, so you get 360 more rows of pixels for the same width. That extra height is why 4:3 frames feel better for faces, for close work, and for anything with a strong vertical line.

The cost is compatibility. Put a 4:3 clip into a 16:9 player and it gets bars on the left and right, or it gets cropped at the top and bottom. Put a 16:9 clip into a vertical feed and it becomes a small strip in the middle of the screen. Neither is wrong, but you should pick the shape before you shoot, not after.

The two numbers that matter for most people are 16:9 and 9:16, because the same footage often has to live in both places. Shoot 16:9 and crop to 9:16 and you keep the middle 31% of the width. Shoot 9:16 and pad to 16:9 and you get wide bars. That trade is the whole reason people shoot vertical first and edit a wide version second.

Changing the width without stretching the picture

Keeping the shape means moving both sides by the same percentage, or solving for the missing side from the ratio. The second way is cleaner because it never drifts. New height equals new width divided by the ratio. New width equals new height multiplied by the ratio.

Start with 1920 by 1080, a 16:9 frame. Halve the width to 960 and the height has to halve too, to 540. Go to 1280 wide and the height is 1280 / 1.7778 = 720. That is the whole 720p case, and it is why 1280 by 720 looks like a smaller copy of 1920 by 1080 rather than a different picture.

Now do it the wrong way. Keep the height at 1080 and set the width to 1280. The frame is now 1280 / 1080 = 1.185, which is close to 6:5. Every circle becomes an oval and every face gets 33% narrower than it should be, because the width shrank by 33% while the height stayed put. This is the mistake that shows up when someone drops a clip into an editor, types a new width, and leaves the height alone with the chain link icon switched off.

On a web page you rarely need to do this by hand. The CSS property aspect-ratio in the CSS sizing specification lets a box hold its shape while one side is fluid, and browsers compute the other side for you. The same math sits underneath: one side, then a division.

What the big platforms ask for

Every platform publishes a preferred shape, and they mostly agree. Horizontal video is 16:9. Vertical video is 9:16. Feed images lean to 4:5 because a taller image takes more of a phone screen, and more screen means more attention.

The numbers in that table are the sizes the platforms quote in their own help pages, and they are all just the ratio scaled up. YouTube's own recommended upload settings list 1920 by 1080 at 16:9 as the standard, with 3840 by 2160 for 4K. The ratio is the rule and the pixel size is a suggestion, which is why a 1280 by 720 upload still plays correctly at 16:9.

Two practical consequences. First, a 16:9 landscape photo dropped into an Instagram feed gets scaled to 1080 wide and 608 tall, so it covers about half the height a 4:5 post would. Second, a vertical clip uploaded to a horizontal player gets pillarboxed, with bars on the sides, unless the player has a vertical mode. Neither platform will stretch your file, which is the one thing you can count on.

Cropping, letterboxing and the pixels you lose

When the target shape differs from the source, you have three options and only two of them keep the picture honest. Crop, which cuts pixels off the sides or top. Pad, which adds bars and keeps every pixel. Stretch, which distorts and should be avoided.

The arithmetic is simple enough to do in your head once you have the ratio. A 1920 by 1080 frame is 2.07 megapixels. Fit it to 1:1 by cropping the width down to 1080 and you keep 1.17 megapixels, so 43.8% of the frame is gone. Fit it to 4:3 at the same width and you need 1440 rows, which the source does not have, so you either pad 360 rows or crop the width down to 1440 and lose 25% of it.

Cropping is the usual answer for social video, and it is why editors talk about a safe area. If you know a clip will be cut to 9:16, keep faces and text inside the middle vertical band, roughly the central 31% of the width, so the crop has something usable in it. Padding is the safer choice for a slide deck or a screen recording, where the edges carry information that a crop would destroy.

Ratios that are close but not equal

Plenty of real screens sit a hair off the standard shapes, and the difference is usually a rounding artifact rather than a design choice. A 1366 by 768 laptop panel comes out at 1.7786, which is 0.05% away from 16:9. That is invisible, and it is not exactly 16:9 either, which is why resizing 1920 by 1080 to 1366 wide gives 768.4 and has to be rounded down to 768.

A 1440 by 900 monitor is 1.6, or 16:10. It is 10% taller than 16:9 at the same width, so a 16:9 video played full screen there gets a thin bar top and bottom, and a screenshot taken on it will not drop cleanly into a 1920 by 1080 timeline without a crop or a pad.

The gaps that catch people out:

• 1440 by 900 is 16:10, not 16:9. The difference is 11% in height.
• 1024 by 768 is 4:3, and it is the shape of a lot of older projectors.
• 2048 by 1080 is 1.896, the cinema 2K frame, not 16:9.
• 1080 by 1350 is 4:5, the shape Instagram prefers for a portrait photo.
• 1080 by 1080 is 1:1, and cropping a 16:9 clip to it throws away nearly half the width.

For encoding, whole even numbers matter more than the exact ratio. Most video codecs work in blocks of 2 or 16 pixels, and an odd dimension gets padded or rejected. When a calculation lands on 719.6, round to 720 and accept the 0.06% shape change rather than shipping a fractional size.

Pixel count, file size and why the ratio is only half the job

Two frames can share a ratio and differ by a factor of four in pixels. That factor is what drives file size, render time and how much a platform has to compress your upload.

720p at 1280 by 720 is 0.92 megapixels. 1080p at 1920 by 1080 is 2.07. 4K at 3840 by 2160 is 8.29, exactly four times 1080p because both sides doubled. At the same number of bits per pixel, a 4K file is about four times the size of the 1080p version of the same shot.

Put a number on it. A one minute clip at 8 megabits a second is 480 megabits, which is 60 megabytes. The same minute at 35 megabits a second, a common 4K rate, is 2,100 megabits or about 263 megabytes. The shape never changed, only the pixel count did.

The other half of the job is what the pixels contain. Upscaling a 1280 by 720 file to 1920 by 1080 keeps the 16:9 shape and adds no detail, because the encoder is inventing the extra pixels by interpolation. Downscaling loses detail permanently. If you have a choice, shoot and edit at the largest size you might need, then export down, because a 4K master gives you a clean 1080p file and a clean 9:16 crop, while a 1080p master gives you neither.