Aspect Ratio Calculator

It is the shape of a rectangle expressed as the relationship between its width and its height, with the size stripped out. Writing 16:9 says that for every 16 units across there are 9 down, and says nothing at all about how big the picture is. That is the whole point: shape and size are separate questions, and a ratio answers only the first. Two numbers separated by a colon is the usual form, but cinema prefers a single decimal against 1 (1.85:1), and both say the same thing.

Because both reduce to the same pair. Divide 1920 and 1080 by their greatest common divisor, 120, and you get 16 and 9. Divide 1280 and 720 by theirs, 80, and you get 16 and 9 again. The test without any division is cross-multiplication: 1920 × 720 and 1080 × 1280 are both 1,382,400, so the two rectangles are the same shape. One simply has 2.07 million pixels in it and the other 0.92 million, which is a question about detail, not about shape.

Multiply or divide by the ratio. To find the height at a known width, divide the width by the ratio's decimal value: 1920 ÷ (16 ÷ 9) = 1080. To find the width at a known height, multiply instead: 1080 × (16 ÷ 9) = 1920. The complication is that the answer is rarely a whole number (1920 at 2.39:1 wants a height of 803.35 pixels), and since there is no such thing as a third of a pixel, something has to give. Rounding down to 803 makes the frame slightly shorter and therefore slightly wider: 2.391:1 rather than 2.39:1, a difference of 0.043%, which this page shows rather than swallowing.

16:9 is 1.778:1 and belongs to television and streaming; it exists because it was a compromise between the 4:3 of old broadcast and the wide shapes cinema was using. 1.85:1, called “flat”, is the narrower of the two standard theatrical release shapes and is achieved by masking the top and bottom of a nearly-full frame. 2.39:1, called “scope”, is the wide one, historically produced with anamorphic lenses that squeeze a wide view onto a normal-shaped negative. The practical consequence is that a scope film on a 16:9 television has thick black bars, a flat film has thinner ones, and a television programme has none. The bars are not a fault, they are the difference between the two shapes.

Because it used to be right. CinemaScope and its successors delivered close to 2.35:1 from the 1950s until SMPTE revised the standard in 1970, mainly to hide splice damage at the frame edges, after which the projected shape became 2.39:1. Habit outlived the change, so plenty of catalogue descriptions, film-stock references, and forum posts still say 2.35 for films that are genuinely 2.39. The two are only 1.7% apart, which is about fourteen pixels of height on a 1920-wide frame: invisible on a sofa, and very visible if you are cutting a matte.

Because 21:9 is marketing shorthand and no shipping panel has that shape. A 2560 × 1080 ultrawide reduces to 64:27, which is 2.370:1; that number is 16:9 cubed, chosen so a 16:9 image inside it scales cleanly. A 3440 × 1440 ultrawide reduces to 43:18, which is 2.389:1. A true 21:9 would be 2.333:1, and neither panel is that. The gap between the two ultrawides is small, roughly 0.8%, but it is enough that a wallpaper or an overlay cut for one leaves a visible strip on the other.

They are the three ways of reconciling a picture with a frame of a different shape. Letterboxing keeps the whole picture and puts bars above and below it, which happens when the picture is wider than the frame. Pillarboxing is the same idea rotated: bars down the left and right, for a picture narrower than the frame. Cropping refuses the bars and enlarges the picture until it covers the frame, throwing away whatever falls outside: a 1920 × 1080 still filling a 1080 × 1350 feed post loses 55% of its width, well over half the shot. The fourth option, stretching, keeps everything and fits everything, at the cost of making every face in the shot the wrong shape.

Because of pixel aspect ratio. Most modern formats use square pixels, so 1920 × 1080 stored is 1920 × 1080 displayed. Anamorphic and legacy broadcast formats do not: DVD-era 16:9 video is stored as 720 × 480 or 720 × 576, which are nothing like 16:9, and carries a flag telling the player to stretch each pixel horizontally on the way out. The same trick appears in modern anamorphic cinematography, where a 2× lens squeezes a 2.39:1 view onto a 1.19:1 negative. So a file's stored resolution is the storage aspect ratio, and only stored resolution multiplied by pixel aspect ratio gives the display aspect ratio, which is the one your eyes judge.

Because of chroma subsampling. Nearly all delivered video uses 4:2:0, which stores full-resolution brightness but only one colour sample for each 2 × 2 block of pixels, on the reasoning that eyes are far more sensitive to brightness detail than colour detail. A frame with an odd width or height has a leftover row or column with no complete colour block to belong to. H.264 and HEVC encoders therefore reject odd dimensions outright or pad them silently, and a padded frame is one row taller than you asked for. If the output is going anywhere near a codec, snap both sides to even numbers; the tiny ratio error that costs is far cheaper than an encode that fails at midnight.

Not if both sides are scaled by the same factor: that is the definition of keeping the ratio, and it is why halving 1920 × 1080 to 960 × 540 is safe. It changes when rounding forces the two sides to move by different proportions, which is most likely on small images and awkward ratios: scaling 1366 × 768 by 150% gives 2049 × 1152, and rounding 2049 up to 2050 to keep it even shifts the shape by about 0.05%. Below a few hundred pixels this stops being academic, because a single pixel is a much larger share of the total.