Pixel Aspect Ratio Calculator (Non-Square Pixels)

Background.

Most aspect-ratio arithmetic quietly assumes that pixels are square — that a 1920 × 1080 frame is 16:9 because 1920 ÷ 1080 = 1.7778. For every modern format that assumption is correct and guaranteed: Recommendation ITU-R BT.709 specifies a pixel aspect ratio of 1:1 for HD, and BT.2020 carries the same square pixels into 4K and 8K. If your pixels are square, the aspect ratio calculator is the tool you want, and this page has nothing to add.

This page is for the cases where they are not. Standard-definition video, DV and DVD masters, HDV, and anything described as anamorphic store frames whose pixels are taller than they are wide, or wider than they are tall, and the frame only takes its intended shape when the display stretches them back. Get that wrong and faces look thin or fat — the single most common symptom of importing old footage into an editor that assumes square pixels.

Three ratios are involved, and one identity connects them. The storage aspect ratio is the stored pixel dimensions, width over height. The pixel aspect ratio is the shape of one pixel. The display aspect ratio is what you actually see, and it is the product of the other two: DAR = SAR × PAR. Give the calculator any two and it returns the third, plus the square-pixel resolution to rescale to.

The standard pixel ratios are not folklore; they are a numbered table in ITU-T Recommendation H.264. `aspect_ratio_idc` 1 is 1:1 square, 2 is 12:11, 3 is 10:11, 4 is 16:11, 5 is 40:33, and the list continues through 24:11, 20:11, 32:11, 80:33, 18:11, 15:11, 64:33, 160:99, 4:3, 3:2 and 2:1, with 255 reserved for an explicitly signalled pair. The specification also requires that the two numbers be relatively prime, which is why this calculator reduces its labels with a greatest-common-divisor rather than rounding to a decimal.

There is one trap that defeats most published pixel-ratio arithmetic, and it is worth stating plainly. Applying the standard 10:11 ratio to a full 720 × 480 frame gives (3/2) × (10/11) = 15/11 = 1.3636, which is not 4:3. That is not a mistake in the ratio: Table E-1's own note explains that the 4:3 shape belongs to the 704 samples that are actually visible, and the extra 8 samples at each edge are horizontal overscan. Do the same arithmetic on 704 × 480 and it lands on exactly 4:3 — and the square-pixel form comes out at exactly 640 × 480. Whenever a pixel-ratio calculation refuses to close on a round number, the 720-versus-704 distinction is the first thing to check.

A second, subtler point: two different conventions are in circulation and both are correct. ITU-T H.264 gives the 704-based ratios (10:11 and 12:11). Recommendation ITU-R BT.500 Table 1-1 lists the same formats measured across the full 720 samples and gives 0.89 for 720 × 483 and 1.07 for 720 × 576. They differ by exactly the overscan, and neither is wrong — they answer slightly different questions about the same signal. This page lets you enter whichever pair your file or your workflow uses, and shows the consequences of the choice rather than picking one for you.

The second mode runs the identity backwards. If you know a frame is 704 × 576 and is meant to be seen as 16:9, the pixel ratio it needs is (16 × 576) ÷ (9 × 704) = 9216 ÷ 6336, which reduces to 16:11 — H.264's `aspect_ratio_idc` 4, recovered from the geometry alone. That is the mode to use when a file has lost its metadata and you have to reconstruct what it should be.

What is pixel aspect ratio calculator?

Pixel aspect ratio (PAR), called sample aspect ratio in the video-coding standards, is the shape of an individual pixel: its width divided by its height. A PAR of 1:1 means square pixels, which every modern display format specifies. A PAR of 10:11 means each pixel is slightly taller than it is wide, so the stored frame has to be stretched horizontally to display correctly.

It matters because it breaks the usual assumption that a frame's shape can be read off its pixel dimensions. The stored dimensions give the storage aspect ratio (SAR); the shape you see is the display aspect ratio (DAR); and DAR = SAR × PAR. ITU-T Recommendation H.264 enumerates the standard values in Table E-1 — 1:1, 12:11, 10:11, 16:11, 40:33 and a dozen more — and requires that the two numbers be relatively prime. Correcting a non-square frame means rescaling it so that the pixels become square while the display shape is preserved, which can be done by widening the width or by shortening the height.

How to use this calculator.

1. 01Enter the stored frame dimensions exactly as the file reports them — the coded width and height in samples, not the shape you expect to see.
2. 02In the first mode, enter the pixel aspect ratio as two numbers: 10 and 11 for 525-line 4:3, 12 and 11 for 625-line 4:3, 40 and 33 for 525-line 16:9, 16 and 11 for 625-line 16:9, 1 and 1 for square.
3. 03Read the display aspect ratio: this is the shape the frame is meant to be seen at, and the number to set in your editor's sequence settings.
4. 04Use the square-pixel width to rescale for a square-pixel workflow — the stored height stays as it is. The square-pixel height output is the equivalent correction applied the other way.
5. 05If the answer does not land on a familiar ratio, check whether you should be using 704 rather than 720 for the width; the extra 16 samples are horizontal overscan and are not part of the 4:3 or 16:9 picture.
6. 06Switch to the second mode when you know how footage is supposed to look but not what pixel ratio it carries — enter the intended display ratio and the calculator returns the pixel ratio in lowest terms.

The formula.

Think of the frame as a grid of Ws by Hs cells, each cell a rectangle PAR times as wide as it is tall. The grid's own shape is SAR = Ws ÷ Hs. Multiplying every cell's width by PAR multiplies the whole picture's width by PAR too, so the shape you see is DAR = SAR × PAR. That is the entire theory; everything else is rearrangement.

Solving for the pixel shape gives PAR = DAR ÷ SAR, which written out on the raw inputs is (darWidth × Hs) ÷ (darHeight × Ws). Because those four numbers are normally integers, the result is an exact fraction, and reducing it by the greatest common divisor produces a ratio in lowest terms — which is exactly what ITU-T H.264 requires when it says that sar_width and sar_height 'shall be relatively prime'. That is why this calculator can print 16:11 rather than 1.4545:1.

Correcting to square pixels can be done in two directions and both are used in practice. Stretching the width gives Ws × PAR by Hs; shortening the height gives Ws by Hs ÷ PAR. Both produce the same display shape, and the choice is about resampling: widening interpolates new columns, shortening discards rows. Pixel counts must be whole numbers, so both are rounded here.

Working the default through: a 704 × 480 frame has SAR = 704 ÷ 480 = 1.4667, which reduces to 22:15 — not a shape anyone would recognise. Its pixels are 10:11, or 0.9091, slightly taller than wide. The display shape is therefore (22/15) × (10/11) = 220/165 = 4/3, exactly 4:3. Correcting by width gives 704 × 10/11 = 640 exactly, so 640 × 480; correcting by height gives 480 × 11/10 = 528, so 704 × 528. Both are exactly 4:3, and both are whole numbers, which is not a coincidence — the standard ratios were chosen so the arithmetic closes on the resolutions people actually use.

The same procedure reproduces every 'without horizontal overscan' example in H.264's Table E-1: 352 × 288 at 12:11, 352 × 240 at 10:11, 528 × 576 at 16:11 and 528 × 480 at 40:33 all land on exactly 4:3, while 352 × 576 at 32:11, 352 × 480 at 80:33, 528 × 576 at 64:33, 528 × 480 at 160:99, 1440 × 1080 at 4:3, 1280 × 1080 at 3:2 and 960 × 1080 at 2:1 all land on exactly 16:9. Thirteen independent checks on one multiplication.

A worked example.

Frequently asked questions.

References& sources.

1. [1]ITU-T Recommendation H.264 (08/2021), Advanced video coding for generic audiovisual services — § E.2.1 and Table E-1 'Meaning of sample aspect ratio indicator': 1:1 (square), 12:11, 10:11, 16:11, 40:33, 24:11, 20:11, 32:11, 80:33, 18:11, 15:11, 64:33, 160:99, 4:3, 3:2, 2:1 and Extended_SAR, with 'sar_width and sar_height shall be relatively prime'. NOTE 1 defines horizontal overscan and states that the 720x576 4:3 entry refers to the visible 704x576 luma samples.
2. [2]Recommendation ITU-R BT.500-15 (05/2023), Table 1-1 — pixel aspect ratio column: 0.89 for 720 × 483 (ITU-R BT.601, 4:3), 1.07 for 720 × 576 (4:3), and 1 for 640 × 480, 1 280 × 720, 1 920 × 1 080, 3 840 × 2 160 and 7 680 × 4 320. International Telecommunication Union.
3. [3]Recommendation ITU-R BT.709-6 (06/2015), Table 1 — aspect ratio 16:9, 1 920 samples per active line, 1 080 active lines, pixel aspect ratio 1:1 (square pixels). International Telecommunication Union.
4. [4]Recommendation ITU-R BT.2020-2 (10/2015) — 3 840 × 2 160 and 7 680 × 4 320 systems, 16:9 aspect ratio, pixel aspect ratio 1:1 (square pixels). International Telecommunication Union.
5. [5]Recommendation ITU-R BT.601-7 (03/2011), Studio encoding parameters of digital television for standard 4:3 and wide-screen 16:9 aspect ratios — the 720-sample active line that non-square pixel ratios derive from. International Telecommunication Union.