Pixel aspect ratio calculator

How to use the pixel aspect ratio calculator

1. Enter the stored frame width and height from the video file or select an interpretation preset.
2. If you know the pixel shape, choose “Display aspect ratio” and enter PAR. If you know the intended display shape, choose “Pixel aspect ratio” and enter DAR.
3. Read the simplified fraction and decimal result, then compare the stored and displayed frame diagrams.
4. Choose which dimension to preserve for a square-pixel export. Copy the measurements or the generated FFmpeg command.

Use the dimensions of the frame being interpreted, after any intentional crop. This is a planning tool: it does not inspect metadata, upload video, crop a file or detect whether an editor has already corrected the pixel shape. Entering a target DAR describes your intended interpretation; it does not prove that interpretation is correct for the footage.

PAR vs SAR vs DAR: what is the difference?

In FFmpeg, SAR means sample aspect ratio, also called pixel aspect ratio. Some other discussions use SAR to mean storage aspect ratio, which is W:H instead. Check the definition before copying a value between tools. This page labels the storage quantity explicitly to avoid that ambiguity.

A PAR of 1:1 means square pixels. Values above 1 make a pixel display wider than it is tall; values below 1 make it narrower. This is unrelated to a screen's device pixel ratio, which compares device pixels with CSS pixels, or to print DPI.

Pixel aspect ratio and display aspect ratio formulas

For 1440 × 1080 intended to display as 16:9, the required PAR is (16 ÷ 9) × (1080 ÷ 1440) = 4:3. To produce square pixels while preserving height, the equivalent raster is 1920 × 1080. Keeping width instead produces 1440 × 810. Both share the same display shape, but they retain different amounts of horizontal and vertical sampling.

Fractions are reduced using integer arithmetic, so 4:3 stays exact. Typing 1.333333 means exactly that six-place decimal, not precisely 4/3. The decimal output is rounded for readability; use the fraction when transferring a known rational PAR into another application.

720×480, 720×576 and 1440×1080 examples

Why do NTSC and PAL PAR tables sometimes disagree?

Legacy video workflows may distinguish the stored raster from its intended active image area. A 720-pixel line and a 704-pixel region cannot both map to the same display width using the same PAR. For example, 704 × 480 with PAR 10:11 displays as exactly 4:3; applying 10:11 across all 720 pixels gives 15:11 instead. That arithmetic difference does not, by itself, mean the source metadata is wrong.

Our 720 presets deliberately map the full raster to 4:3 or 16:9. They are not blanket replacements for an editor's DV, D1 or broadcast presets. Check the file's PAR, capture specification, clean aperture and any existing side padding before changing interpretation. Selecting the 704 preset here only changes calculator inputs; it does not remove the outer pixels from a real 720-pixel recording.

Convert non-square pixels to square pixels

To preserve height, calculate output width = stored height × DAR. To preserve width, calculate output height = stored width ÷ DAR. A square-pixel export makes the raster's W:H match the intended display ratio, so playback no longer depends on a non-square PAR being honored.

Real raster dimensions must be integers. This tool rounds both axes to the nearest positive even pixel for the generated H.264 4:2:0 command and reports the ratio difference. Thus 853.333… × 480 becomes 854 × 480, a small change from exactly 16:9. If exact 16:9 is required, choose a proportional delivery raster such as 1280 × 720 using the video aspect ratio calculator. Very small inputs can have large rounding errors; odd preserved dimensions may also move by one pixel.

Metadata interpretation is not resampling

Changing PAR changes how existing samples should be displayed; it does not create a new raster. Resizing changes the raster itself. Simply forcing PAR to 1:1 without the corresponding resize can stretch or squash the image. Conversely, do not apply a second correction if your editor has already normalized the footage.

Using the FFmpeg square-pixel command

The generated command resamples to the displayed export dimensions, sets sample aspect ratio to 1:1, and encodes H.264 video with AAC audio. Replace the filenames, use a new output file and test a short clip first. CRF 18 is a quality starting point, not lossless output. Encoder, player and platform limits still apply even when a mathematical size is valid.

The command assumes the video raster matches the dimensions entered here, with no additional crop or rotation required. It is intended for progressive footage. Interlaced archival video needs an appropriate field-aware or deinterlacing workflow before resizing; this calculator does not choose field order or deinterlace for you. The basic command is not designed to preserve every subtitle, audio track, color-management setting or archival metadata field.

For a read-only metadata check, run ffprobe -v error -select_streams v:0 -show_entries stream=width,height,sample_aspect_ratio,display_aspect_ratio -of default=noprint_wrappers=1 input.mp4. An unknown or missing SAR is not proof that the intended pixels are square; consult the source format when that information is absent.

Related tools for the next step

Technical references

For filter behavior, see the official FFmpeg setdar and setsar documentation and scale filter reference. For editing context, see Adobe's sequence aspect ratio guide. Preset results in the table are calculations from the explicitly stated raster and target DAR.