How to find the aspect ratio of a video (and calculate it)

To find the aspect ratio of a video, get its width and height in pixels and divide both by their greatest common divisor: 1920×1080 reduces to 16:9, 1080×1920 to 9:16, and 1080×1350 to 4:5. If the result isn't a familiar pair, divide the width by the height and compare decimals instead: 16:9 is 1.78.

The ffprobe and FFmpeg facts come from the ffprobe, filters, and ffmpeg documentation, one size from YouTube's Help page Video resolution & aspect ratios, and the Sume facts from the Video inspect docs and the Sume API reference, all read on 2026-09-28. The ratios are our arithmetic, and anything described as current behavior is read from Sume's code.

How do I calculate the aspect ratio from width and height?

The greatest common divisor (GCD) is the largest whole number that divides both. For 1920 and 1080 it is 120, so the ratio is (1920 ÷ 120):(1080 ÷ 120) = 16:9. To find the GCD by hand, divide the larger number by the smaller and keep the remainder, then repeat with the smaller number and that remainder until nothing is left: 1920 ÷ 1080 leaves 840, 1080 ÷ 840 leaves 240, 840 ÷ 240 leaves 120, 240 ÷ 120 leaves 0. The last number you divided by is the GCD.

The same method as a small calculator, in Python with no imports:

def aspect_ratio(width, height):
    a, b = width, height
    while b:
        a, b = b, a % b  # a ends as the GCD
    return f"{width // a}:{height // a}"

print(aspect_ratio(1920, 1080))  # 16:9
print(aspect_ratio(1080, 1350))  # 4:5

What are the aspect ratios of common video sizes?

Some sizes reduce to pairs nobody uses, and then the decimal is the useful answer. YouTube lists 854×480 among its 16:9 sizes, though it reduces to 427:240: as a decimal it is 1.78, the same as 16:9 to two places. A 1920×804 frame, the 2.39:1 crop from cinematic bars size, reduces to 160:67, or 2.39 as a decimal.

How do I find a video's width and height?

For a clip already on Sume, an unbilled probe gives them: POST /v1/video-inspect with frames: false returns probe.width, probe.height, and probe.rotation. The clip must be your workspace's media.sume.com artifact or asset; get a video's duration, resolution, and fps shows the request.

For any other file, FFmpeg's ffprobe gathers information from media streams and prints it. -select_streams v:0 picks the first video stream, -show_entries stream=width,height limits the output to those two entries, and the csv writer with s=x joins them with an x while p=0 drops the section name:

ffprobe -v error -select_streams v:0 \
  -show_entries stream=width,height -of csv=s=x:p=0 input.mp4
# 1920x1080

Why doesn't the ratio match what I see?

Rotation and non-square pixels can both make a video display in a different shape from its stored pixel size.

• Rotation. A file can carry a rotation value beside its stored width and height, and in current code Sume's probe reads it from the stream's side data or rotate tag. With a quarter turn (90, −90, or 270), swap width and height before you reduce them.
• Non-square pixels. FFmpeg's docs give the display aspect ratio as width ÷ height × the sample (pixel) aspect ratio, SAR. With square pixels (SAR 1:1) that is just width:height; otherwise the frame displays wider or narrower than its pixel counts. A 1440×1080 frame with 4:3 pixels displays at 1440 ÷ 1080 × 4/3 = 16:9.
• A file can also declare its display ratio: FFmpeg's -aspect option takes values such as 16:9 or 1.7777, and with stream copy it changes the ratio stored at container level.
• Sume's probe has no sample or display aspect ratio field, so it can't flag a file with non-square pixels. Its width and height are the stored size.