How much can you compress a video?

Measure the saving rather than assuming a percentage

Saving (%) = (original bytes − output bytes) ÷ original bytes × 100. If a 200 MB source becomes 50 MB, the saving is 75%; the result is one quarter of the original size. If a 20 MB source becomes 24 MB, it is 20% larger. These examples use decimal MB consistently.

A high-rate recording may leave room for a smaller delivery copy, while a heavily compressed source may leave little useful room. Motion, texture, noise, codec, resolution and encoder behaviour all matter. A static-looking tutorial can still need enough detail for tiny text. No category guarantees a 50% or 90% reduction.

Calculate a size budget from bitrate and time

Approximate media size in MB = combined average bitrate in Mbps × duration in seconds ÷ 8. Include audio as well as video, then allow for the container. Decimal units mean 1 MB = 1,000,000 bytes; some operating systems display binary units instead.

Resolution alone does not determine file size. Two 1080p clips can have different durations, rates and content, and very different sizes. A target bitrate is an encoder request rather than a guarantee that every second uses that exact amount of data.

Worked example: fitting a two-minute clip into 25 MB

A 25 MB decimal budget over 120 seconds corresponds to 25 × 8 ÷ 120 ≈ 1.667 Mbps in total. This tool initially reserves 4% for container overhead, leaving about 1.600 Mbps for the encoded streams. With a 96 kbps audio budget, about 1.504 Mbps remains for video.

That calculation explains a budget, not a guaranteed acceptable resolution. Fast action and small tutorial text can be difficult at the same rate that suits a simpler scene. If the limit becomes particularly tight, the implementation may reduce its audio budget or reject the request. Auto can reduce picture dimensions in size mode; a manually selected resolution limits that automatic choice.

The tool checks the completed byte count and can make up to two further encoding attempts if it overshoots. It can still report Target not met. Read that result and inspect the actual file size before sending; progress reaching 100% does not establish that the destination will accept the file.

Why a result can be larger than its source

The quality modes use dimensions, an estimated frame rate and bitrate rules, with an attempted cap based on the source’s average file rate. Minimum bitrate settings, audio choices, container overhead and actual encoder behaviour mean the output is not guaranteed to shrink.

Do not accept a larger output merely because it was called compression. Compare its purpose and quality with the source. Try one lower setting if appropriate, or keep the original when further loss provides no useful benefit. There is no universal MB-per-minute table for a given resolution or mode.

Leave room for the destination’s own rules

Check the recipient service’s current byte limit, supported format and any account-specific restrictions. A file under the limit can still be rejected for duration or codec. Email systems can also count attachment encoding and the rest of the message, so a nominal attachment budget is not always the raw-file allowance.

Choose a target below a strict published limit where practical and verify the finished file in that workflow. This page does not promise current limits for any named app. If you need less duration, trim in a separate editor; the current compressor processes the chosen clip without a trimming control.