How to compress a video on Windows 11 for the size you actually need

Choose the control that matches the job

Constant quality lets the encoder spend different amounts of data on simple and complex scenes. It is useful when you care most about appearance, but it cannot promise a fixed file size. Average bitrate suits a workflow that specifies a rate. Target size is for a file budget.

Resolution, frame rate, codec and audio also matter. Removing detail or frames can reduce data needs, but do not lower them automatically when the recipient needs the original recording characteristics.

Build a smaller delivery file

Keep a source copy and test one representative recording first.

1. Add the video and select an output the destination can play, such as MP4 · H.264.
2. Open Output settings and choose quality, average video bitrate or target size. For size mode, enter your budget in MiB.
3. Review audio bitrate and selected tracks; they share the file budget with the picture.
4. Choose a lower resolution only when it suits the use case. Enable two-pass for a supported software encoder when accurate bitrate allocation matters.
5. Create an encoded sample, compare moving detail and sound, then Convert. Inspect the completed file size and playback.

Leave space for the container and sound

A video file contains more than picture data. Soundtracks, subtitles, metadata and container overhead all use space. Target-size jobs account for the output plan and check the completed file, with a default acceptance tolerance around the requested target.

If an external service has a hard upload cap, choose a target below that cap rather than exactly on it. Check the service’s current limit yourself; a preset name is a convenient starting point, not a promise about an external platform’s rules.

Know when software encoding is useful

Available GPU encoders can help on supported hardware, but two-pass size-oriented processing uses supported software paths. The best choice depends on the encoder, settings and source. The app detects actual hardware encoder availability rather than assuming that every codec works on every GPU.

Avoid the common compression traps

Repeatedly encoding already compressed copies can compound damage. Make alternative test outputs from the original instead. Increasing bitrate cannot recover details that the source never contained. HEVC or AV1 can be useful for suitable players, but an incompatible smaller file does not solve a sharing problem.