How to reduce video file size

Video2 min read

The four levers, in the order that actually saves the most, and how far to push each before it starts to show.

In short

Trim first — it costs nothing. Then cap the resolution, which is the biggest quality-preserving saving. Then compress. Lowering the quality preset is the last resort, not the first.

Most people reach straight for a compressor and push the quality down until the file fits. That works, and it is usually the worst-looking way to get there. Three other levers cost less.

In order of what to try first

  1. 01

    Cut what nobody needs to see

    Duration is linear: remove a third of the clip and you remove a third of the file, at zero quality cost. Dead air at the start, the walk back to the camera, the bit after the point was made. Use Trim Video.

  2. 02

    Cap the resolution

    1080p to 720p removes about 55% of the pixels in every frame. On a phone screen or in a message window the difference is very hard to see, and it beats any quality setting for the size it saves.

  3. 03

    Compress at Balanced

    Re-encode once. Balanced is right for almost everything; it is the setting where most people cannot tell the export from the source at normal viewing size.

  4. 04

    Only then drop the quality preset

    Smallest file is for hitting a hard limit. It softens fast motion and fine texture noticeably, which is why it comes last.

What each step is worth

Roughly what each lever removes
LeverTypical savingQuality cost
Trimming 30% of the length~30%None
1080p → 720pRemoves ~55% of pixels per frameInvisible on a phone; visible on a TV
1080p → 480pRemoves ~80% of pixels per frameClearly softer; fine for a preview or a message
One quality preset harderRoughly halves the bitrateFast motion and texture soften first
Muting the audio~1MB per minuteNone to the picture — but you lose the sound

Mistakes worth avoiding

  • Compressing before trimming. You spend minutes encoding footage you are about to delete.
  • Running the compressor twice. A second pass on your own output stacks the loss for very little gain — go back to the original and use a harder setting once.
  • Compressing hard before uploading to a platform that re-encodes anyway. You are handing it a damaged source.