Change a video’s frame size. Pick a resize mode, give it a number, and export. The output keeps the source container, so an MP4 stays an MP4.
Resize modes
| Mode | What you give it | What comes out |
|---|---|---|
| Fixed width | one width | that width, with the height taken from the source ratio |
| Scale (%) | a percentage | both sides scaled by it |
| Fit inside (W×H) | a width and height | the largest version of the picture that fits inside them, ratio intact |
| Cover (W×H) | a width and height | exactly those numbers, with the overflow cut from the center |
The first three never change the shape of the picture, so nothing is cut and nothing is padded. Only Cover does, which is the trade for getting an exact frame.
Fit inside against Cover
These two take the same pair of numbers and treat them differently.
Fit inside reads them as a limit. A 1920x1080 clip told to fit inside 1080x1080 comes out 1080x608, because that is the biggest it can be without breaking out of the box. It also refuses to upscale: a smaller source is left alone.
Cover reads them as the answer. The same clip told to cover 1080x1080 comes out exactly 1080x1080, with 420 pixels cut from each side.
Use Fit inside when a platform gives you a ceiling. Use Cover when it gives you a spec.
What Cover cuts
The kept area is marked on the preview while you set it up, so the loss is visible before you export. It is always centered, which suits a subject in the middle of the frame and not much else.
Going from 16:9 to 9:16 is the expensive case: the tallest 9:16 slice of a 1920x1080 frame is 608 pixels wide, so 68% of the width goes. To choose which part survives, crop the video first and resize the result.
Upscaling
Fixed width, Scale and Cover will all make a video larger if the numbers say so. The encoder interpolates between the pixels that exist, so no detail comes back, the file grows with the pixel count, and compression artifacts in the source grow with it.
It is worth doing when something rejects anything under a fixed size. It is not worth doing to improve the footage. Fit inside sidesteps the question by never scaling up.
Quality and audio
Changing the frame size rules out a stream copy, so every frame is decoded, drawn at the new size, and encoded again. That adds one generation of compression at a high bitrate. Resizing once is close to invisible; resizing the same file repeatedly is not.
The soundtrack is copied packet for packet when the output container accepts its codec, and re-encoded when it does not. Remove audio drops the track entirely, which is worth doing for a silent background loop.
To change container or codec, run the result through the video converter. When the target is a file size rather than a frame size, the compressor takes that number directly.