Change the speed of a video

This is the one thing to know before using it. The sound is sped up with the picture, so at 2x voices rise in pitch and at 0.5x they drop — the same effect as playing a record at the wrong speed. It is not a fault; it is what happens when you play samples faster.

Keeping the pitch while changing the speed is a different operation called time-stretching, and it is what a video player does when you watch at 1.5x. It works by cutting the sound into fragments and overlapping them, which is real signal processing rather than a setting, and it is not offered here.

So this is right for a timelapse, a slowed clip you want to look at closely, or anything without speech. For a lecture you intend to listen to, changing the speed in the player is the better answer and costs nothing.

Every frame in a video carries a timestamp saying when it should appear. Changing the speed here means dividing those timestamps by the multiplier and writing the frames back untouched. No decoder runs, no encoder runs, and no pixel changes.

That is why a twenty-minute file finishes in seconds where resizing the same file would take minutes, and why the result is bit-for-bit as good as the original. It also means the file size barely changes: the same frames are still in there, just labelled differently.

The arithmetic is the obvious one. A three-minute clip at 2x is ninety seconds; at 0.5x it is six minutes. Slowing down does not add frames — the same frames are simply held for longer, so a heavily slowed clip looks stuttery rather than smooth.

Genuine slow motion needs frames that were recorded, which is why phones offer a high frame rate mode before you film rather than after. Slowing a normal recording is useful for seeing what happened, not for making it look cinematic.

Nothing is added and nothing is thrown away. The timestamps on the existing frames are rewritten so they arrive closer together or further apart, which is why this finishes almost instantly and why every frame the camera recorded is still in the file.

The consequence is arithmetic. A clip at 30 frames a second played at half speed becomes a clip at 15 frames a second, and 15 is where motion starts to look like a slideshow rather than movement. Speeding up has the opposite and harmless effect: the same frames arrive faster, and the result is smoother than the original, not rougher.

A camera records each frame with a small amount of motion blur, matched to how long that frame is on screen. Slow the video down and the blur is still there but the frame now lingers, so fast motion reads as smeared rather than as slow.

This is why proper slow motion is filmed at a high frame rate rather than slowed afterwards, and why a phone's slow-motion mode switches the camera to 120 or 240 frames a second before you press record. Halving the speed of ordinary footage is usually fine; quartering it rarely is, and no amount of processing recovers frames that were never captured.

Because the audio is stretched with the picture, speeding a clip up raises every voice and slowing it down lowers them. None of the four settings here is gentle about it: 2 and 0.5 move speech by a full octave, and 4 and 0.25 by two, which is well past following.

Pitch-corrected speed is a different operation — it resynthesises the audio, which takes real time and is not lossless. When the sound does not matter, muting the result is the cleaner answer, and when it does, the speed change probably belongs in an editor rather than here.