How to use Change Video Speed
- 1Drop in a video file
- 2Choose a speed multiplier — 2 for double, 0.5 for half
- 3Both picture and sound are adjusted together
- 4Pick a quality level for the re-encode
- 5Download the retimed MP4
What Change Video Speed does
Changing playback speed means doing two different things to two different streams. The picture is retimed by rewriting each frame's presentation timestamp — the frames are unchanged, they are simply told to appear sooner or later. That is cheap and exact.
Audio cannot be handled that way. Playing samples faster raises their pitch, which is why a tape played at double speed sounds like a chipmunk. Keeping the pitch while changing the tempo requires resampling the sound properly, and the filter that does it only accepts a factor between 0.5 and 2. Anything outside that range is expressed as a chain of in-range steps — quadruple speed becomes two doublings.
When to use it
- Watching a lecture faster
- 1.5× or 2× makes a recorded talk considerably quicker to get through, and exporting at that speed means it plays that way in any player rather than depending on playback controls.
- Time-lapse from a long recording
- A high multiplier compresses hours into minutes. At those speeds the audio usually stops being useful and is best removed.
- Slow motion for detail
- 0.5× or 0.25× makes a fast movement examinable. Note that no new frames are created, so very slow speeds look stepped rather than smooth.
- Fitting a fixed time slot
- A clip slightly too long for a slot can be nudged to 1.1× or 1.2×, which is barely perceptible and avoids cutting content.
Change Video Speed settings explained
- Speed multiplier
- Values above 1 speed the video up, below 1 slow it down. 2 halves the duration, 0.5 doubles it. The picture and the sound are both adjusted by the same factor so they stay in sync.
- Pitch is preserved
- The audio is retimed rather than resampled, so voices keep their original pitch at any speed. This is what makes a lecture at 1.5× still sound like the person speaking.
- Chained tempo steps
- The audio filter only accepts factors between 0.5 and 2. Larger changes are applied as a sequence — 4× becomes two 2× steps — which is handled automatically but does mean more processing for extreme values.
- Quality level
- High (CRF 22), Balanced (CRF 27) or Smallest file (CRF 32) for the re-encode. The picture has to be re-encoded because the frame timing changes, so this is not a lossless operation.
What happens at each speed
| Speed | A 10-minute video becomes | Notes |
|---|---|---|
| 0.25× | 40 minutes | Visibly stepped — no new frames are created |
| 0.5× | 20 minutes | Usable slow motion for most footage |
| 1.5× | 6m 40s | Speech stays clear and natural |
| 2× | 5 minutes | Speech still followable; one audio step |
| 4× | 2m 30s | Audio rarely useful; two chained steps |
Limits and what to expect
- Slowing down does not create new frames, so low multipliers look stepped rather than smooth.
- The video is always re-encoded, because frame timing changes — this is not a lossless operation.
- Audio beyond about 2× or below 0.5× degrades noticeably even with chained steps.
- Output is always MP4 with H.264, whatever the input format.
- Encoding is single-threaded, so long videos take minutes.
Troubleshooting
- Slow motion looks jerky
- No frames are being invented — a 30 fps clip at 0.25× is still 30 frames spread across four times the duration, so each is held four times as long. Genuine smooth slow motion requires footage shot at a high frame rate to begin with.
- The audio sounds odd at high speed
- Beyond about 2×, tempo adjustment starts to produce artefacts even with the chained steps. For time-lapse work the sound is rarely wanted anyway — remove it with the mute tool and speed up the silent video.
- Audio and video have drifted apart
- Both streams are retimed by the same factor, so they should stay aligned. Drift usually means the source had a variable frame rate, common in screen recordings. Convert it to a constant frame rate first.
- It is taking a long time
- The picture is fully re-encoded because the timing changes, so this costs as much as compressing. Trim to the section you need first — trimming is a stream copy and takes seconds.
Your files and your privacy
Retiming and encoding both run through an ffmpeg build compiled to WebAssembly inside your browser tab. The video never leaves your device, and nothing is queued on a server. The only network request is for the engine itself.
What to use next
FAQ
Does changing the speed affect the pitch?
No. The audio is retimed rather than simply resampled, so voices keep their original pitch at any multiplier. That is what distinguishes this from a tape played fast, where the pitch rises with the tempo.
How do I make a time-lapse?
Use a high multiplier such as 4× or 8×. The audio becomes useless well before that, so mute the video first and then speed up the silent file — it also encodes faster with nothing to retime.
Why does slow motion look jerky?
Because no new frames are created. A 30 fps clip at 0.25× still has 30 frames per original second, each now displayed four times as long. Smooth slow motion requires footage shot at a high frame rate in the first place.
Does this reduce quality?
Yes, some. Changing frame timing means the video has to be re-encoded, which adds a generation of compression loss. Choose High quality if that matters; there is no lossless way to retime a video.
What speed should I use for a lecture?
1.5× is the usual sweet spot — noticeably quicker while speech stays entirely natural. 2× is followable for a clear speaker. Beyond that, comprehension drops faster than the time saved is worth.
Is my video uploaded?
No. All processing happens in your browser through WebAssembly, on your own device, with nothing transmitted.