Fix Audio Sync in Video
Sound arriving before or after the lips move? Load the video, nudge the audio in milliseconds while it plays until speech and picture line up, then save a corrected copy. The picture is copied as-is, so there is no quality loss, and the file never leaves your device.
Last updated:
Green waveform: audio with your shift. Grey outline: where the sound was originally. The orange line is the playhead.
Audio offset
No shift: audio plays as recorded.
Sound comes before the lips move? Use a positive value (delay the audio). Sound comes after? Use a negative value. Keys: ←/→ ±10 ms, Shift ±100 ms, Space play.
Choose a video to start. The demo above shows a flash and a beep so you can feel what an offset does.
ffmpeg -i input.mp4 -itsoffset 0 -i input.mp4 -map 0:v:0 -map 1:a:0 -c copy output.mp4
How to use the Fix Audio Sync in Video
- Choose your video. Drop the file on the box or click to pick it. It is opened locally in your browser.
- Find a clear sync point. Play a moment with a clap, a hard consonant like “p” or “b”, or a door slam. Tick the 4-second loop to replay it.
- Shift the audio. If you hear the sound before you see it, raise the offset (delay the audio). If you hear it late, lower it below zero. Use the ±10 ms or ±1 frame buttons for fine tuning.
- Compare. Toggle “play original timing” to hear the before and after.
- Save. Click Fix sync and save video. The streams are copied into a new file with the corrected offset; download it when it finishes.
Why audio goes out of sync
A fixed audio offset is almost always introduced by the recording chain, not by the file itself. Common causes:
- Separate audio recorder or external mic: the sound file and the camera clip were lined up by eye and are a few frames apart.
- Screen recorders and game capture: video encoding adds latency that the audio path does not, so the sound is ahead of the picture.
- Bluetooth headphones while editing: you synced to delayed monitoring (often 150 to 250 ms), so the exported audio ends up early on other devices.
- Webcams and video calls: USB cameras can deliver frames 100 ms or more after the microphone hears the sound.
- Variable frame rate phone video converted with tools that assume a constant rate: this causes drift that grows over time rather than a fixed offset.
Milliseconds to frames
One frame lasts 1000 ÷ frame rate milliseconds. Use this table to translate a visible offset of a few frames into the value to enter.
| Frame rate | 1 frame | 2 frames | 5 frames | 10 frames |
|---|---|---|---|---|
| 24 fps (film) | 41.7 ms | 83.3 ms | 208 ms | 417 ms |
| 25 fps (PAL) | 40 ms | 80 ms | 200 ms | 400 ms |
| 29.97 / 30 fps | 33.4 / 33.3 ms | 66.7 ms | 167 ms | 334 / 333 ms |
| 50 fps | 20 ms | 40 ms | 100 ms | 200 ms |
| 59.94 / 60 fps | 16.7 ms | 33.4 ms | 83.4 ms | 167 ms |
Tips for finding the exact offset
- Pick a moment with a sharp sound and a clear visual: a hand clap, a slate, a ball bounce, or plosive sounds like “p”, “b” and “m” where the lips close.
- Use the waveform strip: a sharp spike marks the sound, and the grey outline shows where it sat before your shift.
- Get close with ±100 ms, then refine with ±10 ms or ±1 frame. Viewers tolerate late audio better than early audio, so if unsure, err slightly on the late side.
- Check several points (start, middle, end). If the start is fixed but the end is still off, the problem is drift, not a constant delay.
Frequently asked questions
Which direction should I shift the audio?
If the sound is heard before the matching movement (you hear a clap before the hands meet), the audio is early: enter a positive value to delay it. If the sound comes after the movement, the audio is late: enter a negative value so it plays earlier.
How much offset is noticeable?
Broadcast guidelines such as ITU-R BT.1359 found viewers detect audio that leads the picture by about 45 ms and lags it by about 125 ms. Audio that comes early is more noticeable than audio that comes late, so aim for within one frame (about 33 ms at 30 fps).
Does this re-encode or reduce the quality of my video?
No. The default method uses ffmpeg compiled to WebAssembly to remux the file: the video and audio streams are copied bit for bit and only their timing is changed, the same as the desktop command ffmpeg -itsoffset. Only the real-time fallback re-encodes.
What if the audio drifts gradually instead of being off by a fixed amount?
A constant offset fixes delay that is the same from start to end, which is the usual case for cameras, screen recorders and Bluetooth monitoring. Drift that grows over time comes from mismatched sample or frame rates and needs the audio to be time-stretched instead; fix the offset at the start, then check the end of the clip.
Why is the beginning or end of the audio silent or cut after the fix?
Delaying audio inserts a gap at the start, and advancing it removes that many milliseconds from the start of the soundtrack. The video length stays the same.
Is my video uploaded anywhere?
No. The file is read and processed entirely inside your browser. The only download is the free ffmpeg.wasm engine (about 31 MB) the first time you save, which your browser can then reuse.
Which formats work?
MP4, MOV, MKV and WebM are supported. The output keeps the same container as the input. Previewing needs a format your browser can play; H.265/HEVC files may not preview in every browser but can still be saved.
Are all the tools on FreeToolCabin tested for quality and accuracy?
Yes! Every tool on our site undergoes automated Q/A testing in real browser environments to ensure high precision, correct output, and zero errors. If you happen to spot a bug or have a suggestion, please contact us at [email protected] and our team will review it right away.