English ▾
Free, Fast & Secure Online Tools
Lips move, sound arrives late, and the video is unwatchable. Before trying anything, answer one question — it splits the problem cleanly in two and tells you which half of this article applies.
Watch the start, then watch ten minutes in. Is the gap the same, or is it bigger?
A gap that stays the same is a constant offset. A gap that grows is drift. They have nothing to do with each other.
The audio and video are both playing at the correct speed; one simply starts at the wrong moment. The causes are all about the path the sound takes, not about the file’s internals.
Bluetooth. This is the most common cause and the easiest to test. Wireless audio has to be encoded, transmitted, buffered and decoded, and that round trip takes a noticeable fraction of a second. Players and televisions compensate, but imperfectly, and the amount varies by codec and device.
The test takes five seconds: plug in wired headphones. If the sync is suddenly perfect, the file was never the problem.
Display processing. Televisions apply motion smoothing, noise reduction and upscaling, all of which delay the picture by a few frames. The sound goes straight out. Most sets have an audio delay setting precisely because of this, and “game mode” usually bypasses the processing entirely.
An offset recorded in the file. A container can specify that an audio track begins slightly before or after the video. Usually this is correct and compensates for something real; occasionally a recording or editing tool writes it wrong, and every player faithfully reproduces the error.
If wired headphones and a different player both show the same offset, the file genuinely carries it, and the fix is to shift the audio track and re-save.
This one is in the file, and it means the audio and video disagree about how fast time passes.
Variable frame rate. The dominant cause. Screen recorders, phone apps and game capture tools often record at a frame rate that changes moment to moment — more frames when a lot is happening, fewer when the picture is still. The audio, meanwhile, runs at a fixed sample rate, because sound has no equivalent of a still moment.
Play that file and most players cope. Edit it, convert it, or upload it somewhere that re-encodes, and the pipeline frequently assumes a constant frame rate — so it places frames at even intervals that were never evenly spaced. By ten minutes in, the picture is seconds away from the sound.
This is why a screen recording looks fine until you cut it, and then the second half is unusable.
Sample rate mismatch. Audio is recorded at 44.1 kHz or 48 kHz. If a tool believes a 48 kHz track is 44.1 kHz, it plays it about 9% too slowly — and 9% of ten minutes is nearly a minute of drift.
Joining clips with different settings. Concatenating a 30 fps clip and a 25 fps clip, or 44.1 kHz and 48 kHz audio, without re-encoding gives you a file whose timing changes partway through. Everything after the join is wrong, and increasingly so.
Almost every drift problem resolves the same way: re-encode with a constant frame rate, 30 or 60 fps, and a consistent audio sample rate of 48 kHz. That forces the file into a uniform timeline instead of one that assumes the reader will interpret variable timing correctly.
Do this before editing, not after. Once an editor has laid variable-rate footage onto a fixed timeline, the drift is baked into the cut and fixing it afterwards means redoing the edit.
Before re-encoding, inspect the file to confirm what you are dealing with — it will show whether the frame rate is variable and what the audio sample rate is. Two minutes of looking prevents re-encoding something that was never the problem.
Rule out the playback path first: wired headphones, a different player, a different device. Only if the offset survives all three is it in the file.
If it is, the fix is to shift the audio by the measured amount. Measure it on something with a sharp transient — a clap, a door closing, a hand hitting a table — where the exact frame is unambiguous. Guessing at “about a third of a second” produces a file that is wrong in the other direction.
Audio that is in sync, drifts out, then snaps back into sync is usually a damaged or incomplete file — a recording that was interrupted, or a download that lost a chunk. The player is skipping, not drifting. Re-encoding often repairs the timeline; if the data is genuinely missing, nothing will.