Audio that drifts out of sync after converting a video is usually caused by variable frame rate (VFR) source footage being forced to a constant frame rate, by a mismatched audio sample rate, or by a start-time offset that the converter dropped. If the delay is the same throughout, it is an offset you can correct with a simple shift; if it gets worse over time, it is a drift problem and needs the timing handled properly during conversion.
This guide helps you identify which kind of sync problem you have, explains the common causes, and gives specific fixes for each, including FFmpeg commands for the most frequent cases.
Step one: offset or drift?
Play the converted file and check sync at the start, middle and end. This single observation tells you most of what you need to know.
| Symptom | Type | Likely causes |
|---|---|---|
| Same delay from start to finish | Constant offset | Start-time mismatch, encoder delay, edit list ignored, audio delay setting |
| Fine at the start, worse over time | Drift | VFR source, wrong frame rate, sample rate mismatch, dropped frames during capture |
| Jumps out of sync at a specific point | Discontinuity | Gap or corruption in the source, joined clips, a broadcast recording glitch |
| Only wrong in one player | Playback issue | Player or hardware decoder handling, not the file |
Before changing anything, check the source in the same player. If the original is also out of sync, conversion did not cause it - although conversion can still fix it.
Cause 1: variable frame rate sources
This is the most common cause of drift. Phones, screen recorders and some game capture software record at a variable frame rate: the frame timing changes depending on load or scene content, and each frame carries its own timestamp. Many video editors and some converters assume a constant rate. If a converter treats a 29.4 fps average as exactly 30 fps, or rebuilds timing without respecting the timestamps, the video slowly runs ahead of or behind the audio.
Check whether your source is VFR with a media info tool; look for "Variable" frame rate mode or a minimum and maximum frame rate that differ. To convert VFR footage to constant frame rate while keeping sync, let FFmpeg duplicate or drop frames based on timestamps:
ffmpeg -i input.mp4 -vf fps=30 -c:v libx264 -crf 20 -c:a copy output.mp4
The fps filter places frames according to their real timestamps, so audio and video stay aligned. Alternatively, keep the output VFR if your target player handles it; most modern players do. Our guide to changing video frame rate covers the difference between these approaches in detail.
Cause 2: wrong frame rate interpretation
Footage at 23.976 fps that is treated as 24 fps will drift by roughly 3.6 seconds per hour. The same happens between 29.97 and 30, or 59.94 and 60. Make sure the output rate matches the source exactly - in FFmpeg, specify fractional rates such as 24000/1001 rather than rounded numbers.
Be careful with -r on the input side. Placed before -i, it overrides the source timing and changes the playback speed of the video without touching the audio, which causes exactly this kind of drift.
Cause 3: audio sample rate problems
Audio drift can also come from the audio side. A recording made at 48 kHz but labelled 44.1 kHz will play roughly 9% too slowly, and at a lower pitch. More subtly, cheap capture hardware sometimes records at a rate slightly off nominal, so the audio is technically 48 kHz but has a few extra or missing samples per second. FFmpeg can stretch or squeeze audio to fit its timestamps:
ffmpeg -i input.mkv -c:v copy -af aresample=async=1000 -c:a aac -b:a 192k output.mkv
The aresample=async option corrects audio timing against the timestamps by adding or removing samples, which fixes slow drift and small gaps.
Cause 4: start-time offsets and edit lists
A constant offset often comes from streams that start at different times in the source. MP4 and MOV files can include edit lists that tell the player to skip the first few milliseconds of audio (encoders like AAC add a short priming delay), and transport streams from TV recordings frequently start audio and video at different timestamps. If a converter ignores this information, the whole output is shifted.
For a constant offset, the quickest fix is to shift one stream without re-encoding. This delays the audio by 0.3 seconds:
ffmpeg -i input.mp4 -itsoffset 0.3 -i input.mp4 -map 0:v -map 1:a -c copy output.mp4
To move the audio earlier instead, put -itsoffset before the first input (the one used for video). Find the right value by testing: watch a clear event such as a clap, a door closing or lip movement on a hard consonant.
Cause 5: joined clips and damaged sources
When several clips are concatenated, small differences in audio length between clips accumulate, so sync gets progressively worse at each join. Re-encoding the audio during concatenation, rather than stream copying, usually fixes this. Recordings with dropouts - from a weak broadcast signal or a dropped network stream - can jump out of sync at the damaged point; re-encoding with aresample=async or repairing the gap before conversion helps.
When it is the player, not the file
If the file is in sync on one device and not another, the file is probably fine. Bluetooth headphones add noticeable latency; some TVs process video more slowly than audio; hardware decoders occasionally mishandle unusual files. Test on a computer with wired audio before re-encoding. If the file will not play at all on some devices, see what to do when a converted video won't play.
Preventing sync problems
- Check the source frame rate mode before converting, and handle VFR deliberately.
- Keep the original frame rate unless you have a reason to change it.
- Avoid input-side
-runless you intend to change playback speed. - Test a full-length encode at the end, not just the first minute - drift hides early on.
- When remuxing, consider whether the new container preserves the start offsets; our guide to codecs and containers explains why remuxing can behave differently from re-encoding.
Doing it with Movie Encoder
Movie Encoder's media info view shows the frame rate, codecs, sample rates and stream details of every source, so you can spot VFR footage and mismatches before converting. You can choose constant or variable frame rate output, convert frame rates, set the audio sample rate and apply audio delay correction to fix a constant offset, with the command preview showing the exact FFmpeg arguments.
Short sample encodes let you check sync on a section before committing to the whole file, and the original is never overwritten, so you can try a correction and compare without risk.
Frequently asked questions
Why does audio drift further out of sync as the video plays?
Gradual drift usually comes from variable frame rate footage converted to a constant rate without respecting timestamps, or from a frame rate rounded from 29.97 to 30. A constant delay, by contrast, is a simple offset.
How do I fix a constant audio delay without re-encoding?
Use FFmpeg's -itsoffset option to shift the audio stream and copy both streams unchanged. Test the value against a clear sound such as a clap.
Could the problem be my player rather than the file?
Yes. Bluetooth headphones and some TVs add latency, so test the file on a computer with wired audio before re-encoding.
