A looping video can sound in sync at the start and wrong later, or it can be consistently early or late from the first moment. To tell which problem you have, test the same file in OBS locally and in a private or unlisted YouTube stream, checking at the beginning, near the end and just after the loop restarts.
Start with an OBS Media Source, enable Loop, and confirm that its audio reaches the mixer track you intend to send. Do not add a timing offset until you know whether the mismatch is fixed, accumulating or specific to the loop boundary; those symptoms point to different checks.
Add and loop the OBS Media Source
Before opening OBS, play the original file in a local media player from beginning to end. Listen at the opening, near the end and across the transition back to the start. If the audio begins late, ends early, or has a pause or abrupt cut at that transition in the file itself, OBS will reproduce that content problem when it loops. A playback setting cannot make an imperfect edit seamless.
In OBS, add a Media Source to the scene that will carry the video. Select the local file, then enable Loop in the source properties. The OBS Media Sources guide describes Media Source as a way to play local media and explains that looping restarts the file after playback completes. OBS labels may vary between releases, so check the current source properties if the wording differs.
Once the source is added, confirm that playback is advancing and that the picture is the one you expect to broadcast. Then let it reach the end rather than seeking straight to the loop: the restart itself is part of the test. For a channel built around a single devotional set or ambience video, a Media Source is the direct path. If you are playing a playlist instead, OBS also documents a VLC Video source with a loop-playlist option, which depends on VLC being installed. Choose the source type that matches your material rather than changing tools to chase a sync symptom.
A 4K60 file puts more work through decoding and encoding than a smaller or lower-frame-rate file. That does not by itself mean that video resolution caused an audio problem, but a machine under load is worth watching during a long test. If you are deciding how a continuous music channel should be arranged beyond this one scene, the practical setup points in setting up an always-on Indian music channel are useful context; they do not replace this file-specific sync check.
Route the source audio to the intended mixer track
A Media Source can provide both picture and sound, but the sound still needs to be present in the mix that OBS sends to the stream. Start playback and look at the OBS Audio Mixer. The Media Source should show activity when the file contains audible sound. Check that it is not muted and that its fader is not turned down so far that it is effectively absent.
Open the source’s audio properties and check its monitoring and output behaviour against your scene. If you use advanced audio properties, verify that the source is assigned to an active audio track that is included in the stream’s output. A common false diagnosis is to hear a local monitoring path while the stream is receiving a different track, or to see a moving meter that is not routed to the live mix. The OBS Audio Mixer guide explains mixer controls and track routing; the Audio Sources guide covers how OBS handles sources that provide audio.
Use headphones during the preflight if practical. Listening through speakers can create an apparent delay because you hear monitoring, room sound and the video display through different paths. This is a listening issue, not proof that the transmitted audio is offset. Keep the test listening arrangement the same when comparing recordings and the received stream.
If other scene audio is also active, temporarily make the test scene simple enough to identify the Media Source clearly. You are not trying to redesign the full channel mix; you are isolating whether this source’s sound is reaching the expected output. Once the route is confirmed, restore the normal scene and make a second test there, because added sources and processing may change what you hear.
Record locally and test a private or unlisted stream
A local recording and a YouTube test stream answer related but different questions. The local recording shows what OBS is producing with the chosen scene, frame rate, encoder and audio route. The private or unlisted stream shows what happens after YouTube receives and plays that output. If the recording is already out of sync, investigate OBS and the file before treating YouTube as the cause. If the recording is aligned but the received stream is not, compare the stream output and playback conditions as well.
Use the same file, scene, output frame rate, encoder and audio path that you plan to use for the real broadcast. Record long enough to include the last part of the file and its restart. Then create a representative private or unlisted test in YouTube Live Control Room. YouTube’s encoder settings guidance recommends testing with audio and movement similar to the planned stream, and monitoring stream health during an event. A short opening-only test will not show a problem that appears after a long playback or on the loop.
Write down what you notice rather than relying on memory: whether the voice or beat is early or late, whether the amount seems unchanged, and whether anything changes at the restart. Use the same identifiable cue at each check, such as a drum hit, a hand clap or a spoken consonant next to a visible movement. Watching a visual event with no clear audio transient makes small differences hard to compare.
The test should also use the intended stream configuration. YouTube’s published recommendation for H.264 at 4K/2160p and 60 fps is a 35 Mbps video bitrate, with CBR and a recommended two-second keyframe interval; YouTube says not to exceed four seconds. Those are YouTube ingest recommendations, not evidence that your computer or connection can sustain them. The page also lists AAC or MP3 audio and recommends 44.1 kHz for stereo. Check the current official guidance before relying on settings, as these recommendations can change.
For 4K, YouTube says the “improve for low latency” option is unavailable and the stream is optimised for normal latency. Do not plan a 4K60 workflow on the assumption that viewers will receive it with low latency. If you are comparing encoders or file formats, the discussion of H.264 and HEVC for 24/7 YouTube loops can help frame that choice, but codec choice is not a substitute for measuring sync in your own output.
Compare sync at three points, including the loop
Use three checkpoints in both the local recording and the received test stream:
| Checkpoint | What to compare | What it can reveal |
|---|---|---|
| Start of playback | A visible action against its matching sound | A file or source that begins with a fixed offset |
| Near the end | The same kind of audio/visual cue later in the run | Whether the mismatch has grown during playback |
| Just after the loop | A cue after the file restarts, compared with the first pass | A discontinuity or restart-specific change |
For the start check, choose a cue close to the file’s opening, but allow playback to settle if the source takes a moment to start. Compare the sound with the image in the recording first. Then check the same cue on the YouTube test stream. You are not looking for laboratory precision from a casual listen; you are looking for a repeatable difference you can classify.
Near the end, compare another identifiable cue with the opening observation. A bhajan vocal entering on a visible title change, a beat beside a repeating animation, or a person speaking while their mouth is visible can work. If the file contains no clear cue near its end, note that limitation and use the closest reliable event rather than pretending a vague impression is a measurement.
Immediately after the loop point, compare the restarted passage with the first passage. Listen for a gap, overlap, clipped first sound or a new offset. A stream that is aligned through the first pass but jumps on each restart is not the same problem as a steady delay throughout. You can also compare the sound at the file’s final moment with its first moment; a deliberate fade or pause in the content may be audible even when the timing machinery is functioning as designed.
Repeat the comparison on the local recording and the received stream. Keep the same scene and output settings for both. If you hear a mismatch only while monitoring live in OBS but cannot find it in the recording or stream, check the monitoring path and listening setup before adjusting the source. The distinctions matter because an adjustment that improves one symptom may worsen another.
Tell a fixed offset from accumulating drift
A fixed offset means the same sound seems early or late by about the same amount at the beginning and near the end. For example, if a visible clap and its sound appear separated from the first checkpoint by roughly the same interval at the final checkpoint, you have a candidate for a constant timing adjustment. Record whether audio is early or late, and avoid applying a value to every source when only the Media Source is affected.
Accumulating drift is different: the alignment gets progressively worse as playback continues. A small mismatch at the start may become plainly noticeable near the end. Adding a fixed delay cannot correct a difference that keeps changing. Re-check the original file, source timing and audio path; also watch for decoding pressure and inconsistent audio handling. The available official guidance does not establish one universal cause or corrective value for every file and computer.
A third pattern is an error that appears at the loop boundary and then remains or repeats on each pass. Check the final and first moments of the source file, including their audio duration and intended edit. A restart can expose a discontinuity that is part of the content or playback behaviour; a fixed offset applied to the whole source cannot remove a one-time jump at the boundary.
Network or encoder symptoms are worth noting but should not be confused with proof of audio drift. OBS’s stream connection troubleshooting guide describes dropped frames as a connection or bitrate-sustainability issue. If OBS reports dropped frames or YouTube reports unstable stream health, address that separately and repeat the sync test. The warning alone does not show that network instability caused the audio mismatch.
For a long-running channel, the same discipline matters after preflight. A system that worked for a short test may behave differently during a full night if its workload, connection or source changes. Advice on restarting a looping ASMR stream automatically concerns recovery from interruption, not audio timing, but it is relevant when you are planning what to check during an always-on broadcast.
Adjust only the affected source, then retest
If the evidence points to a stable offset, adjust the timing of the affected audio source in OBS in small increments. Depending on how the source is presented in your version of OBS, this may mean setting a sync offset in Advanced Audio Properties. Confirm first which audio source you are changing; do not add delay to a microphone, music bed or the entire mix when the Media Source alone is out of step. Make one change at a time so you can tell whether it helped.
Repeat the complete test after an adjustment: record locally, send a private or unlisted stream, and compare at the start, near the end and after the loop. A change that aligns the opening but makes the end worse has not solved accumulating drift. A change that aligns both sides of the loop but leaves a jump at the restart has not repaired a boundary discontinuity. Return to the original value if the result is less consistent rather than stacking more changes on top of it.
If the mismatch grows, inspect the source before trying a succession of offsets. Play the file through in a local player, check whether the final audio and video durations line up, and verify that the audio sample rate and output path are consistent. Review OBS load indicators and the computer’s performance while the 4K60 file is playing. If the source itself is malformed or has a poorly edited ending, replace or prepare a corrected file and test it; OBS’s Loop setting only repeats playback.
If the local recording looks aligned but the received stream does not, retain both examples and check YouTube’s stream health messages, output settings and playback conditions. If both are aligned but viewers report a difference, ask them to compare another device or browser before changing your production timing. Viewer playback buffering can affect what an individual hears, and a report from one playback path is not the same evidence as a repeatable offset in the source recording.
For operators who cannot leave a computer running just to keep a fixed file on air, StreamNeo removes the need to keep OBS open on that computer: you upload the video, provide the YouTube stream key, and the broadcast can continue with the computer switched off. That addresses the burden of leaving a machine running, not source preparation or OBS audio-timing controls; the file still needs to be checked for its own alignment and loop transition.
Before committing, compare the operating options on the pricing page. When the file and channel are ready, start free — 24-hour trial, no card.
FAQ
Why does my audio drift out of sync during a YouTube livestream?
First establish whether the mismatch is fixed, gradually growing or tied to a loop restart. Compare a local OBS recording with a private or unlisted stream at the beginning, near the end and just after the loop. A growing or boundary-only error needs diagnosis of the file, source timing and playback path rather than an arbitrary fixed offset.
How do I loop a video in OBS without the audio going out of sync?
Add the local file as a Media Source, enable Loop and verify that its audio is routed to an active stream mixer track. Then test the full pass and restart in a recording and a representative private or unlisted stream. Those checks can identify a problem but do not guarantee identical behaviour for every file or setup.
Should I add an audio sync offset to fix a loop-boundary jump?
Only use an offset when the evidence shows a stable mismatch through the playback, and adjust only the affected source. A jump that appears at the restart may come from the file’s ending, opening or restart behaviour, which a constant delay will not repair. Retest the beginning, end and loop point after every change.
Does 4K60 require a different audio sync setting?
There is no universal 4K60 sync value established by the guidance. YouTube’s 4K ingest recommendations concern stream settings and delivery, while the sync diagnosis depends on the file, OBS source and output path. Test your intended configuration and check current YouTube guidance before going live.