If audio is drifting out of sync on a YouTube RTMP loop, first establish whether the mismatch stays roughly constant or grows as the stream runs. Those patterns need different investigations: a measured, stable offset may suit an OBS sync adjustment, while progressive drift calls for checking clocks, sources, routing and the path to YouTube.
Make a local OBS recording and compare the same moment with the YouTube output. If both are out of sync, start with what OBS is sending; if the local recording is sound but YouTube is not, inspect encoder health, ingest messages and delivery. An RTMP loop is not automatically the cause.
Fixed offset or progressive drift?
A fixed offset is a mismatch that remains about the same across the segment you inspect. For example, a speaker’s mouth may move before the corresponding voice by roughly the same amount near the start and later in the recording. A progressive mismatch changes: the sound may begin in step, then appear increasingly early or late as the programme continues.
These are useful descriptions, not diagnoses. A stable offset can arise at different points in the audio or video path, and a growing error does not by itself prove that RTMP is responsible. Record when the mismatch is first noticeable and whether it worsens during one pass through the file, after a loop boundary, or throughout the session. Note the source type, such as an OBS Media Source, a VLC playlist or a captured device.
Do not reach for an OBS sync offset just because the stream looks wrong at one moment. A fixed delay control shifts one signal relative to another; it does not establish why the timing changed, and a setting that helps at the start can make the end worse if the error is progressive. The distinction is the first practical test because it tells you whether a delay adjustment is even a sensible experiment.
Also separate audio/video synchronisation from stream latency. YouTube playback may arrive later than the content being produced, particularly depending on the selected latency mode, but that capture-to-playback delay is not the same as audio gradually moving out of step with video. YouTube explains latency options in its live-stream latency guidance. Judge whether sound matches the picture in the same playback, not how long it takes to reach your screen.
Make a local OBS recording for comparison
A local recording helps divide the problem into two branches. It captures what OBS produces before you draw conclusions from the version played back through YouTube. OBS recommends recording a normal session and listening back; YouTube likewise advises checking the encoder output and local archive when a live stream appears or sounds wrong. Use a representative stretch long enough to show the reported drift, not only a brief opening clip.
Keep the test controlled. Use the same scene, media file, audio routing and output settings as the live loop. Start the recording, note the time or a recognisable cue, and let it run through the point where sync usually becomes wrong. Then compare that same cue in the local recording and the YouTube stream or archive. Do not change an audio delay, sample rate and source at once: if the result changes, you will not know which adjustment mattered.
| Local OBS recording | YouTube playback or archive | Where to investigate first |
|---|---|---|
| Drifts in the same way | Also drifts | Media source, audio route, capture device and clock behaviour inside or before OBS |
| Stays in sync | Drifts or breaks up | Encoder output, ingest warnings and outbound delivery path |
| Has a stable mismatch | Similar stable mismatch | Find where it enters the signal path, then measure before trying an offset |
| Stays in sync | Appears different only during live viewing | Compare a recording or archive and account for playback latency before changing settings |
The table narrows the search; it does not prove a single cause. If the local file is already wrong, YouTube cannot be the first place that particular mismatch appears. If it is clean, investigate the encode and delivery branch rather than randomly changing the media source. For a wider 24/7 workflow context, the guide to rotating video blocks on a YouTube stream may help you identify whether an error lines up with a scheduled source change.
Check the media source and audio routing
In OBS, open the source that provides the loop and confirm it is the intended file or playlist. Check that its playback speed is normal, the expected audio track is selected, and looping is enabled if the programme is meant to repeat. For a Media Source, OBS documents both its Loop property and an option to restart playback when the source becomes active. Those settings affect playback behaviour, so verify what the scene actually uses rather than assuming a source with a familiar name points to the right file. See the OBS Media Sources guide.
If VLC Video supplies a playlist, confirm the playlist order and its loop behaviour as well. Its network-caching control concerns playback from network sources; changing it is not a general sync repair. Make one change at a time and make another short recording after each change. A loop boundary is especially informative: if the offset changes abruptly at that point, compare the source’s restart behaviour and audio track with what happens during uninterrupted playback.
Trace the audio path from its origin to the OBS mixer. Ask which source supplies the loop’s sound, and whether that same input or output device is also enabled globally under Settings → Audio or added separately to the scene. Capturing one device by two routes can produce duplicate sound or echo. That is different from progressive drift, but it can make it difficult to judge which signal is late. OBS describes audio inputs, outputs and scene sources in its audio sources documentation, and its audio mixer guide explains how to inspect mixer activity.
Listen at successive points in the route. If you can listen to the original file outside OBS, compare it with the OBS monitor or local recording. The point is not to decide by ear that a device is faulty; it is to find the earliest place where the mismatch becomes audible. Monitoring itself can add confusion, so compare recorded material as well as live monitoring, and avoid treating a delayed headphone monitor as proof that the recording is delayed.
For a prerecorded devotional or radio programme, the file itself may already contain a mismatch. Check a copy outside the stream workflow, especially near the start and end of the section you are looping. A guide to broadcasting prerecorded Hindi talk-radio shows covers the broader source preparation problem; the diagnostic principle here is simply to distinguish a file defect from a live routing or timing issue.
Inspect capture devices and clock behaviour
If OBS receives audio from a microphone, interface, USB device or other capture source, check whether the problem follows that device. A useful controlled comparison is to record the same video while using a different available audio source, then restore the original arrangement. Do not buy an interface or adapter on the assumption that hardware will cure progressive drift; the evidence you have so far should guide whether a device needs inspection.
On Windows, OBS’s Audio Input/Output Capture source documents a “Use Device Timestamps” option intended to prevent desynchronisation. Treat it as a test, not a universal fix. Change only that setting, record a representative segment, and compare the result. Its relevance depends on the operating system, capture source and device; it is not a general switch for a file-based Media Source or every OBS installation. Details are in the OBS audio sources documentation.
Clock behaviour is one possible explanation when independently timed audio and video sources gradually diverge, but the symptom alone does not identify a clock fault. If a file-based source and a captured device are combined, test them separately where practical. If the drift follows one capture device across tests, note the device and operating system for further troubleshooting; if it remains with the media file, return to its playback and routing path.
A long-running channel makes repeatable tests more useful than a late-night guess. Keep a short note with the source, whether the local recording drifts, when the mismatch begins, and what single setting changed. A Mac mini stream troubleshooting guide addresses a different failure, but its practical lesson applies: record the operating-system and change context rather than assuming a symptom has one cause.
Compare local output with YouTube playback
When the local OBS file is in sync but YouTube playback is not, look beyond source settings. Check the encoder’s preview or output, its health indicators and the Live Control Room for warnings around the period in question. YouTube’s encoder troubleshooting guidance recommends checking the encoder, CPU load, Live Control Room errors and local archive. If output appears healthy there, assess the outbound connection; if not, investigate the encoder or routed sources.
Use an archive or a later playback to compare the same cue, rather than relying only on a live player that may be buffering or using a different latency setting. If the archive is also wrong while the local file is clean, that points towards the encode or ingest branch, but it still does not isolate the precise fault. Save the time, relevant warning text and the local comparison result before changing settings.
Check the current YouTube encoder baseline and messages. YouTube’s live encoder guidance specifies RTMP or RTMPS, constant bitrate (CBR), a recommended keyframe interval of two seconds and a maximum interval of four seconds. It specifies AAC or MP3 audio and sample rates of 44.1 kHz for stereo or 48 kHz for 5.1 surround. Treat these as ingest requirements and checks, not proof that a gradual sync problem has been solved. YouTube’s current live encoder settings and Live Control Room messages are the place to verify the applicable guidance.
If you run a file-based radio or music loop, compare the sound and picture at a recognisable cue in the local recording and platform output. Album art and video changes can make visual comparisons easier; the guide on showing album art on a YouTube internet radio stream addresses that presentation detail. Do not infer audio timing from an image change alone if the visual is static or the playlist has a transition.
Measure before applying an OBS sync offset
An OBS sync offset is a timing adjustment for a stable difference between audio and video. Before using it, measure the mismatch in a recording: choose a clear event, such as a visible clap, a spoken consonant paired with a mouth movement, or a sharp impact, and compare its sound and picture near the start and later in the segment. Repeat with another event if available. You need to know whether the difference is consistent, not merely that something looks wrong.
If the same signal is consistently early or late, change one relevant delay control in a small, documented test and record again. Keep the original value noted so you can reverse the change. Verify both an early and later section, and check a loop boundary if the source restarts. A setting that aligns one point but worsens another indicates that a static offset is not describing the whole fault.
Do not apply an arbitrary number copied from someone else’s setup. The measurement depends on your source, device and routing. Most importantly, a static sync offset is not a proven fix for progressive drift: the reviewed guidance does not establish that shifting a track once will correct a mismatch that continues to grow. If the error changes over time, return to the source, capture clock, encoder or delivery investigation instead of compensating at one moment.
Review encoder health and ingest messages
If the local recording remains in sync, inspect the encoder during a controlled test. Confirm that the intended audio and video sources reach the encoder, review CPU load and output status, and note any dropped frames or audio-related warnings shown by OBS or the encoder. A local recording and an outgoing stream can use different paths, so a clean local result is a clue, not a guarantee that the live output is clean.
Read YouTube’s Live Control Room messages rather than guessing from the player. A warning about unsupported format, audio stream count, sample rate or frame rate should be addressed against the current official guidance. The recommended keyframe interval and audio sampling values described above help establish a supported ingest baseline; they do not independently explain every instance of drift. Change only the setting implicated by an observed warning, then repeat the local and YouTube comparison.
If the encoder output looks healthy but platform playback still differs, check the outbound connection and any network interruptions noted during the same period. A brief drop can interrupt delivery without proving that network conditions caused a gradually increasing audio/video mismatch. Keep the exact timing of the symptom and connection warnings together, and avoid changing unrelated audio delays to compensate for delivery trouble.
For channels that must keep a prerecorded loop running while your own computer is off, StreamNeo removes the need to leave OBS running on your desktop; it does not remove the value of checking the prepared file, stream key and YouTube playback when you are diagnosing sync. Whichever workflow you use, test a representative section and verify the result on YouTube before relying on an overnight run.
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
Can an OBS sync offset fix audio that keeps drifting?
It can be tested when the mismatch is stable, but a fixed offset is not an established cure for progressive drift. Compare early and later points in a recording before changing it; if the gap grows, investigate the source, capture device, clocks or output path.
Does YouTube stream latency cause audio and video to go out of sync?
Latency describes the delay between capture and playback, not necessarily a change in the relationship between sound and picture. Compare audio and video within the same local recording or YouTube playback, and account for the selected latency mode before diagnosing drift.
What if my local OBS recording is in sync but YouTube is not?
Check the encoder output, CPU load and Live Control Room warnings, then compare an archive or later playback with the same cue in your local file. If the encoder looks healthy, investigate outbound delivery; the comparison narrows the branch but does not identify a specific cause by itself.
What should I check first in an OBS loop?
Confirm the actual media file, playback speed, audio track and loop or restart settings. Then check whether the same audio device is captured twice, and record a test after each change so you can tell whether the mismatch moved or changed.