If audio and video are out of sync by the same amount from the beginning, you may be dealing with a fixed delay. If they start aligned and separate further as a YouTube Live loop runs, that is progressive drift, and an OBS Sync Offset cannot correct its growing timing difference.
Start by comparing OBS preview, a local recording and the YouTube playback result. Then check sample rates and the particular audio source, before treating performance or network warnings as possible clues rather than proof of a cause.
Is the audio offset fixed or growing?
Watch for a change in the gap, not just a gap. A fixed offset is a consistent delay: for example, the same visible mouth movement is followed by the same audio delay near the start and later in the stream. Progressive drift begins with closer alignment and worsens with time. A short clip may not make that pattern clear, so compare a cue near the start with another after the stream has run long enough for the symptom to appear.
For a prerecorded loop, choose a moment that is easy to recognise both visually and audibly: a spoken word, a hand clap or a sharp musical attack. Note whether the sound leads or trails the picture, and whether the separation changes. Avoid judging only from a viewer's impression that something “feels late”; playback buffering can make the whole stream arrive later without changing the relationship between its picture and sound.
Four patterns are worth keeping separate:
| What you observe | What it may indicate | First useful check |
|---|---|---|
| The same audio delay is present throughout | A stable offset | Compare repeatable cues, then consider source-specific Sync Offset. |
| Picture and sound begin aligned, then separate | Progressive drift | Check source timing and sample-rate consistency; reproduce in a local recording. |
| The whole live programme reaches you late, but sound matches the picture | Stream latency | Check latency and playback conditions, not Sync Offset. |
| Choppiness, dropped frames or disconnections occur too | A performance or connection concern may coexist | Review OBS and YouTube diagnostics without assuming they explain the drift. |
These are diagnostic clues, not conclusions. More than one issue can be present, and the same visible symptom can come from different places in the signal path. The useful first result is knowing whether the mismatch grows and where you can reproduce it.
Compare OBS preview, local recording and playback
Compare the same cue in three places: the OBS preview, a local recording made by OBS, and the viewer playback on YouTube. If the preview and recording both grow out of sync, focus first on the source, device timing, routing or computer path. If the local copy stays aligned but YouTube playback does not, inspect the live output and viewer playback conditions before changing a capture device.
Make a local recording while reproducing the loop rather than relying on memory. A recording offers a way to revisit the start and later cue at normal speed. It can also show whether the issue exists before YouTube receives the stream. Check that the recording includes the same audio source and scene or output path as the live broadcast; a recording made from a different mix may not represent the stream accurately.
For a test stream, use private or unlisted visibility where appropriate, and avoid changing several settings between comparisons. Keep the same video, scene, source routing and approximate run length. If you have a second viewer device, compare playback there too, but remember that browser, device and connection differences may affect the viewing experience.
A repeatable test is more useful than a single observation. Record the time at which the mismatch first becomes noticeable, whether the audio is ahead or behind, and how it changes later. If you use OBS for a prerecorded broadcast, the guide to streaming prerecorded videos with OBS on Windows 11 covers the broader setup; for this fault, keep your investigation centred on the specific point where timing changes.
Check audio sample rates and the affected source
A sample rate describes how often digital audio is represented per second. When devices and software in the same path use inconsistent settings, timing or conversion problems may be worth investigating. YouTube's Live Streaming API documentation lists 44.1 kHz and 48 kHz as recommended audio sample rates and flags sample-rate mismatches between primary and backup streams as configuration issues. Those recommendations do not mean that selecting either value will automatically resolve every drift symptom.
Check the audio settings in OBS and the formats selected for the devices that actually feed the stream: an audio interface, capture card, microphone or system audio source. Keep the participating path consistent at a supported rate where your devices allow it. If you use primary and backup ingestion, verify that their audio settings match as well. Do not change every device just because one source is affected; identify which source is routed into the stream and test that path first.
In OBS, inspect the mixer and source properties to establish which input carries the audio. A video loop may combine audio embedded in the media with music, a microphone or system sound. If only one of those sources appears out of step, isolate it in a controlled test rather than adjusting the entire mix. Muting or removing an unrelated source for a test can help localise the fault, provided you restore the intended mix afterwards.
On Windows, OBS documents a “Use Device Timestamps” property for audio input and output capture sources. It is intended to use timing information from the device to help prevent desynchronisation, and OBS notes it is off by default. Test it only on the source that appears implicated, if the control is available. Change one setting, repeat the same test, and compare the result; there is no basis for toggling it on every source as a universal fix.
If you need to ask for help, collect the OBS version, operating system, device or capture-card model, source routing, sample rate in OBS and relevant devices, and whether timestamps were enabled. Include whether the local recording drifts, how long the test ran before the change appeared, and an OBS log from a representative reproduction. Specific evidence is more useful than buying an audio interface, capture card or cable before a fault has been identified.
Use Sync Offset only for stable delay
OBS's Sync Offset is for shifting an audio source by a fixed amount relative to video. It is useful when the delay is stable and repeatable: a cue is consistently late by the same amount at the beginning and later in the programme. A fixed shift can align that source at both points, assuming the rest of the signal path remains consistent.
It cannot correct progressive drift. If a source is increasingly late, applying one offset may make one moment line up while leaving an earlier or later moment wrong. Changing the offset again and again to chase the moving gap can hide the symptom at one point without addressing why it grows. Return to the source, rate and timing comparisons before considering a fixed adjustment.
When the evidence points to a stable offset, use a clear sync cue and make small, controlled changes in OBS Advanced Audio Properties. Check more than one point in the loop after each adjustment, then verify the result in the output that matters. Do not copy a millisecond value from another creator or device guide: the required value depends on the capture path, hardware and settings. OBS support discussions and vendor examples may help explain where the control is, but their particular values are not transferable diagnoses.
Keep changes reversible. Note the original setting, make one adjustment, and record what changed. If the apparent delay differs between playback devices or changes as the stream continues, an offset is unlikely to be the right first action. A useful rule is to calibrate only after you have established that the gap is stable.
Inspect OBS performance and stream health
Performance and connection problems deserve a separate check. OBS says GPU overload can affect rendering and encoding performance, and recommends reducing workload when output is choppy or overloaded. Its connection troubleshooting guide discusses dropped frames and intermittent disconnections in relation to the network path to the ingest server or available bitrate capacity. Either kind of warning can accompany a sync complaint, but neither by itself proves why audio is drifting.
Look at OBS statistics or status indicators during a representative run. Note whether rendering or encoding lag appears, whether frames are dropped, and whether the connection disconnects or recovers. If output is choppy or OBS reports overload, reduce the demands on the computer for a controlled test—for example, close unrelated heavy applications or simplify a scene—then compare again. If the concern is dropped frames or unstable connectivity, investigate the connection and the bitrate you can sustain rather than changing audio offset blindly.
YouTube's stream health guidance recommends testing with audio and video motion similar to the real event, then monitoring stream health and reviewing messages while live. Use a test representative of your loop: a static image with silence is not a meaningful check if the real channel has music, movement and several audio sources. A clean health display is useful, but it does not establish that the source timing is correct or rule out gradual drift.
Stream latency is another distinct issue. YouTube describes latency as the time between capture and playback for viewers; it concerns how late the whole live stream arrives. Lower latency uses less read-ahead buffering and can make viewers more sensitive to problems between encoder and player. That is not the same as audio progressively moving relative to its video. If a viewer sees the whole loop late but speech, music and picture remain together, investigate latency and playback instead of adding Sync Offset.
For a computer that must remain on for a continuous broadcast, reducing local setup and overnight monitoring demands may be relevant once you have established that the media and channel are ready. StreamNeo addresses the specific burden of keeping your own computer running by turning an uploaded video into a YouTube-only live stream that continues with your computer switched off; it does not diagnose or guarantee a fix for a source file whose audio drifts.
Check the YouTube playback result
After checking the source and encoder path, examine the viewer result. Compare the beginning and later cue on the same playback device and, if possible, another connection or device. Wait for the playback to settle before making a judgement: buffering or a temporary playback interruption can affect when you see content without necessarily changing the encoded audio/video relationship.
If your OBS preview and local recording remain aligned while YouTube playback diverges, note whether the issue is consistent across viewers. Check YouTube's stream-health messages and the ingest configuration. For a primary and backup stream, verify that configuration, including the audio sample rates, is consistent. If only one viewer reports the issue, playback conditions may be contributing; do not assume that a physical audio device is at fault from that alone.
If you are still unsure how playback buffering affects what a viewer sees, the explanation of what a video player does can help separate delivery and playback from the source signal. Keep the distinction practical: latency means later arrival; drift means the relative timing between picture and sound changes. A delayed but internally aligned programme calls for a different investigation from one whose audio gradually goes out of sync.
Test changes with a short stream
Do not treat “short” as a particular duration. Run the test long enough for the reported fault to appear, using the same loop and a beginning-and-later cue. If the drift usually takes time to become noticeable, a quick preview cannot rule it out. A private or unlisted broadcast can let you inspect the real path without making an unfinished troubleshooting session public.
Use a simple change log: write down the baseline settings, the one variable changed, the test conditions and the result. Change one thing at a time—sample rate, device timestamp behaviour, scene load or connection conditions—so you can tell which difference mattered. If several changes are made together, even an improved result does not show which one helped, and a later regression is harder to diagnose.
When the picture and audio remain aligned in the local recording but not in YouTube playback, preserve that comparison along with the time and health messages. When both local and live outputs drift, preserve the local file and OBS log. In either case, give support the source routing and device details rather than reporting only that YouTube is out of sync. If the issue does not reproduce reliably, describe what you observed and avoid presenting a guess as the cause.
A stable 24/7 loop also needs checks beyond sync. The guide to checking whether a 24/7 lecture stream is still playing offers a related way to think about ongoing playback monitoring. For a broader comparison of how local and hosted operation affect a continuous broadcast, compare options for 24/7 YouTube streaming; neither step replaces diagnosing the audio path, but both can inform how you supervise a long-running channel.
When you have a repeatable diagnosis and the channel is ready, compare the operating options before committing.
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 OBS Sync Offset fix audio that gradually goes out of sync?
No. Sync Offset shifts audio by a fixed amount, so it can help with a consistent delay but cannot correct a mismatch that grows over time. Check the source timing, sample rates and local recording first.
Should I use 44.1 kHz or 48 kHz?
YouTube's Live Streaming API lists both as recommended sample rates. Check the settings across OBS and the devices in the actual audio path, and match primary and backup streams if you use them; neither rate is a universal cure.
What if the local recording is in sync but YouTube playback is not?
That points you towards the live output, stream health, ingest or viewer playback conditions, but does not prove which one is responsible. Compare playback on another device or connection and note YouTube's health messages before changing source hardware.
Does a late YouTube stream mean the audio is drifting?
Not necessarily. Latency describes how long the whole stream takes to reach a viewer; drift describes audio and video becoming progressively less aligned with each other. Check a cue at the start and later to distinguish the two.