When audio from a prerecorded video reaches you before the matching picture, measure the difference with a short OBS recording and adjust that source’s Sync Offset. Do not start with a standard millisecond value: the right correction depends on the file and the path through your particular setup.
A fixed offset can address a steady lead or lag. If the mismatch grows while the video plays, or appears only after the stream leaves OBS, a larger fixed offset is unlikely to solve the underlying problem. Test each part of the path before going live.
Add the prerecorded file as an OBS Media Source
In OBS, add the local video to the scene as a Media Source. This source type is intended for media files played within an OBS scene. OBS documents common local video formats such as MP4, MOV, MKV and WebM, along with audio formats including MP3, AAC and WAV. Check the current OBS Media Sources documentation if your file format or the source properties are unclear.
Give the source a useful name, such as “Opening bhajan file” or “Local news loop”, especially if the scene has more than one video or audio source. A named source makes it easier to adjust the audio you actually hear rather than a microphone, desktop-audio device or another clip. If the media source is selected, its preview should show the intended file.
Review the playback options before testing. Loop restarts the file after it finishes. Restart playback when source becomes active restarts it when the source is in the current scene and visible. For a one-time sermon segment, you may want it to play from the beginning when shown; for a continuous ambience scene, you may want it to loop. Choose behaviour that matches the sequence you plan to broadcast, because switching scenes or restarting a source during testing can change what you hear.
Keep the test simple. If the file contains its own audio, begin with that audio rather than adding a separate capture device or routing it through other software. Extra paths make it harder to identify where a timing difference starts. If you are building a longer sequence of worship recordings, this guide to alternating sermons and worship recordings on a church YouTube Live stream covers the broader scene-planning problem; first establish that one file plays correctly.
Make a short test recording
Choose a section of the video with an event that is easy to match in sound and picture. A hand clap, a door closing, a drum strike or a person visibly starting a word can work. Avoid judging sync from background music alone: sustained notes and soft ambience do not provide a clear instant to compare.
Record a short sample in OBS that includes the event, then stop the recording and inspect it in a player or editor. A local recording is useful because it captures OBS’s output without requiring you to infer timing while watching the preview. OBS community troubleshooting advice also recommends recording and reviewing the tracks when diagnosing sync; see the OBS discussion of video and audio sync problems.
Use the same scene, file and audio path you intend to use for the stream. If you only record the media source but the live scene also includes another audio source, the recording may not represent what viewers will hear. Conversely, do not add devices just to make the test look more like a complex final setup if they are not part of the intended broadcast.
Write down what you observe before changing anything: whether the sound is early or late, whether the gap seems steady, and where in the clip you checked. You do not need a precision timing instrument. The goal is to establish direction and whether the difference remains similar, not to claim a universal measurement for other files or systems.
If you run a continuous channel from a laptop, this short recording is still worth doing before leaving the show unattended overnight. The practical considerations in making a 24/7 YouTube lofi stream using a laptop in India include the wider always-on setup, but a sound check should begin with the clip and output you actually plan to use.
Decide whether sound leads or lags
Find the visible instant and compare it with the sound. If a person’s mouth begins to open or a hand strikes an instrument, does the sound arrive first, at about the same time, or after the visible event? Repeat the check at another obvious event in the same recording to make sure you are not reacting to an ambiguous moment.
A consistent lead means the sound is early by roughly the same amount at the beginning and later in the clip. That is the case where delaying the relevant audio source may help. A consistent lag means sound follows the picture. Delaying that audio further moves it in the wrong direction; investigate a video delay or the playback and capture path instead.
A difference that changes over time is a different problem. If the first event is close but later events are increasingly out of step, that is drift, not a single fixed offset. A fixed delay can align one point while leaving another worse. Compare the beginning and a later point before deciding which case you have.
Also note which output you inspected. An OBS recording can show whether the timing is already wrong in the recorded OBS output. If that recording looks right but a live viewer reports a mismatch, the issue may be farther along the path, or related to that viewer’s playback. Do not assume that a local correction proves what YouTube receives or what every viewer sees.
For a channel that mixes music and speech, a small mismatch can be more obvious on a close-up singer than on a landscape or static artwork. Use an event from the actual programme where possible. The audio timing in one clip is not a calibration value for every other source or scene.
Set the relevant Sync Offset
In OBS, open Edit → Advanced Audio Properties and locate the audio source that corresponds to the prerecorded file. The Sync Offset (ms) field is the millisecond delay control discussed in OBS support material; the OBS forum thread on forcing an audio delay identifies the setting. The reported sign behaviour and outcomes can depend on the source and release, so confirm the result with a new recording rather than relying on a rule of thumb.
If the file’s sound consistently leads, apply a small test adjustment to delay that audio, then record again. Change one thing at a time and compare the same event. If your first adjustment makes the sound later than the picture, reduce or reverse the correction as appropriate for the installed OBS version and source. Do not copy a number from a forum post as a preset: it describes someone else’s signal path, not yours.
If sound is already late, do not add an audio delay. Inspect whether the video source needs a delay or whether another part of the playback path is adding latency. OBS community discussions describe both audio delay and video-source filtering as possible adjustment points, but the correct one depends on which signal is early. Make a fresh test after any change.
Only adjust the source whose sound is out of sync. If the Media Source’s audio and a separately captured device are both active, changing a global or unrelated source can create a new mismatch or echo. If you are unsure which source contributes the sound, mute one source at a time during a recording test and listen for the change. Restore the intended mix before the live test.
Keep a brief note of the tested value and result, along with the file and scene used. The value is a calibration for that combination, not a general OBS setting. If you replace the file, change the scene’s audio routing or use a different playback device, repeat the test instead of assuming the previous correction still applies.
Replay and inspect the corrected output
Record a new sample after every adjustment. Replay the same event and compare the beginning and a later event. A good result is not merely that one clap looks right: speech, music or another clear event later in the clip should remain aligned as well. If the correction improves one point but worsens another, return to the drift checks rather than adding more delay.
Listen as well as watch. A video player’s controls, display processing or playback performance can affect what you perceive, so use the same player and conditions for before-and-after checks when possible. If an editor allows you to inspect audio and video tracks separately, it can help you locate the difference, but no special purchase is required for the basic test.
Check that the recording contains the audio you intended to adjust. OBS may have a media source, desktop audio and a separate microphone active at once. If a microphone captures sound from speakers, it can produce a delayed duplicate that sounds like sync error. Listen for echo and check the sources in the mixer before concluding that the prerecorded file itself needs another offset.
For a continuous programme, also test a point after a scene transition or a media-source restart if that is part of the show. A source configured to restart when it becomes active may begin at a different point than you expect. A loop can also introduce a new start boundary. Verify that the corrected timing persists through the actual playback behaviour, not only in a clip started once by hand.
Verify the actual stream path
A local OBS recording checks OBS’s recorded output; it does not establish what a viewer receives through YouTube Live. When practical, make a controlled live test and inspect the received stream on a separate device or connection. Use a private or otherwise controlled test if appropriate for your channel, and check a clear audio-visual event rather than judging from a long stretch of music.
Compare the results by stage: does the local OBS recording have the mismatch, does it appear in the received live stream, or does it appear only on one viewer’s device? This distinction helps avoid correcting OBS for a playback issue downstream. The available sources do not establish a particular YouTube ingest delay or timing behaviour for your stream, so do not attribute a measured discrepancy to YouTube without evidence from your own test.
Keep the test output and notes until the channel is ready. They give you a comparison point if the issue returns after changing the file, audio device, scene or OBS version. If the local recording is aligned but the received stream is not, gather those specifics before changing offsets again; a local adjustment that only happens to compensate on one viewing path may make the source wrong elsewhere.
If the channel is meant to run unattended, decide how you will check the actual stream after making changes. A person watching a local preview is not the same as confirming the public or controlled viewer path. For a separate guide to pre-broadcast checks, see how to test a YouTube radio livestream privately before going public.
If repeated manual restarts, a running computer or overnight checks are themselves the problem for a prerecorded channel, StreamNeo removes that specific computer-running burden by turning an uploaded video into a YouTube live stream that can continue with your computer switched off; it does not replace checking sync in the stream path you intend viewers to use.
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
Should I use a standard Sync Offset value?
No. Measure the discrepancy in your own short recording and adjust from that result. A value that helped another file or setup is not evidence that it will correct yours.
Does a positive offset always delay the sound?
OBS support discussions describe Sync Offset as a millisecond delay control, and users report positive values delaying audio. Confirm the direction in your installed release with a short test, because the result can depend on the source and setup.
What if the sound is late rather than early?
Do not delay the audio further. Check whether a video-source delay or another part of the playback path is responsible, then record a fresh test after changing one thing.
What if sync is correct locally but wrong on the live stream?
A local recording only shows OBS’s recorded output; it cannot prove what viewers receive. Check a controlled live output separately, note where the difference first appears, and avoid assuming a particular YouTube ingest behaviour without evidence.