If audio is out of sync on a 24/7 vMix stream, first compare vMix’s local output or recording with YouTube playback. Change an audio delay only if the mismatch is present in the local output; if it appears only on YouTube, check the encoder, stream health and upload path first.
A sound that is early needs a different correction from one that is late, and a fixed offset differs from drift that grows over time. Use the same visible movement and sound as a test, record what you observe, and make one small change at a time. India is not, by itself, evidence of a particular cause: measure the connection where the stream actually runs.
Locate the first point where sync breaks
Start with a repeatable test, not a guess. Use a visible action paired with a sound, such as a hand clap in front of a camera, or a clearly timed beat and movement in a test video. YouTube recommends testing with audio and movement representative of the real stream. A quiet devotional image with no obvious movement will not make a useful lip-sync test, so add a brief test segment before relying on it.
Compare three views if possible: vMix’s programme output, a local recording, and YouTube playback. The local programme view helps establish what vMix is producing at that moment; the recording gives you something to replay and inspect carefully; YouTube shows what viewers receive. If you cannot record locally, at least compare the vMix output and the Live Control Room preview, while remembering that a preview is not the same evidence as a saved file.
Note whether the audio is early or late, and whether the gap stays about the same or changes with time. An early sound can point to separately connected audio arriving before the video. A mismatch that gradually increases may indicate a timing or delivery problem that a single fixed input adjustment will not solve. Do not call a changing gap a fixed offset just because one moment looks correct.
| What you observe | Where to investigate first | What not to do yet |
|---|---|---|
| vMix output and local recording are both out of sync | Input routing and the relevant vMix audio input timing | Do not change a YouTube delivery setting to compensate |
| Local output is right, YouTube playback is wrong | Stream health, encoder load, archive and upload conditions | Do not blindly shift audio in vMix |
| The gap changes during the run | Compare beginning and later evidence; inspect stream and connection conditions | Do not assume one fixed delay will correct the whole run |
| Only one source is wrong | That source’s path and input settings | Do not offset every audio source |
A useful example is a bhajan channel with a camera showing a singer and a separate audio mixer feeding vMix. If the singer’s mouth and the local recording already disagree, investigate that production path. If the recording is aligned but YouTube playback is not, keep the vMix offset unchanged while checking delivery evidence.
Compare vMix output and local recording
Make the local comparison as close as possible to the stream setup. Check the same camera, audio source, frame rate and scene that are in the live programme. A test made with a different input or a different routing arrangement may hide the very mismatch you are trying to locate. YouTube’s encoder setup guidance recommends a representative test before a live broadcast; use that principle even if your stream has already started and you are preparing a controlled diagnosis.
In vMix, watch the programme output and, where practical, create a local recording that covers the test cue. Replay it rather than relying only on what you hear while producing. Live monitoring is affected by the delay of the display, audio device and playback path; a recorded file can help distinguish a production sync error from a monitoring-path delay. Keep the original recording unchanged before making a correction, so you can compare the before and after.
If the local recording is consistently misaligned, identify which inputs contribute to the image and sound. Are camera video and audio embedded in one capture input, or does the picture come from a camera while sound arrives from a separate interface or mixer? If the sources are separate, their paths can introduce different timing. If a single input is wrong, inspect that input and its connection rather than adding a delay to unrelated sources.
For a 24/7 channel, a brief, planned test segment is safer than changing settings during an important programme. If you can, use a known test scene during a quieter interval, tell anyone monitoring the channel what is being tested, and note the time. This is useful for a continuous Marathi sermon stream, where a clean, repeatable opening test can establish whether the playback chain is aligned before the long run.
Check input timing before changing the offset
vMix’s audio delay is configured for an individual audio input. Its audio-delay guide describes using the Audio Mixer, opening the affected input’s settings, selecting General, and adjusting Delay in milliseconds. If sound arrives before the matching picture, increasing that input’s delay can bring the two together. If sound is late, adding more delay moves it further in the wrong direction; first check the routing and timing path rather than copying another operator’s setting.
Open the Audio Mixer and identify the input that carries the sound you are testing. Click that input’s settings and find General and Delay. Make a small adjustment, replay the same cue, and record the result. Repeat only if the measured result is still off. The vMix guide does not supply a universal delay value for every camera, capture card or audio interface, so a number from someone else’s setup is not a diagnosis for yours.
Also check whether the delay is stable across the duration of the test. If a clap lines up at the start but is noticeably out later, a fixed offset may make one moment look right while making another worse. Inspect the input format, routing and device behaviour, and compare recordings at more than one point. Do not try to correct a growing timing difference by repeatedly increasing an input delay.
If video and sound arrive through separate devices, confirm that you have selected the audio input associated with the picture in question. A delay applied to the wrong input may make one source appear better while shifting another. vMix’s documented example of a combined path is camera audio carried through a capture-card input. A compatible capture card with embedded audio may be worth considering only after confirming that your camera, connection standard and current routing support it; hardware is not the default fix for a setting that has not yet been measured.
Keep a simple note of the starting delay, each change, the cue used, and whether the sound was early or late. If the test becomes less reliable after a change, return to the recorded starting setting and examine the path again. This makes a useful distinction between a result that improved on evidence and a change that merely sounded different while you were monitoring.
If only YouTube playback is affected
When vMix output and the local recording are aligned but YouTube playback is not, leave the input delay alone for now. Compare the outgoing encoder view with what appears in the Live Control Room and inspect stream-health messages. YouTube’s stream troubleshooting guidance points operators towards checking the encoder output, dashboard errors, CPU load, local archive files and outbound connection. Use the message timestamps to find whether the problem coincides with a particular period of the broadcast.
The Live Control Room can show stream health and metrics; use the live-stream metrics page to understand what YouTube exposes for the stream. A health warning is evidence to investigate, not proof that it caused the lip-sync issue. Write down its wording and time, then compare it with the local recording and any change in the playback sync. If the playback is delayed but the local file is not, keep the two paths separate in your notes.
Check an archive or local recording around the same time. A healthy local file with a YouTube-only mismatch moves attention towards encoding or delivery. If both the local recording and YouTube version are wrong, return to input timing. If no local evidence exists, avoid making a confident diagnosis based on the viewer’s device alone: browser buffering and monitoring delay can affect what one person sees.
It is also useful to test from a second playback device or network when practical, without changing the live encoder. That does not prove where the fault lies, but it can establish whether the symptom is consistent across viewers. For a 24/7 radio stream over JioFiber, for example, the actual upload conditions at the streaming location matter more than a separate speed test run elsewhere or at a quieter time.
Review encoder load and upload conditions
A stream can look correct inside vMix and still be affected after it leaves the computer. Check vMix’s CPU load and whether the encoder reports errors or drops around the time the sync problem begins. Avoid changing several encoder settings together: if the result changes, you will not know whether the cause was load, format or connection.
Measure upload at the location and under the conditions where the stream runs. YouTube recommends that total streaming bitrate fit within available upload bandwidth with 20% headroom, and warns that a shared network can reduce the capacity available to the stream. Treat this as a recommendation from YouTube’s streaming tips, not as a claim about what a particular Indian connection should deliver. Repeat the check when other devices or users are active if that reflects normal operating conditions.
Check whether the stream shares its connection with office users, phones, cloud backups or other video traffic. A speed test showing good upload at midday does not describe a busy evening if the connection is shared or variable. Compare the test with stream-health timestamps and the time an audio mismatch appears. The location in India is a reason to measure the real route and usage pattern, not evidence that a particular provider or region causes sync drift.
Review the encoder format against current YouTube guidance rather than relying on a setting copied from an older profile. YouTube’s encoder settings page lists supported codecs and recommends constant bitrate, a two-second keyframe interval that should not exceed four seconds, and a 44.1 kHz audio sample rate for stereo; it lists 128 kbps as a recommended stereo audio bitrate. These are platform recommendations, not a promise that changing them will cure sync. Confirm that the current stream-health messages and your chosen encoder settings agree before adjusting them. vMix’s Help 27 streaming-quality page identifies AAC as the audio option when FFMPEG is selected, so check the encoder mode you actually use.
vMix’s Network Buffer is about the maximum time vMix can buffer; server and player buffers add further latency. vMix notes that increasing the buffer may help when network speed is unreliable. It can also increase delay, and it is not a substitute for correcting input lip sync. Treat buffer settings as a delivery-stability choice, not as an audio offset, and change them only after recording the original value and observing the effect.
Preserve evidence and plan around archive limits
For a continuous stream, evidence needs a plan. Keep a local recording or archive when practical, check that it is growing, and inspect the segment around the time a viewer reports a problem. A small test file can help establish normal sync before an overnight run; a later segment can show whether timing changes. Note the date and time, scene or source, relevant input delay, stream-health status and any encoder or network change.
Do not assume that YouTube will produce one complete archive for a 24-hour broadcast. YouTube’s encoder setup page says streams under 12 hours are automatically archived. That statement does not promise one complete file for a longer run. Check the current behaviour for your channel and decide how to preserve local evidence across long sessions. If your content is a Ram bhajan stream from India, planning what you can replay after an overnight issue is part of the operating setup, not an optional extra.
A local recording also has limits: it consumes storage and must itself be checked. Confirm that it is still being written, that the relevant audio is present, and that you can locate the affected interval later. A recording that stops before the incident cannot answer whether the sound drifted. Keep a short sample of the actual cue and, where appropriate, a longer segment that reveals gradual change.
Separate observations from conclusions. “Audio was early in the local recording at the clap near the start” is useful evidence. “The internet caused it” is a hypothesis until you correlate it with a network or stream-health change. That distinction helps an operator, a volunteer or a technician continue the investigation without repeating blind changes.
Monitor changes through a full run
After any adjustment, replay the same test cue and compare it with the original. Change only one variable at a time: input delay, encoder format, network buffer or connection conditions. Record the time and setting, then observe both the vMix output and YouTube playback. If you change the input delay and the stream buffer together, a better result does not tell you which one mattered.
For a 24/7 channel, a fix should be checked beyond the first minutes after a change. Monitor at the start of a run and later in the run, especially if the symptom has previously grown over time or appeared only during busy network periods. YouTube recommends continuous monitoring of audio and video quality and stream health. Arrange who will notice a warning overnight and what evidence they should preserve; an unattended channel still needs a practical way to detect and document a fault.
If the stream is generated from a single uploaded video rather than a live vMix production, there may be no reason to change an input offset at all. StreamNeo can remove the burden of leaving your own computer on to relay a prepared file continuously, while the sync diagnosis still depends on checking the file and the YouTube playback separately. It is a YouTube-only option, not a substitute for testing an incorrectly prepared source file.
Once you have a measured correction, save the working settings and keep the before-and-after evidence together. Recheck after changing a camera, audio interface, scene, encoder profile or network arrangement; each can alter the signal path. A note of what worked on this exact setup is more useful than a copied millisecond value from a different channel.
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 change vMix audio delay if YouTube playback is out of sync?
Not until you compare YouTube with vMix output or a local recording. If those are aligned and only YouTube playback is wrong, first inspect stream health, encoder load, archive evidence and the upload connection. A vMix offset could make the local output wrong while leaving the delivery cause untouched.
What delay value should I use in vMix?
There is no universal value for every setup. Use a repeatable visible action and sound, establish whether the sound is early or late, and adjust the affected input in small measured steps. Recheck the same cue after each change.
Can a larger Network Buffer correct lip sync?
A network buffer affects buffering and can add latency; it is not the same as the per-input audio Delay used to align sound and picture. vMix says a larger buffer may help where network speed is unreliable, but assess it as a stability setting and verify the result rather than using it as a blind sync correction.
Will YouTube archive my whole 24-hour stream?
Do not assume it will. YouTube’s setup guidance says encoder streams under 12 hours are automatically archived, but that does not promise one complete archive for a 24-hour broadcast. Check the current behaviour for your channel and keep local evidence when practical.