Make the decision from the signal path

Fix continuity before sync. A constant offset can be measured and corrected on the relevant source; an offset that grows over time points to dropped frames, clock differences, unstable devices or workload and should not be hidden with one delay value.

There is no universal delay value for a product. Measure the actual camera, capture and audio paths, and separate a constant offset from drift that grows during the recording.

Eight decision factors

Check these in order. Offset type and Reference event decide whether the route works; the remaining factors refine operation and convenience.

  • Offset type: Determine whether lip sync is constant from the start or drifts during the recording.
  • Reference event: Use a visible clap or other sharp event captured by both picture and microphone.
  • Source ownership: Identify which camera/capture and audio source paths create the relative delay.
  • Monitoring: Judge the recorded file, not a potentially delayed monitor preview.
  • Frame/drop health: Check OBS rendering, encoding and network/dropped-frame diagnostics.
  • Sample/device stability: Keep audio devices and routes stable before measuring.
  • Correction location: Apply one measured sync offset at the source/bus that needs it.
  • Duration: Retest for the full intended session length and after reconnect/reboot.

Constraints that change the answer

Start with Growing drift. The conditions below can change the choice or the next troubleshooting branch.

  • Growing drift: A fixed delay cannot solve a changing clock or dropped-frame problem.
  • Duplicate audio: Two paths can sound like sync error; remove duplication first.
  • Destination-only issue: Compare local recording with destination playback before changing sources.
  • Mode change: Camera/capture resolution and frame rate changes can alter latency.
  • Wireless monitoring: Do not use Bluetooth playback as the timing reference for source sync.

A practical connection and configuration workflow

Follow the steps in order. Confirm the hardware in the operating system before selecting it in OBS or another routing application, and change one cable, mode or route at a time.

  • Step 1: Record a local baseline with a visible clap.
  • Step 2: Inspect frame-by-frame in a player/editor and measure picture versus chosen audio source.
  • Step 3: If offset is constant, apply the smallest correction to the correct source.
  • Step 4: If it grows, remove overload/unstable routing and repeat without a delay patch.
  • Step 5: Run a long representative recording and destination test.

Examples

A camera video path through HDMI capture may arrive later than a microphone connected directly to an interface. Measure that exact mode and apply a controlled offset only after both paths are stable.

If the first minute is aligned and minute thirty is not, delete the fixed-delay hypothesis. Test simpler device routes, host load and continuous USB/capture behavior.

Use OBS routing guide and pre-stream checklist for the adjacent checks.

Four common failure modes

Match the symptom below using a local recording, then change one variable per attempt.

  • Constant lip-sync offset: Measure and correct the relevant source once, then document the mode.
  • Drift grows over time: Investigate clocks, USB stability, dropped frames and redundant conversions.
  • Only stream is out of sync: Compare local file and destination; network/player behavior may be downstream.
  • Echo mistaken for sync: Remove duplicate microphone or program path before timing adjustments.

Prove the result

Record a visible clap and inspect the local file frame by frame. Apply a measured offset only when the same difference remains constant from the beginning to the end.

If the offset grows, remove duplicate paths and investigate dropped frames, device stability and host load. Repeat for the full intended duration before saving the correction.

Equipment for these connections

Start with the needed role. Check the complete profile before adding a product to your setup.

Sources