1. Record the test signal
Send this page through the same display, capture, mixer, encoder, and return path you want to check. Record the final output with its audio.
Stopped
- Record at the same frame rate and through the same output path used in production.
- Step through the recording and count from the white flash frame to the audio transient.
- Repeat at least three times. A changing result suggests drift or buffering, not one fixed offset.
2. Enter the measured offset
Describe what the recording shows. The worksheet uses exact fractional rates for 23.976, 29.97, and 59.94 conversions.
Audio is early. Add about 66.73 ms of delay to the audio path, then record and measure again.
Repeat checks
No comparison saved yet.
3. Apply, record, and verify
Audio early
Delay audio by the measured magnitude. This matches the normal positive audio-delay workflow documented by systems such as vMix. Verify the encoded or recorded output, not only local monitoring.
Audio late
Delay video by the measured magnitude if the system supports it, or reduce an existing audio delay. Do not enter a negative audio value unless the exact platform documents that behavior.
Fixed vs variable: a stable repeated number can be corrected with a static delay. If the measured relationship changes over time, investigate clocks, sample rates, capture timing, buffering, load, or network behavior instead.
Primary references
- EBU R 037 describes concurrent flash/blip markers and broadcast sound-to-picture timing limits.
- vMix: set audio delay documents adding delay in milliseconds when audio arrives before video.
- OBS source API reference confirms that OBS sources expose a sync-offset control; verify actual UI behavior in the installed OBS version.