The Zoom R16 recorded audio drifts by ~3.6ms/min. This means that if you use the R16 to record people talking, and try to synchronize it with video, the Zoom recording will lag ~15ms in only 5 minutes. That will result in seeing people's lips move slightly before you hear them speak.
Does it matter? Yes. The timing drift is audible by human ears almost immediately, with it turning into a delay type effect within minutes. This could be corrected with software if the drift were consistent, but it isn't. There are some periods when it seems to lag twice as slow, then it catches back up to standard drift times.
This test was done using 3 hours recording time because I didn't expect to see any significant data before at least 45 minutes. But if you want to repeat the test using your video gear, you really only need to record for 15 minutes before you see easy to measure drift.
| Time | Difference | Drift | |
|---|---|---|---|
| (Minutes) | (Samples) | (Samples) | (Milliseconds) |
| 0 | 0 | 0 | 0 |
| 15 | 1,974 | 1,974 | 41.1 |
| 30 | 3,929 | 1,955 | 40.7 |
| 45 | 5,866 | 1,937 | 40.4 |
| 60 | 10,986 | 5,120 | 106.7 |
| 75 | 12,888 | 1,902 | 39.6 |
| 90 | 14,777 | 1,889 | 39.4 |
| 105 | 18,259 | 3,482 | 72.5 |
| 120 | 20,134 | 1,875 | 39.0 |
| 135 | 22,003 | 1,869 | 38.9 |
| 150 | 23,867 | 1,864 | 38.8 |
| 165 | 28.933 | 5,066 | 105.5 |
| 180 | 30,792 | 1,859 | 38.7 |