Empty blocks of audio data #
|Problem ID||Manufacturer||Affected Firmware||Affected Hardware||Status|
|FL002||Frontier Labs||2.99||BAR||Minor problem|
Sometimes BARs start recording but no actual signal is recorded, until at some point it starts being captured. Even though there is no signal, some digital noise is captured, indicating that the sensor does think it is recording.
The simplest answer is that the microphone is either faulty or not attached. Normally we’d see some indication of the microphones health in the log files but in this case there’s no mention of anything about the microphone.
We’ve encountered two cases of this oddity, both with the same sensor:
- In the first case, the recording started capturing true signal at 00:07:00.953 into the recording
- In the second case, the recording started capturing true signal at 00:07:00.953 into the recording. Though in this case the first two 30 minute recordings captured no signal at all.
The fact that both recordings started capturing signal within one millisecond leads me to believe this not simply a microphone connection issue.
Minor Problem It most likely is just damaged hardware. This is mainly being documented in case we discover the behaviour again.
Status with vendor #
FL is not aware of the issue.
Effects of the problem on common tools #
Acoustic Workbench (Ecosounds, A2O) #
- Most recent update: There was a bug with our background noise calculation in AP.exe which has been resolved https://github.com/QutEcoacoustics/audio-analysis/issues/187
- We harvest and manipulate these files without error
- We should be doing better QA for incoming audio and then rejecting or trimming files that have periods of silence in them.