Vanuatu - Githubissues

Current status on Vanuatu.

Here is an example metadata file for a given child :

Date 8th September 2018\
\
Participant says:\
Recording 1: Start time: 13.00\
Recording 1: End time: 18.39\
Recording 2: Start time:  06.52\
Recording 2: End time:  11.29\
\
On the recording there are 2 recordings \
Child21_1 = 1 hour 27 mins\
Child21_2 = 2 hours 42 mins\
\
Second recording was made on 4th & 5th October 2018\
\
Participant says:\
Recording 1: Start time: 16.00\
Recording 1: End time: 16.30\
Recording 2: Start time:  06.10\
Recording 2: End time:  09.39\
\
On the recording there are 3 recordings \
Child21_3 = 2 hour 24 mins\
Child21_4 = 3 hours 4 mins\
Child21_5 = 2 hours 37 mins

Clearly start/end time are not reliable, as they are not consistent with recordings durations at all, which was confirmed by Heidi. But worse than that, they are not always supplied (cf. the second recording period for which there are 3 files but only 2 recording start/end times)

Therefore, I don't think we can set start_time for the recordings at all, or it could be highly unreliable (not just inaccurate, but sometimes completely off and random)

Questions @alecristia :

Should we concatenate these files at all or just keep them split ? (i'd go for keeping them split)
Should we attempt to define a start_time, even though it is highly inaccurate and unreliable, even impossible to infer any value sometimes ?

LAAC-LSCP / datasets

Vanuatu #16