This plugin detects if the user is engaged in a conversation or not. It follows a duty cycle of 1 minute audio collection, 3 minutes pause. It does not store the raw audio (it's disabled). This plugin was developed as a collaboration between Cornell and Dartmouth College for the StudentLife project and later extended to support the converstations start/end at the Center for Ubiquitous Computing at the University of Oulu.
Parameters adjustable on the dashboard and client:
- status_plugin_studentlife_audio: (boolean) activate/deactivate plugin
ACTION_AWARE_PLUGIN_CONVERSATIONS_START Broadcasted when we detect the start of a conversation
ACTION_AWARE_PLUGIN_CONVERSATIONS_STOP Broadcasted when we detect the end of a conversation
content://com.aware.plugin.studentlife.audio_final.provider.audio_final/plugin_studentlife_audio_android
| Field | Type | Description |
|---|---|---|
| _id | INTEGER | primary key auto-incremented |
| timestamp | REAL | unix timestamp in milliseconds of sample |
| device_id | TEXT | AWARE device ID |
| datatype | INTEGER | identifier for what is in blob_feature: 0 = voice/noise and volume, 1 = audio features, 2 = conversations. |
| double_energy | REAL | amplitude of audio sample (L2-norm of the audio frame) |
| inference | INTEGER | 0 = silence, 1 = noise, 2 = voice, 3 = unknown |
| blob_feature | BLOB | audio raw sample. not stored by default (would be huge!) |
| double_convo_start | REAL | UNIX timestamp of beginning of sample (if datatype = 2) |
| double_convo_end | REAL | UNIX timestamp of ending of sample (if datatype = 2) |