Skip to main content

Voice Recording

Voice Recording start screen

Voice recording activities allow participants to record audio, which can be used for speech-based assessments or qualitative data collection. Recordings are stored as MP3 files.

ActivitySpec: lamp.recording

Configuration​

  1. Navigate to the Activities tab and click + Add.
  2. Select Voice Recording.
  3. Select the target group.
  4. Title the activity.
  5. Save.

The following labels can be customized by clicking the pencil icon under "Activity Settings":

  • Record label — Text displayed on the record button (e.g., "Press to record").
  • Re-record label — Text displayed on the re-record button.

This is useful for providing specific instructions on what to record or including a participant safety disclaimer.

Usage​

The participant sees a prompt and records up to 2 minutes of audio. They can pause and resume the recording. If unsatisfied, they can press "Clear" and re-record. When satisfied, they press "Upload" to submit.

Ready to recordRecording in progressPlayback ready
Ready — Press microphone to startRecording — Stop or pausePlayback — Review before uploading

Use Cases​

  • Speech pattern analysis in relation to mood or symptom changes
  • Qualitative data collection (participant narratives)
  • Pronunciation or language assessments
  • Clinical speech assessments

Data​

temporal_slices​

FieldDescription
temporal_slices[0].valueThe recorded audio as a base64-encoded string (MP3 format)

Example​

{
  "timestamp": 1649859300000,
  "activity": "voice_recording_id",
  "duration": 30000,
  "static_data": {},
  "temporal_slices": [
    {
      "item": null,
      "value": "<base64-encoded audio data>",
      "type": null,
      "duration": 30000,
      "level": null
    }
  ]
}

Cortex Features​

voice_survey (raw). No primary or secondary Cortex features process voice data. For custom speech analysis, access raw recordings via cortex.raw.voice_survey.

View in Portal | Python SDK | API Reference