Speaker Detection

Automatic Speaker Detection

The platform automatically detects and distinguishes between different speakers in your recordings using advanced voice recognition technology.

How Speaker Detection Works:

1. Voice Analysis: The system analyzes voice characteristics (pitch, tone, speed)
2. Speaker Separation: Different speakers are identified and separated
3. Label Assignment: Each unique speaker receives a label (SPEAKER_1, SPEAKER_2, etc.)
4. Confidence Scoring: The system assigns confidence scores to speaker identifications

Speaker Labels

Speakers are automatically labeled during transcription:

• SPEAKER_1: First detected speaker
• SPEAKER_2: Second detected speaker
• SPEAKER_3: Third detected speaker
• And so on for additional speakers

These labels are temporary until you assign actual names to speakers (see Speaker Management section).

Speaker Identification

Assign real names to detected speakers:

1. Navigate to the Speakers section in the recording details
2. Click on a speaker label (e.g., SPEAKER_1)
3. Enter the speaker's name
4. Optionally add their email address
5. Save the assignment

Once assigned, the speaker's name will appear throughout the transcript instead of the generic label.

Speaker Merging

If the system incorrectly identifies one person as multiple speakers, you can merge them:

1. Go to the Speakers section
2. Select the speakers you want to merge
3. Click Merge Speakers
4. Choose which speaker name to keep
5. Confirm the merge

All transcript segments from the merged speakers will be combined under the selected name.

Speaker Statistics

View detailed statistics about each speaker's participation:

• Speech Duration: Total time each speaker spoke
• Percentage: Percentage of total conversation time
• Number of Segments: How many times each speaker contributed
• Average Segment Length: Average length of each speaker's contributions

These statistics help you understand participation levels and engagement in your meetings.

Speech Duration by Speaker

The platform tracks how much each speaker contributed to the conversation:

• Visual Charts: Bar charts and pie charts show speaker participation
• Time Breakdown: See exact minutes and seconds for each speaker
• Comparison: Easily compare participation levels between speakers
• Export: Export speaker statistics for reporting

This feature is particularly useful for:

• Meeting Analysis: Understanding who contributed most
• Training: Identifying team members who may need encouragement to speak
• Compliance: Tracking participation in important meetings