The Task
The client requested the collection of 750 unique audio recordings of children’s laughter, crying, and speech within one month. Each child could participate only once, eliminating the possibility of using the same actors multiple times. Strict quality and diversity requirements added complexity to the task.
The Solution
To ensure an efficient data collection process, we divided it into several stages:
-
- 01
-
Defining Data Requirements:
- Each child needed to provide five recordings: two speech samples, two laughter samples, and one crying sample.
- Each audio file had to be 20 seconds long.
- Participants’ ages ranged from 0 to 4 years, with specific quotas for each age group.
-
- 02
-
Data Collection Approach:
- A pilot phase using the Yandex.Toloka platform proved to be too slow.
- We switched to an in-house collection strategy, engaging parents through social media and childcare institutions.
- To verify the authenticity of the audio, we required submissions in video format to confirm that the laughter, crying, and speech genuinely belonged to a child and that there were no repeated participants.
-
- 03
-
Data Collection Approach:
- A pilot phase using the Yandex.Toloka platform proved to be too slow.
- We switched to an in-house collection strategy, engaging parents through social media and childcare institutions.
- To verify the authenticity of the audio, we required submissions in video format to confirm that the laughter, crying, and speech genuinely belonged to a child and that there were no repeated participants.
-
- 04
-
Data Verification and Processing:
- Initial validation by our team.
- Audio processing: Extracting sound from video files and segmenting recordings into 20-second clips.
- Final verification to ensure compliance with all requirements.
The Results
750 unique audio recordings were collected within the deadline.
The dataset met the required diversity and authenticity standards.
The client successfully validated the data and was fully satisfied with the outcome.
