Italian Speech Recordings From Mobile Phone Guidance

A collection of 9.8 hours of Italian speech data collected from 351 native speakers in quiet environments. Each speaker contributed 50 sentences, covering categories like in-car scenes, smart home, and speech assistant interactions.

Use Cases

• Train an automatic speech recognition (ASR) model on Italian speech data from 351 speakers.
• Analyze speech patterns across different recorded categories such as in-car scenes and smart home commands.
• Develop a speech synthesis model using the 50-sentence samples per speaker for voice cloning applications.
• Build a multi-category classifier for spoken commands using the dataset's scene labels (e.g., in-car, smart home).

Strengths

• Data collected from 351 Italian native speakers, providing a substantial speaker pool.
• 9.8 hours of validated audio volume for model training.
• Recordings made in quiet environments to ensure audio clarity.
• Content covers multiple practical categories including in-car and smart home scenes.

Limitations

• The dataset is described as a sample of a larger paid dataset, implying limited scope.
• Specific column definitions, file formats, and exact sentence counts are unknown.
• Potential for limited demographic or geographic diversity within the 351-speaker pool.

Provenance

Source

Nexdata

Collection Method

Collected via mobile phone guiding from Italian native speakers in quiet environments.

Freshness

Last updated on 2025-04-23.

Geography

Italy (inferred from data collection from Italian native speakers).