North American English Speech Recordings From 302 Speakers

Encompassing 201 hours of North American English speech data from 302 speakers. Recordings were made in quiet indoor environments using PC and Android mobile phone devices, capturing phrases and sentences across various scenes.

Use Cases

• Train an automatic speech recognition model on 201 hours of North American English speech from 302 speakers.
• Analyze acoustic variations in speech recorded from different device types, such as PC and Android mobile phone.
• Develop a speaker verification or diarization system using audio from 302 distinct North American speakers.
• Study the phonetic characteristics of English phrases and sentences recorded in quiet indoor environments.

Strengths

• Contains 201 hours of validated speech audio.
• Includes data from 302 distinct North American speakers.
• Recordings were captured in a controlled, quiet indoor environment.
• Uses multiple recording device types (PC and Android mobile phone).

Limitations

• The dataset is described as a sample of a larger paid dataset, so the available scope is limited.
• Specific recording dates, speaker demographics, and detailed transcription formats are not provided.
• The full dataset description is hosted externally, requiring a page visit for complete details.

Provenance

Source

Nexdata

Collection Method

Speech recordings collected from 302 speakers in quiet indoor settings using PC and Android mobile phones.

Freshness

The dataset card was last updated on 2025-04-24.

Geography

North America