Simple projects require native American speakers to participate in the recording -- 2
Budget: $10 – $30 USD
1. Dataset Introduction
This database collects interactive commands and wake-up words completed by adults using English at different speech rates, and provides training data for improving speech recognition.
2. Dataset description
Sample size: 200 people
Recording requirements: Normal speaking speed and slightly fast speaking speed, 4 times each, 8 times in total, 368 words/person in total
Sample attributes: male to female ratio 1:1, age 18-50, mainly 20-35
Collection equipment: mobile phone equipment
Collection scene: indoor quiet environment
Collecting text: the wake-up word for mainstream artificial intelligence manufacturers
Geographical requirements: Accent: American, clear and standard everyday English pronunciation (especially exclude Indian and Southeast Asian accents)
This database collects interactive commands and wake-up words completed by adults using English at different speech rates, and provides training data for improving speech recognition.
2. Dataset description
Sample size: 200 people
Recording requirements: Normal speaking speed and slightly fast speaking speed, 4 times each, 8 times in total, 368 words/person in total
Sample attributes: male to female ratio 1:1, age 18-50, mainly 20-35
Collection equipment: mobile phone equipment
Collection scene: indoor quiet environment
Collecting text: the wake-up word for mainstream artificial intelligence manufacturers
Geographical requirements: Accent: American, clear and standard everyday English pronunciation (especially exclude Indian and Southeast Asian accents)