About

We are looking for native English speakers from specific regions to participate in a short but high-quality wake-up word recording project. The data will be used to train a high-performance English voice wake-up model. If you enjoy simple, well-structured audio tasks — this one is extremely quick! Project Overview You will record 80 short wake-up phrases, each containing only two words (e.g., “Hi BMW”, “Hey BMW”, etc.). Each phrase must be recorded one time per file, creating a total of 80 audio files. The entire task takes around 10–15 minutes, and the compensation is $25 USD. Eligible Regions / Native Speakers We specifically need native speakers from: United Kingdom United States Canada Australia Ireland New Zealand Germany (English speakers) Recording Requirements Environment Quiet, echo-free room No background noise (fans, AC, conversation, traffic, etc.) Preferably a professional or controlled space (voice lab, home studio, etc.) Equipment High-fidelity microphone (built-in laptop mics or phone mics are NOT acceptable) Performance Avoid swallowing sounds, slurring, mouth clicks, or noise Speak naturally, clearly, and consistently Audio Format Requirements Format: WAV Sampling Rate: 16,000 Hz Bit Depth: 16-bit Bit Rate: 256 kbps Channel: Mono One wake-up phrase per audio file File Naming Convention Use the following format for each file: [WakeWord][RecorderID][Gender][Age][Country]_[RecordingNumber].wav Examples: Hi BMW_001_M_25_Germany_01.wav Hey BMW_034_F_37_America_05.wav we must provide a text of each audio file (80 lines total). If your accent affects pronunciation, please note it briefly. Payment $25 USD upon submission of all 80 recordings Task duration: only 10–15 minutes only. we will relase payment after feedback!