Job Type: Freelance | Remote | Short-term

Location: Global (Participants must be native Hong Kong Cantonese speakers)
Category: Data Collection

Objective:

We are looking for 1,000 native Hong Kong Cantonese speakers to provide key words speech data for training a TinyML model. The project involves recording specific voice commands using a mobile device in a quiet environment.

Project Specifications:

  • Speakers: 1,000 native Hong Kong Cantonese speakers.
  • Age Groups:
    • 18-30 years: 45% ±5%
    • 31-45 years: 45% ±5%
    • Above 45 years: 10% ±5%
  • Gender Balance: 50% male, 50% female.
  • Recording Device: Mobile phone.
  • Recording Environment: Quiet, noise-free background.
  • Commands:
    • 100 total voice commands (50 English, 25 Cantonese, 25 Mandarin).
    • Each command will be no longer than 5 characters.
    • Command folders will have English translations for Cantonese and Mandarin phrases.
  • Audio File Specifications: 16kHz, 16-bit, mono WAV files.
  • Speaking Speed: Normal conversational speed with audio clip lengths varying within ±10%.

Requirements:

  • Must be a native speaker of Hong Kong Cantonese.
  • Comfortable with recording voice commands in a quiet environment.
  • Ability to follow precise instructions for audio quality and recording speed.
  • Must have access to a mobile phone for the recording.

Proposed Rate:

Please include your proposed rate per recording in your application form.

Compensation:

Final compensation details will be based on the number of valid recordings and the agreed rate. Payment will be competitive.

How to Apply:

If you're interested in participating, please fill out the application form below:

Google Form Link