English Audio Data Collection for Asr Training by Us military veterans or military school students

We are seeking a skilled freelancer or team to manage and execute a large-scale audio data collection project. The collected data will be used to train an Automatic Speech Recognition (ASR) model.

Project Requirements:

Language & Dialect: Standard English from across the continental United States.

Number of Speakers: We aim to recruit a large number of speakers, with the goal of maximizing diversity and coverage.

Speaker Demographics:
- Age Range: 20s to 40s
- Gender Ratio: Approximately 85% male, 15% female
- Region: Diverse representation from across the continental United States
- Ethnicity Distribution: Target distribution is White American (50%), African American (30%), Hispanic American (10%), Asian American (10%).

Recording Environment: Recordings must be conducted in an indoor setting with minimal background noise to ensure high audio quality.

Recording Duration: Each speaker should provide a minimum of 2 hours of recorded audio.

Speech Segments: Each individual audio file should contain speech segments between 15 to 20 seconds in duration.

Total Target: The overall project goal is to collect a total of 1,000 hours of high-quality audio data.

Speech Content Type: Speakers will be required to read provided scripts (script content will be supplied by us).

Audio Format: All collected audio must be delivered in WAV format, Mono channel, 16kHz sample rate, and PCM 16-bit encoding.

The ideal candidate will have experience in managing data collection projects, recruiting participants based on specific demographic criteria, and ensuring adherence to strict technical specifications for audio recording and formatting. Strong organizational and project management skills are essential for coordinating the recruitment and recording process for a large number of speakers.