Audio Producer

We’re looking for skilled voice talents to record 60 minutes of scripted audio that demonstrates a range of emotions (e.g., excitement, contempt, confusion) and speaking styles (e.g., narrative, whispered, reading to children).

These recordings will be used exclusively for internal text-to-speech (TTS) model training and evaluation.

No public broadcast

No commercial or advertising use

No voice cloning or likeness creation

• Duration: 1 hour of finished, usable speech (solo recordings only; no dialogues)
• Script: Fully provided (machine-translated text)
• Languages: Various — specify your native language when applying

Important: Scripts may contain minor translation imperfections. Please read them as written, but aim to maintain a natural, believable performance.

• Deliver genuine, natural, and expressive performances — not overly acted or robotic.
• Maintain clear pronunciation and a steady pace.
• Adjust tone and mood to fit the intended emotion or style described in each section.

• Format: WAV, stereo, 16-bit
• Sample rate: 48 kHz
• Audio quality:
• Clean recording environment (noise floor ≤ −60 dB)
• No reverb, echo, or room noise
• No clipping or distortion
• Minimal plosives or mouth noise
• Processing: Please submit raw recordings (no EQ, compression, denoising, or normalization).

Please submit the following with your proposal:

1. A short audition using the attached sample script (in the language you’re applying for).
2. A brief note about your recording setup (microphone, room treatment, etc.).
3. Your rate for one hour of finished, approved audio.