English Voice Recording Contributors
Budget: $15 – $25 USD
I’m expanding a multi-speaker speech corpus that will be used to train both Text-to-Speech and Automatic Speech Recognition models for better naturalness, prosody, and accent robustness.
What I need from you is straightforward: read the script I supply, record it clearly in English, and return the raw, high-quality audio files.
The work is open to any location; what matters is that you have native-level or expert fluency and can deliver a clean, confident American accent for this round of data collection. If you already have some experience recording voice for AI or similar projects, that’s a plus, but not a strict requirement—consistent diction and the ability to follow technical guidelines matter most.
You will receive:
• A script with sentence prompts and basic pronunciation notes.
• File-naming and audio-format instructions (16-bit WAV, mono, 48 kHz).
• A short reference clip so you can match energy and pacing.
I will check each submission for:
• Accurate script adherence (no skipped or altered words).
• Even volume, no clipping or background noise.
• Natural, conversational delivery—no robotic or rushed reading.
Note- Using AI is prohibited.
Once your sample passes review, we can schedule the full batch of lines and set milestones for delivery. Looking forward to adding your voice to the dataset!
What I need from you is straightforward: read the script I supply, record it clearly in English, and return the raw, high-quality audio files.
The work is open to any location; what matters is that you have native-level or expert fluency and can deliver a clean, confident American accent for this round of data collection. If you already have some experience recording voice for AI or similar projects, that’s a plus, but not a strict requirement—consistent diction and the ability to follow technical guidelines matter most.
You will receive:
• A script with sentence prompts and basic pronunciation notes.
• File-naming and audio-format instructions (16-bit WAV, mono, 48 kHz).
• A short reference clip so you can match energy and pacing.
I will check each submission for:
• Accurate script adherence (no skipped or altered words).
• Even volume, no clipping or background noise.
• Natural, conversational delivery—no robotic or rushed reading.
Note- Using AI is prohibited.
Once your sample passes review, we can schedule the full batch of lines and set milestones for delivery. Looking forward to adding your voice to the dataset!