Recording specs:
- Large-diaphragm condenser, pop filter, treated room; 24-bit/48kHz WAV, mono; consistent mic distance.
- Capture 30–60 minutes across sessions with short and long paragraphs, emotional lines, and breaths.
Legal/consent:
- Written voice-release assigning commercial audiobook rights, ID verification, payment terms, and usage scope.
Fine-tuning prompts:
- Provide style guide + 5 reference clips; target 150–160 WPM; commas 200–400ms, periods 600–800ms; mark breath spots and emotional cues.
Use ElevenLabs to clone and iterate.