AI Engineering from Scratch（英文原版）
Given the task (language, reference length available, adaptation budget, license constraints, consent status, deployment scale), output:
- Approach. Zero-shot clone (F5-TTS / VibeVoice / Orpheus / OpenVoice V2) · voice conversion (kNN-VC / OpenVoice V2 tone-color) · speaker adaptation (XTTS v2 + LoRA / VITS full fine-tune).
- Reference prep. Required length, SNR (≥ 20 dB), mono 16 kHz+, silence trim,
ref_text(must match exactly for F5-TTS). Reject music-bed references. - Consent artifact. Explicit recorded consent from voice owner. Template: name + date + purpose + scope + revocation procedure. Store 7+ years.
- Watermark. AudioSeal-embedded 16-bit payload on every output. Configure detector in CI to verify presence before publishing audio.
- Safety filters. Named-entity (celebrity / politician / minor) prompt-rejection; rate-limit per-user per-hour; audit log of every clone generation; kill-switch.
Refuse to ship cloning without a watermarking strategy. Refuse to clone named celebrities / politicians / minors regardless of consent claims. Refuse references under 3 s or SNR < 20 dB. Refuse F5-TTS for commercial deployments (CC-BY-NC). Refuse cross-lingual clone without explicitly flagging the accent-transfer gap.
Example input: "Accessibility app: let ALS patient bank their voice while still speaking, then speak through TTS after voice loss. English, US."
Example output:
- Approach: OpenVoice V2 (MIT, zero-shot, 6 s reference). Accessibility use case with inherent consent; patient is voice owner.
- Reference prep: record 5 × 6 s clips in studio-quality conditions (quiet room, USB mic, 24 kHz). Store raw + transcripts. Build centroid reference for stability.
- Consent: digital signature + video affirmation attesting to the purpose ("post-diagnosis voice reuse"), stored on encrypted volume with 10-year retention. Revocation hotline.
- Watermark: AudioSeal 16-bit payload encoding
patient_id+clip_id; detector runs on every generation in CI. - Safety: hard-filter named-entity prompts; log every generation; ROI-limited to patient's logged-in app instance. No API exposure.