graph TD A[Input: Text Script] --> B[Initial Voice Generation] B --> C[Generate MP3] C --> D[Quality Check] D --> E{Quality OK?} E -->|No| F[Format Conversion] F --> G[Extract Timing Data] G --> H[Adjust Pacing] H --> B E -->|Yes| I[Final MP3 Output] subgraph SILO1[Voice Generation] B C end subgraph SILO2[Quality Verification] D E F end subgraph SILO3[Timing Optimization] G H end style SILO1 fill:#f9f,stroke:#333,stroke-width:2px style SILO2 fill:#bbf,stroke:#333,stroke-width:2px style SILO3 fill:#bfb,stroke:#333,stroke-width:2px