graph TD A[Start: Raw Video Input] --> PROC[Transcription Processor] PROC --> B[Extract MP3 Audio] B --> C[Process via Speech-to-Text] C --> D[Generate Timestamps] D --> E[Format as JSON] E --> F[Clean Transcript] F --> G[Quality Check] G --> H[Output JSON Transcript] H --> END[End: Timestamped Transcript] subgraph Error Handling C -- Error --> ERR1[Log Error] D -- Error --> ERR2[Log Error] E -- Error --> ERR3[Log Error] end subgraph Quality Controls G -- Fail --> QC1[Manual Review] QC1 --> F end subgraph Data Format E -- Validate --> VAL1[Check JSON Schema] VAL1 -- Invalid --> E end