⚡ Master Prompt Directive

📋 Copy Master Prompt

[ALIBABA QWEN3.8-LIVETRANSLATE REAL-TIME DUBBING & SUBTITLE ARCHITECTURE]

SYSTEM ENGINE: Alibaba Qwen3.8-LiveTranslate + Confucius4 Latency Orchestrator
LATENCY BENCHMARK: 2.3s Audio-to-Audio / 800ms Subtitle Streaming
STREAM PROTOCOL: WebSocket Bi-directional PCM Audio Stream (16kHz 16-bit Mono)

STREAMING ARCHITECTURE:
1. CLIENT AUDIO CAPTURE:
   - Voice Activity Detection (Silero VAD) buffers audio in 200ms chunks
   - Emits raw PCM packets over secure WebSocket connection
2. QWEN3.8-LIVETRANSLATE PIPELINE:
   - Streaming ASR: Incremental acoustic phonetic recognition
   - Neural Machine Translation: Context-aware streaming sentence completion
   - Zero-Shot Voice Cloning: Clones speaker timbre, emotional tone, and pitch contour
   - Streaming TTS Synthesizer: Emits synchronized target-language audio stream
3. CONFUCIUS4 SYNCHRONIZATION:
   - Dynamic time-stretching compensates for syllable length differences
   - Emits real-time word-level ASS/WebVTT subtitles with sub-100ms drift

SAMPLE SYSTEM DIRECTIVE:
"Translate incoming streaming English audio to Spanish. Match original speaker emotional intensity and vocal timbre. Maintain 2.3-second buffer limit. Deliver synchronized bilingual subtitle payloads."

NEGATIVE PROMPT:
robotic monotone voice, acoustic hallucination, audio clipping, dropped sentences, severe latency drift, garbled foreign pronunciation.