Today's session covered MapleLogic Studio — Major Session (Apr 23-24), SDH Subtitle Pipeline (biggest change), Veo Audio Prompt Improvement. Here's a summary of what was accomplished and what's still in progress.

• Replaced ALL guess-based SRT generation with Gemini SDH transcription
• Pipeline: Veo generates video → concat → upload to Gemini Files API → Gemini watches/listens → returns real SRT with dialogue + sound effects
• SDH prompt gives Gemini the actual script (expected dialogue per scene) so it doesn't mislabel speech as sound effects
• Consecutive same-sound merging added to prompt (avoids 4x [snoring] entries)
• Timestamp clamp: entries past video duration are trimmed automatically
• Env var: STUDIO_SRT_METHOD=gemini (default) or timing (fallback)