feat: Improve podcast maker UX and fix bugs

Frontend:
- Add progress modals with educational content for analysis and voice cloning
- Improve tab navigation in AnalysisPanel (combine Titles, Hook, CTA into one tab)
- Fix tab styling to make inactive tabs visible
- Fix avatar 'Make Presentable' not updating preview (blob URL handling)
- Improve mobile responsiveness for avatar tabs
- Clean up verbose console logging (AnalysisPanel, demoMode, RobustCamera)
- Add sequential progress messages instead of cycling

Backend:
- Fix 'Depends object has no attribute get' error in auth and image editing
- Use get_session_for_user instead of get_db outside FastAPI DI context
- Reduce WARNING logs to DEBUG in audio handler
- Add proper emphasis boolean handling in script generation
- Add missing fields to PodcastScene and PodcastSceneLine models
- Fix voice cloning cost estimate display issue

backend/api/podcast/handlers/audio.py

@@ -391,9 +391,9 @@ async def serve_podcast_audio(
         raise HTTPException(status_code=400, detail="Invalid filename")
     
     user_id = require_authenticated_user(current_user)
-    logger.warning(f"[Podcast] serve_podcast_audio called: user_id={user_id}, filename={filename}")
+    logger.debug(f"[Podcast] serve_podcast_audio called: user_id={user_id}, filename={filename}")
     audio_path = _resolve_podcast_media_file(filename, "audio", user_id)
-    logger.warning(f"[Podcast] Resolved audio path: {audio_path}")
+    logger.debug(f"[Podcast] Resolved audio path: {audio_path}")
     
     return FileResponse(audio_path, media_type="audio/mpeg")