---

name: Video Assembly tier: 3 load_policy: task-specific description: Assemble final video from audio, images, and subtitles version: 1.0.0 parent_skill: production-operations

Video Assembly Skill

The Visual Wrapper for the Audio Journey

This skill handles assembling the final video from mastered audio, scene images, and VTT subtitles.

---

Purpose

Create YouTube-ready video files that complement the hypnotic audio experience.

---

Video Standards

---

Input Requirements

---

Scene Image Generation

Primary Method: Stable Diffusion (Default)

Copypython3 scripts/core/generate_scene_images.py sessions/{session}/

Alternative: Midjourney Prompts

Copypython3 scripts/core/generate_scene_images.py sessions/{session}/ --midjourney-only

Alternative: Stock Images

Copypython3 scripts/core/generate_scene_images.py sessions/{session}/ --method stock

Image Specifications

---

VTT Subtitle Generation

Copypython3 scripts/ai/vtt_generator.py sessions/{session}

VTT Format

CopyWEBVTT
Kind: captions
Language: en

1
00:00:00.000 --> 00:00:05.500
Welcome to this healing journey.

2
00:00:06.000 --> 00:00:12.000
Find a comfortable position and
allow your eyes to close.

Subtitle Guidelines

---

Video Assembly Command

Copypython3 scripts/core/assemble_session_video.py sessions/{session}/

This automatically:

• Sequences images based on script sections
• Adds cross-fade transitions
• Syncs subtitles to audio
• Outputs to output/video/session_final.mp4

---

Manual FFmpeg Assembly

For custom control:

Copy# Create video from images with audio
ffmpeg -y \
  -framerate 1/10 \
  -pattern_type glob -i 'images/uploaded/*.png' \
  -i output/{session}_MASTER.mp3 \
  -c:v libx264 -r 24 -pix_fmt yuv420p \
  -c:a aac -b:a 320k \
  -shortest \
  output/video/session_final.mp4

---

Image-to-Section Mapping

Images should correspond to script sections:

---

Transition Effects

---

Output Files

---

Video Overlay Generation

Generate supporting graphics:

Copypython3 scripts/core/generate_video_images.py sessions/{session}/ --all

Creates:

• title_card.png - Video intro screen
• sections/section_*.png - Chapter transitions
• outro.png - End screen
• social_preview.png - Social sharing

---

Quality Verification

After assembly:

Copy# Check video properties
ffprobe -v error -show_format -show_streams output/video/session_final.mp4

# Play with VLC to verify sync
vlc output/video/session_final.mp4

Quality Checklist

• Resolution is 1920x1080
• Frame rate is 24 fps
• Audio syncs with subtitles
• Transitions are smooth
• No visible artifacts
• Duration matches audio
• File size reasonable (<2GB for 30 min)

---

Troubleshooting

---

Integration with Pipeline

Before (dependencies):

• Audio mastered ({session}_MASTER.mp3)
• Scene images ready (images/uploaded/)
• VTT subtitles generated (output/subtitles.vtt)

After (next steps):

• YouTube packaging

---

Related Resources

• Skill: tier3-production/audio-mixing/ (input)
• Skill: tier3-production/youtube-packaging/ (next step)
• Doc: docs/STOCK_IMAGE_SOP.md
• Script: scripts/core/assemble_session_video.py