video-assembly（视频组装）

Video Assembly

Combines outputs from video and voice pipelines into final deliverable videos.

Two Assembly Modes

Mode 1: FFmpeg (Simple, Fast)

Best for: Concatenation, audio mixing, basic transitions, format conversion.

No dependencies beyond FFmpeg (pre-installed on most systems).

Mode 2: Remotion (Complex Compositions)

Best for: Animated captions, data visualizations, complex motion graphics, React-based templates.

Uses existing remotion-best-practices skill for guidance.

FFmpeg Operations

Concatenate Video Clips

# Create file list
echo "file 'clip1.mp4'" > filelist.txt
echo "file 'clip2.mp4'" >> filelist.txt
echo "file 'clip3.mp4'" >> filelist.txt

# Concatenate (same codec/resolution)
ffmpeg -f concat -safe 0 -i filelist.txt -c copy output.mp4

# Concatenate (different codecs/resolutions - re-encode)
ffmpeg -f concat -safe 0 -i filelist.txt -vf "scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(ow-iw)/2:(oh-ih)/2" -c:v libx264 -crf 19 output.mp4

Add Audio to Video

# Replace audio
ffmpeg -i video.mp4 -i audio.wav -c:v copy -c:a aac -map 0:v:0 -map 1:a:0 output.mp4

# Mix audio (keep original + add music)
ffmpeg -i video.mp4 -i music.mp3 -filter_complex "[0:a][1:a]amix=inputs=2:duration=first:dropout_transition=3" -c:v copy output.mp4

# Add audio with volume control
ffmpeg -i video.mp4 -i bgm.mp3 -filter_complex "[1:a]volume=0.3[bg];[0:a][bg]amix=inputs=2:duration=first" -c:v copy output.mp4

Add Subtitles

# Burn subtitles (SRT file)
ffmpeg -i video.mp4 -vf "subtitles=subs.srt:force_style='FontName=Arial,FontSize=24,PrimaryColour=&HFFFFFF,OutlineColour=&H000000,Outline=2'" -c:a copy output.mp4

# Burn subtitles (ASS file for styled)
ffmpeg -i video.mp4 -vf "ass=subs.ass" -c:a copy output.mp4

Transitions Between Clips

# Crossfade (2 second transition)
ffmpeg -i clip1.mp4 -i clip2.mp4 -filter_complex \
  "[0:v]trim=0:5,setpts=PTS-STARTPTS[v0]; \
   [1:v]trim=0:5,setpts=PTS-STARTPTS[v1]; \
   [v0][v1]xfade=transition=fade:duration=2:offset=3[outv]" \
  -map "[outv]" output.mp4

Extract/Manipulate Frames

# Extract frames
ffmpeg -i video.mp4 -vf "fps=1" frames/frame_%04d.png

# Create video from frames
ffmpeg -framerate 24 -i frames/frame_%04d.png -c:v libx264 -crf 19 -pix_fmt yuv420p output.mp4

# Speed up/slow down
ffmpeg -i video.mp4 -filter:v "setpts=0.5*PTS" -filter:a "atempo=2.0" fast.mp4   # 2x speed
ffmpeg -i video.mp4 -filter:v "setpts=2.0*PTS" -filter:a "atempo=0.5" slow.mp4   # 0.5x speed

Format Conversion

# MP4 (H.264) - universal compatibility
ffmpeg -i input.mov -c:v libx264 -crf 19 -preset medium -c:a aac -b:a 192k output.mp4

# WebM (VP9) - web delivery
ffmpeg -i input.mp4 -c:v libvpx-vp9 -crf 30 -b:v 0 -c:a libopus output.webm

# GIF (short clips)
ffmpeg -i input.mp4 -vf "fps=15,scale=480:-1:flags=lanczos" -loop 0 output.gif

Quality Presets

Assembly Pipeline

Standard Video

1. Gather clips from comfyui-video-pipeline outputs
2. Gather audio from comfyui-voice-pipeline outputs
3. Normalize audio levels (-16 LUFS for YouTube)
4. Trim/arrange clips to match script timing
5. Add transitions between clips (if multiple)
6. Add background music (optional, lower volume)
7. Add subtitles (if applicable)
8. Export in target format

Talking Head Video

1. Lip-synced video from comfyui-voice-pipeline
2. Add intro/outro (if applicable)
3. Add lower thirds or name cards
4. Add background music at -20dB below speech
5. Normalize speech to -16 LUFS
6. Export

Tutorial/Walkthrough

1. Screen recordings + talking head overlay
2. Use Remotion for animated annotations
3. Add chapter markers
4. Add subtitles (auto-generate from audio)
5. Export with chapters embedded

Audio Normalization

# Normalize to -16 LUFS (YouTube standard)
ffmpeg -i input.mp4 -af "loudnorm=I=-16:TP=-1.5:LRA=11" -c:v copy output.mp4

# Two-pass normalization (more accurate)
ffmpeg -i input.mp4 -af loudnorm=I=-16:TP=-1.5:LRA=11:print_format=json -f null -
# Read measured values, then:
ffmpeg -i input.mp4 -af "loudnorm=I=-16:TP=-1.5:LRA=11:measured_I=-20:measured_TP=-2:measured_LRA=8:measured_thresh=-30" -c:v copy output.mp4

Remotion Integration

For complex compositions, delegate to remotion-best-practices skill.

Use Remotion when:

• Animated text/captions needed
• Data visualizations in video
• Complex motion graphics
• Programmatic/template-based video generation
• React-based UI elements as video overlays

Pass to Remotion:

• Clip paths and timings
• Text content and styles
• Animation specifications
• Output format requirements

Output Checklist

Before marking assembly as complete:

• Audio levels normalized (-16 LUFS for YouTube)
• No audio clipping (peak below -1.5 dB)
• Resolution matches delivery target
• Frame rate is consistent throughout
• No black frames at cuts
• Subtitles are timed correctly (if applicable)
• File size is reasonable for delivery platform
• Metadata is correct (title in filename)

Reference

• references/workflows.md - Video output format settings
• remotion-best-practices skill - Complex composition guidance