Rajbharti06/Ultimate-Video-Editing-Skills

Ultimate Video Editor — God-Level Skill

You are a 50+ year veteran editor with mastery of every editing discipline. You think in cuts, breathe in keyframes, and dream in color spaces. Every decision is intentional. Every frame earns its place. You can color grade a movie, animate like Disney, sound design like Skywalker Sound, and edit with the invisible precision of a Thelma Schoonmaker.

---

WHEN TO ACTIVATE

This skill activates for ANY request involving:

• Video editing, cutting, trimming, arranging footage
• Color grading, color correction, look development
• Animation, motion graphics, kinetic typography
• Sound design, audio mixing, music editing
• Subtitle/caption creation and styling
• Transitions, compositing, overlays, VFX
• Speed ramping, slow motion, timelapse
• Export and delivery for any platform
• Automated/AI-assisted editing workflows

Decision Router:

1. Editing raw footage? → Parts 1-3 (Mindset, Narrative, Cuts)
2. Color work? → Part 5 + color-science/color-grading-master.md
3. Animation/motion? → Part 8 + references/motion-design.md
4. Sound work? → Part 6 + sound-design/sfx-guide.md + sound-design/audio-mixing.md
5. Subtitles/captions? → Part 7 + automation/auto-captions.md
6. Pro effects (glow, velocity, flash)? → ffmpeg-recipes/pro-effects.md
7. Text overlays & typography? → ffmpeg-recipes/text-overlays.md
8. Beat-synced editing? → automation/beat-sync.md
9. Batch processing? → automation/batch-pipeline.md
10. Social media optimization? → social-media/viral-editing.md + social-media/platform-specs.md
11. Programmatic video? → programmatic/remotion-guide.md + programmatic/moviepy-guide.md
12. AI-powered workflow? → ai-tools/ai-workflow.md
13. Full production? → Part 12 (Pro Workflow) → route to each part

---

PART 1: THE EDITOR'S MINDSET

The Three Laws (Non-Negotiable)

1. Story is king. Every cut serves the narrative. If it doesn't advance story, emotion, or rhythm — kill it.
2. Audio leads, visuals follow. Cut decisions come from speech boundaries, music beats, and silence gaps. Drill into visuals only at decision points. (Source: video-use)
3. Verify before presenting. Watch your own output at 1x speed. If you wouldn't ship it, don't show it.

The Invisible Cut Principle

The best edit is one the viewer never notices. Cuts should feel like natural eye movements. The viewer should be carried by story, not distracted by technique.

Decision Framework (Ask Before Every Cut)

1. What does the viewer FEEL at this moment?
2. What information do they NEED?
3. What's the RHYTHM of this section?
4. What comes NEXT and how do we bridge to it?

Pro Heuristics

The 6-Second Rule — Every 6 seconds, something should change: cut, camera move, new info, music shift, or visual change. This keeps attention in the short-form age.

The Breath Principle — Great editing breathes. Tension needs release. Fast sections need slow follow-ups. Silence after noise is powerful. Don't fill every second.

The Kuleshov Effect — Meaning is created by juxtaposition. A face + food = hunger. A face + coffin = grief. Same face, different meaning. Use this for storytelling power.

Sound Leads Image — The ear processes faster than the eye. Bring in audio 2-4 frames before the visual cut. This is why J-cuts feel so natural.

The 180-Degree Rule — Keep camera on one side of the action line. Crossing disorients the viewer. Bridge with a neutral shot (overhead/frontal) when you must cross.

Continuity of Energy — Match energy levels across cuts. Don't jump from calm to chaos without a bridge shot or audio transition. The eye and ear need preparation.

---

PART 2: NARRATIVE STRUCTURE

Three-Act Structure (Universal)

Act	Share	Purpose	Pacing
Act 1 — Setup	~25%	Hook + establish context, build anticipation	Medium-fast, 2-3s shots
Act 2 — Development	~50%	Main content, escalating tension/interest	Varied, 3-5s average
Act 3 — Resolution	~25%	Climax + emotional payoff + closure	Slow build → peak → breathe

Shot Pacing Rules

• Golden rhythm: 3-4 seconds per shot average
• Fast cuts (<2s): montage, energy, excitement, music-synced
• Medium holds (3-6s): dialogue, establishing, narrative
• Long holds (>8s): emotion, landscape, tension, breathing room
• Vary rhythm: alternate fast/slow. Monotonous pacing = death of engagement

Content Mix (for vlogs/docs — adjust per material)

Food/action 30-40% + Scenery 30% + People/interaction 15% + B-roll/transitions 15%

Deep dive: director/narrative-structure.md

---

PART 3: THE CUT — TYPES & WHEN TO USE

Every Cut Type a Pro Editor Knows

Cut Type	When to Use	Technique
Hard cut	Default. Clean transitions between related shots	Cut on action or dialogue pause
J-cut	Audio leads visual. Build anticipation	Next scene's audio starts 0.5-2s before visual
L-cut	Visual leads audio. Smooth dialogue flow	Current audio continues over next visual
Jump cut	Energy, time compression, social media style	Same framing, skip time. Intentional only
Match cut	Poetic connection between scenes	Match shape, movement, or color across cut
Smash cut	Shock, comedy, dramatic contrast	Abrupt shift in tone/energy
Cutaway	Context, humor, emphasis	Brief insert shot (1-3s), then return
Cross-cut	Parallel action, tension building	Alternate between simultaneous events
Montage	Time compression, training sequences, travel	Rapid succession, beat-synced
Invisible cut	Hidden transition, long-take illusion	Whip pan, object pass, rack focus hide
Axial cut	Emphasis, dramatic zoom	Same angle, different focal length
Freeze frame	Emphasis, intro, comedy, ending	Hold frame + text/narration

Hard Rules for Cutting (Source: video-use)

1. Never cut inside a word. Snap to word boundaries from transcript.
2. Cut on action. Mid-motion cuts feel invisible.
3. Cut on beat. Music-synced cuts on downbeats.
4. Pad every cut edge. 30-200ms padding absorbs ASR timestamp drift (50-100ms typical).
5. 30ms audio fades at every cut boundary. Prevents pops/clicks.

   afade=t=in:st=0:d=0.03,afade=t=out:st={dur-0.03}:d=0.03
   

6. Per-segment extract → lossless concat. Don't double-encode when adding overlays.
7. Word-level verbatim ASR only. Never SRT/phrase mode (loses sub-second gap data).
8. Cache transcripts per source. Never re-transcribe unless the source file changed.

What to Cut

• Filler words: "um", "uh", "like", "you know", "so", "basically", "right", "I mean"
• Dead air / silence gaps > 0.8s (keep intentional pauses)
• False starts, repetitions, verbal stumbles
• Self-corrections ("I mean", "wait no", "actually")
• Off-topic tangents (unless charming or character-building)

---

PART 4: TRANSITIONS

Rule: Less is More

90% of transitions should be hard cuts. Fancy transitions are seasoning, not the meal.

Complete Transition Guide

Transition	Duration	When to Use	Energy
Hard cut	0ms	Default. Same energy, related content	Neutral
Cross dissolve	300-800ms	Time passage, dream/memory, gentle shift	Soft
Fade to black	500-1500ms	Chapter end, major time jump, finality	Low
Dip to white	300-600ms	Flashback, heavenly, high-energy	High
Wipe	400-800ms	Playful, retro, geographic movement	Medium
Zoom transition	200-500ms	Energy burst, social media, travel	High
Whip pan	150-300ms	Location change, comedy, energy	Very high
Morph/match	400-1000ms	Artistic, shape/color matching	Medium
Glitch	100-300ms	Tech, gaming, modern, disruption	High
Slide/push	300-600ms	Clean, corporate, organized	Medium
Iris	300-600ms	Vintage, spotlight, dramatic reveal	Medium
Light leak	200-500ms	Film aesthetic, dreamy, warm	Soft
Luma/color key	400-800ms	Creative, music video, experimental	Variable

FFmpeg commands for every transition: patterns/transitions.md

---

PART 5: COLOR GRADING

Color Tells Emotion — The Master Table

Emotion/Genre	Temperature	Saturation	Contrast	Blacks	Highlights
Warm/happy	+15-30 warm	+10-20%	Medium	Natural	Warm
Cold/tense	-15-30 cool	-10-20%	High	Crushed	Cool
Nostalgic/vintage	+10 warm	-15-25%	Low-medium	Lifted	Rolled off
Cinematic (Hollywood)	Slight warm	Normal	Medium-high	Slightly lifted	Rolled off
Documentary	Neutral	Normal	Natural	Natural	Natural
Horror/thriller	Cool blue-green	Desaturated	Very high	Crushed	Harsh
Romance	Warm golden	+10%	Low-medium	Lifted	Soft glow
Sci-fi	Cool blue/teal	Selective	High	Deep	Harsh
Wes Anderson	Pastel warm	Highly saturated	Medium	Lifted	Warm
David Fincher	Desaturated teal	Low	High	Crushed green	Cool
Michael Bay	Orange/teal split	High	Very high	Deep	Blown warm
Blade Runner	Neon + dark	Selective neon	Very high	Crushed	Neon bleed

FFmpeg Color Grading Recipes

# Warm Cinematic (Hollywood blockbuster)
-vf "eq=brightness=0.03:contrast=1.1:saturation=1.15,colorbalance=rs=0.05:gs=0.02:bs=-0.03,curves=m='0/0.04 0.5/0.5 1/0.96'"

# Cool Moody (thriller/drama)
-vf "eq=contrast=1.2:saturation=0.85,colorbalance=rs=-0.05:gs=0.0:bs=0.08,curves=m='0/0.05 0.5/0.45 1/0.95'"

# Vintage Film (lifted blacks, rolled highlights, warm)
-vf "curves=m='0/0.06 0.25/0.22 0.5/0.5 0.75/0.78 1/0.94':r='0/0.08 1/0.95':b='0/0.04 1/0.92',eq=saturation=0.85"

# Wes Anderson (pastel, warm, saturated)
-vf "eq=saturation=1.3:brightness=0.05,colorbalance=rs=0.08:gs=0.04:bs=-0.02,curves=m='0/0.08 0.5/0.52 1/0.95'"

# Teal & Orange (blockbuster look)
-vf "colorbalance=rs=0.1:gs=-0.05:bs=-0.1:rh=0.05:gh=-0.03:bh=-0.08:ms=-0.05:bs=0.1,eq=contrast=1.15:saturation=1.2"

# Black & White Cinematic
-vf "hue=s=0,eq=contrast=1.3:brightness=0.02,curves=m='0/0.03 0.3/0.25 0.7/0.75 1/0.97'"

# Day for Night
-vf "eq=brightness=-0.15:contrast=1.2:saturation=0.6,colorbalance=rs=-0.1:gs=-0.05:bs=0.15"

# Bleach Bypass (desaturated, high contrast — war film look)
-vf "eq=contrast=1.4:saturation=0.5:brightness=-0.05,curves=m='0/0 0.25/0.15 0.75/0.85 1/1'"

# Cross-processed (fashion/music video)
-vf "curves=r='0/0.1 0.5/0.6 1/0.9':g='0/0 0.5/0.45 1/1':b='0/0.15 0.5/0.5 1/0.85'"

# Golden Hour Simulation
-vf "colorbalance=rs=0.12:gs=0.06:bs=-0.08,eq=brightness=0.04:saturation=1.2,curves=m='0/0.02 1/0.98'"

Scene Matching Workflow

1. Sample reference — extract a frame from the "hero" shot
2. Analyze — measure average RGB, contrast ratio, saturation
3. Apply correction — use colorbalance + eq + curves to match
4. Verify — A/B compare corrected shot against reference

# Extract color analysis from a reference frame
ffmpeg -i reference.mp4 -vf "select=eq(n\,0)" -frames:v 1 ref_frame.png
ffprobe -f lavfi -i "movie=ref_frame.png,signalstats" -show_entries frame_tags -of csv 2>&1 | head -20

Full color science deep dive: color-science/color-grading-master.md

---

PART 6: SOUND DESIGN

Audio Hierarchy (Sacred Order)

1. Dialogue — always intelligible, -12 to -6 dBFS peak, -16 LUFS integrated
2. Music — supports emotion, -18 to -24 dBFS under dialogue, -12 alone
3. Sound effects — punctuation and emphasis, -12 to -15 dBFS
4. Ambience — world-building atmosphere, -30 to -24 dBFS

Music Editing Rules

• Cut on the beat. Always. No exceptions.
• Music dips 6-12dB during dialogue (sidechain ducking)
• Fade in: 500ms-2s. Fade out: 1-3s. Never abrupt stops.
• Match music energy curve to visual energy curve
• Key/chord changes are natural scene change points
• BPM guides cut rhythm: 120 BPM = one beat every 500ms

Audio Processing Chain (FFmpeg)

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

# Remove background noise + normalize
ffmpeg -i input.mp4 -af "afftdn=nf=-25,loudnorm=I=-16:TP=-1.5" output.mp4

# Music ducking (auto-lower music when voice is present)
ffmpeg -i voice.wav -i music.wav -filter_complex \
  "[1]sidechaincompress=threshold=0.02:ratio=8:attack=200:release=1000[duck];[0][duck]amix=inputs=2" output.wav

# De-essing (reduce sibilance)
ffmpeg -i input.mp4 -af "equalizer=f=6000:t=q:w=2:g=-6" output.mp4

# Add room tone / ambience under cuts
ffmpeg -i main.mp4 -i room_tone.wav -filter_complex "[1]volume=0.15[amb];[0:a][amb]amix=inputs=2" output.mp4

# Audio fade at cut points (MANDATORY — prevents pops)
-af "afade=t=in:st=0:d=0.03,afade=t=out:st={duration-0.03}:d=0.03"

Full audio processing reference: ffmpeg-recipes/audio-processing.md

---

PART 7: SUBTITLES & CAPTIONS

Subtitle Rules

1. Max 2 lines, ~42 characters per line
2. Display time: minimum 1s, maximum 7s
3. Reading speed: 15-20 characters/second
4. Position: bottom-center safe area (10% from bottom edge)
5. Font: Bold sans-serif, white with black outline or dark semi-transparent background

5 Subtitle Styles

Style	Look	Best For
Classic	White text, 2px black outline	Documentary, corporate, interviews
Modern	White on semi-transparent dark bar	YouTube, tutorials, how-to
Social	2-word UPPERCASE animated chunks, colored highlight	TikTok, Reels, Shorts
Cinematic	Thin light sans-serif, in letterbox area	Film, premium content
Karaoke	Word-by-word color highlight, synced to audio	Music videos, lyric content

FFmpeg Subtitle Recipes

# Classic style (white + black outline)
ffmpeg -i input.mp4 -vf "subtitles=subs.srt:force_style='FontName=Arial,FontSize=24,PrimaryColour=&H00FFFFFF,OutlineColour=&H00000000,Outline=2,Bold=1,Alignment=2,MarginV=30'" output.mp4

# Modern style (dark background bar)
ffmpeg -i input.mp4 -vf "subtitles=subs.srt:force_style='FontName=Helvetica,FontSize=22,PrimaryColour=&H00FFFFFF,BackColour=&H80000000,BorderStyle=4,Outline=0,Shadow=0,MarginV=25'" output.mp4

# Cinematic (thin, elegant)
ffmpeg -i input.mp4 -vf "subtitles=subs.srt:force_style='FontName=Futura,FontSize=20,PrimaryColour=&H00E0E0E0,OutlineColour=&H40000000,Outline=1,Bold=0,Alignment=2,MarginV=15'" output.mp4

Social/karaoke styles require programmatic rendering — use video-use or Remotion skills.

CRITICAL: Subtitles are ALWAYS applied LAST in the filter chain. Otherwise overlays hide captions.

---

PART 8: MOTION DESIGN & ANIMATION

The Three Pillars (Source: LottieFiles/motion-design-skill)

Pillar	Question	Drives
Emotional Intent	What should the viewer FEEL?	Easing, timing, amplitude
Visual Narrative	What's the micro-story?	Setup → Action → Resolution
Motion Craft	How do we make it believable?	Physics, secondary motion, arcs

Three Motion Layers (flat animation = missing layers)

• Primary: Main action the viewer's eye follows
• Secondary: Supporting richness (shadows shifting, icons reacting)
• Ambient: Background life (gradients pulsing, particles drifting)

8-Step Animation Checklist

1. Emotional target? — joy, calm, urgency, elegance
2. Motion Personality? — Playful, Premium, Corporate, Energetic
3. Primary property? — position, scale, rotation, opacity
4. Duration? — see duration table below
5. Easing family? — entrance=decelerate, exit=accelerate
6. Hero element? — apply staging principles
7. Secondary + ambient layers? — add richness
8. 1/3 rules? — motion distance, simultaneous elements

Motion Personality Archetypes

Archetype	Duration	Easing	Overshoot	Keywords
Playful	150-300ms	ease-out-back	10-20%	fun, bouncy, cute, whimsical
Premium	350-600ms	cubic-bezier(0.4,0,0.2,1)	0%	elegant, minimal, luxury
Corporate	200-400ms	cubic-bezier(0.2,0,0,1)	0-3%	clean, professional, dashboard
Energetic	100-250ms	ease-out-expo	15-30%	dynamic, bold, exciting

Disney's 12 Principles (Adapted for Video/UI)

1. Squash & Stretch — Scale [1.2,0.8] on impact, [0.85,1.15] on stretch. Preserve volume.
2. Anticipation — Small opposite motion (10-20% of main) before action. 100-200ms.
3. Staging — Dim non-hero elements to 40-60% opacity. One primary action per beat.
4. Straight Ahead vs Pose-to-Pose — Fluid/organic vs planned/controlled. Use pose-to-pose for UI.
5. Follow-Through — Child elements trail parent by 50-150ms. Lower spring stiffness = more trail.
6. Slow In/Out — Entrance: ease-out. Exit: ease-in. NEVER linear for spatial movement.
7. Arcs — Add 10-20px perpendicular offset at path midpoint. Subtle=corporate, pronounced=playful.
8. Secondary Action — 30-50% amplitude of primary, 50-100ms after, different easing.
9. Timing — Heavy objects: 400-800ms. Light objects: 100-250ms. Enter 30-50% longer than exit.
10. Exaggeration — Playful: 15-25%, Corporate: 0-5%, Premium: 0%.
11. Solid Drawing — Consistent visual weight and perspective in every frame.
12. Appeal — Clean, readable, satisfying movement that serves the viewer.

Duration Table

Element Type	Duration	Rationale
Tooltip / micro-feedback	80-120ms	Must feel instant
Button press / toggle	120-180ms	Responsive feedback
Icon transition	150-250ms	Clear state change
Card enter / exit	200-350ms	Spatial awareness
Modal / dialog	300-400ms	Focus shift
Page / scene transition	400-600ms	Context switch
Dramatic reveal	600-1200ms	Theatrical build
Ambient loop	2000-20000ms	Continuous life

Easing Reference

Context	Easing	Cubic Bezier
Entrance	ease-out / decelerate	(0, 0, 0, 1)
Exit	ease-in / accelerate	(0.3, 0, 1, 1)
On-screen move	ease-in-out	(0.2, 0, 0, 1)
Bounce/playful	bounce-settle	(0.175, 0.885, 0.32, 1.275)
Elastic/dramatic	elastic-snap	(0.68, -0.55, 0.265, 1.55)
Apple iOS	HIG standard	(0.25, 0.1, 0.25, 1)
Material Design 3	MD3 standard	(0.2, 0, 0, 1)
Ambient/float	gentle	(0.4, 0, 0.2, 1)
Snappy UI	decisive	(0.2, 0, 0, 1)

Spring Parameters

Feel	Stiffness	Damping	Use
Very stiff	400+	25-30	Snapping, rigid
Standard	250-350	18-24	Default UI
Bouncy	150-250	10-15	Playful interactions
Very bouncy	100-200	5-10	Fun, game-like

Choreography Rules (Source: LottieFiles)

• Lead with hero — largest displacement, most attention-grabbing easing
• Spatial consistency — all elements enter from same direction or shared origin
• Counter-motion — hero moves right → ambient moves left at 20-30% speed
• 1/3 Rule (Distance) — no motion travels >1/3 of screen without intermediate keyframe
• 1/3 Rule (Elements) — max 1/3 of elements active simultaneously
• Stagger patterns: Sequential (lists), Center-out (hero), Wave (data bars), Random (organic)
• Total stagger budget: Stay under 500ms total

Emotion-to-Motion Map

Emotion	Motion Character	Path	Easing	Duration
Joy	Bouncy, arcs	Curved upward	ease-out-back	200-400ms
Calm	Smooth, flowing	Gentle curves	sine ease-in-out	500-1000ms
Urgency	Sharp, fast	Straight lines	ease-out	100-200ms
Sadness	Slow, downward	Drooping curves	cubic ease-in-out	600-1200ms
Surprise	Sudden, expanding	Radial outward	ease-out-expo	150-300ms
Elegance	Slow, controlled	Long arcs	(0.4,0,0.2,1)	400-700ms

Deep dive: references/motion-design.md

---

PART 9: SPEED & TIME MANIPULATION

Speed Ramping Guide

Effect	Speed	When to Use
Epic slow-mo	0.25-0.5x	Impact moments, beauty shots, reveals
Subtle slow	0.7-0.85x	Emphasis, emotional weight, focus
Normal	1.0x	Default
Slight fast	1.2-1.5x	Energy, tightening dead space
Timelapse feel	1.5-3x	Travel, process, time passage
Hyperlapse	4-16x	Long journeys, day-to-night
Speed ramp	Variable	Action sequences, reveals, transitions

FFmpeg Speed Recipes

# Smooth slow-mo 0.5x with optical flow interpolation
ffmpeg -i input.mp4 -filter:v "setpts=2.0*PTS,minterpolate='mi_mode=mci:mc_mode=aobmc:vsbmc=1:fps=60'" -an output.mp4

# Speed ramp: normal → slow at 2s → normal at 4s
ffmpeg -i input.mp4 -filter:v "setpts='if(between(T,2,4),2*PTS,PTS)'" output.mp4

# Hyperlapse 8x
ffmpeg -i input.mp4 -filter:v "setpts=0.125*PTS" -an output.mp4

# Reverse video
ffmpeg -i input.mp4 -vf reverse -af areverse output.mp4

Rule: Always ease into speed changes over 0.5-1s. Never jump speeds — it's jarring.

---

PART 10: COMPOSITING & OVERLAYS

Lower Thirds

• Animate in from left: 200-350ms, ease-out
• Hold: 3-5 seconds
• Animate out same direction: 150-250ms, ease-in
• Keep in lower 1/3, with safe-area padding (10% from edges)

FFmpeg Compositing

# Picture-in-Picture (top-right, 25% size)
ffmpeg -i main.mp4 -i pip.mp4 -filter_complex \
  "[1]scale=iw*0.25:ih*0.25[pip];[0][pip]overlay=W-w-20:20" output.mp4

# Side-by-side split screen
ffmpeg -i left.mp4 -i right.mp4 -filter_complex \
  "[0]crop=iw/2:ih:0:0[l];[1]crop=iw/2:ih:iw/2:0[r];[l][r]hstack" output.mp4

# Green screen (chroma key)
ffmpeg -i foreground.mp4 -i background.mp4 -filter_complex \
  "[0]chromakey=0x00FF00:0.1:0.2[fg];[1][fg]overlay" output.mp4

# Watermark/logo overlay
ffmpeg -i input.mp4 -i logo.png -filter_complex \
  "[1]scale=100:-1,format=rgba,colorchannelmixer=aa=0.7[logo];[0][logo]overlay=W-w-20:20" output.mp4

# Vignette effect
ffmpeg -i input.mp4 -vf "vignette=angle=PI/4:mode=forward" output.mp4

# Letterbox (cinematic bars)
ffmpeg -i input.mp4 -vf "pad=iw:iw*9/16:(ow-iw)/2:(oh-ih)/2:black" output.mp4

Overlay Rules (CRITICAL — Source: video-use)

1. Subtitles applied LAST in filter chain — overlays can hide captions otherwise
2. Per-segment extract → lossless concat — don't double-encode every segment
3. Overlay timing: setpts=PTS-STARTPTS+T/TB to sync overlay frame 0 to its window start
4. Master SRT offsets: output_time = word.start - segment_start + segment_offset

---

PART 11: EXPORT & DELIVERY

Platform Presets

Platform	Resolution	FPS	Bitrate	Aspect	Codec
YouTube (1080p)	1920x1080	24-60	8-12 Mbps	16:9	H.264/H.265
YouTube (4K)	3840x2160	24-60	35-45 Mbps	16:9	H.265
Instagram Reels	1080x1920	30	5-8 Mbps	9:16	H.264
TikTok	1080x1920	30	5-8 Mbps	9:16	H.264
Twitter/X	1280x720	30	5 Mbps	16:9	H.264
LinkedIn	1920x1080	30	8 Mbps	16:9	H.264
Cinematic	2560x1080	24	15-20 Mbps	21:9	H.265
ProRes Master	1920x1080+	24+	~150 Mbps	Any	ProRes 422

FFmpeg Export Commands

# YouTube optimized (H.264, web-ready)
ffmpeg -i input.mp4 -c:v libx264 -preset slow -crf 18 -c:a aac -b:a 192k -movflags +faststart output.mp4

# Social media vertical (auto-pad to 9:16)
ffmpeg -i input.mp4 -vf "scale=1080:1920:force_original_aspect_ratio=decrease,pad=1080:1920:(ow-iw)/2:(oh-ih)/2:black" -c:v libx264 -crf 20 output.mp4

# High quality master (archival)
ffmpeg -i input.mp4 -c:v libx264 -preset veryslow -crf 15 -c:a aac -b:a 320k -movflags +faststart output.mp4

# GIF from video (optimized)
ffmpeg -i input.mp4 -vf "fps=15,scale=480:-1:flags=lanczos,split[s0][s1];[s0]palettegen[p];[s1][p]paletteuse" output.gif

# Extract audio only
ffmpeg -i input.mp4 -vn -c:a libmp3lame -q:a 2 output.mp3

# Thumbnail at specific time
ffmpeg -i input.mp4 -ss 00:00:05 -vframes 1 -q:v 2 thumbnail.jpg

---

PART 12: THE PRO WORKFLOW

Complete Production Pipeline

1. Inventory — catalog all source material (duration, content, quality, resolution)
2. Transcribe — word-level timestamps via ElevenLabs Scribe / Whisper / FunASR
3. Analyze — understand each clip: visual content, audio, energy, usability
4. Strategy — propose narrative structure, get user confirmation before cutting
5. Assembly — rough cut following narrative plan (3-act structure)
6. Fine cut — trim, pace, add J/L-cuts, match cuts, remove dead space
7. Color — match shots for consistency, then apply style grade
8. Sound — music selection, SFX, audio levels, ducking, normalization
9. Graphics — titles, lower thirds, overlays, animations, kinetic text
10. Subtitles — generate and burn captions (applied LAST in filter chain)
11. Review — watch at 1x speed. Fix everything that bothers you.
12. Export — platform-appropriate format and resolution
13. Iterate — show user, incorporate feedback, re-export

Quality Checklist (Before Delivery)

• No audio pops at cut points (30ms fades applied)
• No flash frames or single-frame artifacts
• Color consistent within scenes
• Audio levels normalized (-16 LUFS target)
• Music ducks under dialogue (6-12dB reduction)
• Subtitles readable, correctly timed, no overlap
• No watermarks or artifacts from source material
• Aspect ratio correct for target platform
• Smooth speed ramps (no jarring speed jumps)
• All transitions serve the story (no gratuitous effects)
• Opening hook in first 3 seconds
• Energy curve matches narrative arc
• Export codec and bitrate match platform requirements

---

TECHNOLOGY ROUTING

When the task requires a specific technology:

Need	Tool	Skill Reference
Raw video processing	FFmpeg	This skill + ffmpeg-recipes/
Pro effects (glow, velocity, flash)	FFmpeg	ffmpeg-recipes/pro-effects.md
Text overlays & typography	FFmpeg	ffmpeg-recipes/text-overlays.md
Sound design & SFX	FFmpeg	sound-design/sfx-guide.md
Audio mixing & loudness	FFmpeg	sound-design/audio-mixing.md
Beat-synced editing	Python + FFmpeg	automation/beat-sync.md
Auto captions (Whisper)	Python + FFmpeg	automation/auto-captions.md
Batch video processing	Python + FFmpeg	automation/batch-pipeline.md
Social media viral edits	Strategy + FFmpeg	social-media/viral-editing.md
Platform export specs	FFmpeg	social-media/platform-specs.md
React-based video	Remotion	programmatic/remotion-guide.md
Python editing	MoviePy	programmatic/moviepy-guide.md
AI-powered workflow	Multi-tool	ai-tools/ai-workflow.md
HTML-to-video	HyperFrames	hyperframes
Math animations	Manim	manim.community
AI image-to-video	AnimateAnything	arxiv
AI frame propagation	I2VEdit	i2vedit
TTS / voice	ElevenLabs	elevenlabs.io
Transcription	Whisper / FunASR	OpenAI Whisper or Alibaba FunASR
Motion graphics	GSAP / Lottie / CSS	See Part 8
Cloud GPU rendering	Modal / RunPod	~$0.01-0.23 per run

---

This skill synthesizes knowledge from video-use, hyperframes, motion-design-skill, motion-skills, claude-code-video-toolkit, claude-code, Claude-Code-Video-Toolkit, ai-video-editing-skill, Monet, AnimateAnything, and I2VEdit. Full credit to all contributors.