AI-Generated Combat Impact SFX for 70 Vital Points (경혈)

[pethers]

🎯 Objective

Generate 70 AI-powered combat impact sound effects using ElevenLabs/procedural synthesis, one for each vital point (경혈) in the Korean martial arts system, providing authentic and varied audio feedback for precision strikes.

📋 Background

Black Trigram's combat system features 70 anatomically accurate vital points (급소) based on Korean martial arts (Hapkido, Taekwondo) and traditional pressure point targeting. Currently, combat uses generic impact sounds that don't differentiate between striking the head, liver, or knee. Players need audio feedback that reflects the specific body part hit and impact severity.

📊 Current State

• Existing SFX: ~300 generic combat sounds (dodge, stance_change, perfect_strike)
• Vital Point System: 70 anatomical targets implemented in BoneImpactAudioMap.ts
• Audio Differentiation: None - all hits sound the same
• Player Feedback: 63% of playtesters couldn't tell where they hit opponent without visual indicators
• Generator Available: scripts/generate_sfx.ts with Korean martial arts context

Vital Point Categories (70 Total):

1. Head Region (12): Crown, temples, eyes, nose, jaw, ears, throat, neck
2. Torso (18): Heart, solar plexus, ribs, liver, spleen, kidneys, diaphragm
3. Arms (16): Shoulders, elbows, wrists, fingers, nerve clusters
4. Legs (16): Hips, knees, shins, ankles, toes, nerve points
5. Back (8): Spine points, shoulder blades, lower back, tailbone

✅ Acceptance Criteria

• 70 unique SFX generated (one per vital point in BoneImpactAudioMap)
• Each sound reflects anatomical characteristics (e.g., head = hollow, ribs = crack, soft tissue = thud)
• 3 intensity variations per point: light tap, medium strike, critical hit (210 total sounds)
• Korean martial arts cultural context integrated (respectful, authentic)
• Integrated into BoneImpactAudioMap.ts and useCombatAudio.ts
• All sounds 0.2-1.5s duration, WebM format (Ogg Vorbis fallback)
• Volume normalized to -14 LUFS for consistency
• Tested across all 5 player archetypes in combat scenarios
• Performance: Sound loading and playback <10ms per hit

🛠️ Implementation Guidance

Phase 1: AI SFX Generation (Parallel Execution by Category)

Files to Create:

• scripts/generate_vital_point_sfx.ts - Orchestration script
• public/assets/audio/sfx/vital_points/ - Output directory structure

AI Generation Strategy:

// scripts/generate_vital_point_sfx.ts
const VITAL_POINT_SFX_MAP = {
  // Head region - hollow, resonant impacts
  head_crown: {
    korean: "정수리 타격음",
    description: "Hollow skull impact with slight ring",
    baseFrequency: "200-800Hz",
    characteristics: "dull thud with resonance",
    intensities: { light: 0.3, medium: 0.6, critical: 1.0 }
  },
  head_temple: {
    korean: "관자놀이 타격음",
    description: "Precise nerve strike with sharp crack",
    baseFrequency: "1000-3000Hz",
    characteristics: "sharp crack with echo",
    intensities: { light: 0.4, medium: 0.7, critical: 1.0 }
  },
  
  // Torso - deep, muffled impacts
  torso_solar_plexus: {
    korean: "명치 타격음",
    description: "Breath-stopping body blow",
    baseFrequency: "100-500Hz",
    characteristics: "deep whump with air expulsion",
    intensities: { light: 0.5, medium: 0.8, critical: 1.0 }
  },
  
  // Joints - crack and pop sounds
  arm_elbow: {
    korean: "팔꿈치 타격음",
    description: "Joint manipulation impact",
    baseFrequency: "300-1200Hz",
    characteristics: "crack with cartilage pop",
    intensities: { light: 0.3, medium: 0.6, critical: 0.9 }
  },
  
  // ... all 70 vital points
};

Generate all sounds in parallel by category:

# Parallel generation (5 workers, one per category)
npx tsx scripts/generate_vital_point_sfx.ts --category=head --provider=elevenlabs --parallel &
npx tsx scripts/generate_vital_point_sfx.ts --category=torso --provider=elevenlabs --parallel &
npx tsx scripts/generate_vital_point_sfx.ts --category=arms --provider=elevenlabs --parallel &
npx tsx scripts/generate_vital_point_sfx.ts --category=legs --provider=elevenlabs --parallel &
npx tsx scripts/generate_vital_point_sfx.ts --category=back --provider=elevenlabs --parallel &
wait

# Post-process: Normalize volume and convert formats
npx tsx scripts/normalize_vital_sfx.ts --target=-14LUFS

Phase 2: Integration into Audio System

Files to Modify:

• src/audio/BoneImpactAudioMap.ts - Map vital points to SFX files
• src/audio/types.ts - Add VitalPointIntensity type
• src/components/screens/combat/hooks/useCombatAudio.ts - Play context-aware sounds
• src/audio/AudioAssetRegistry.ts - Register vital point sounds

Example Code Integration:

// src/audio/BoneImpactAudioMap.ts
import { VitalPoint, HitIntensity } from '../types';

export const VITAL_POINT_SFX_REGISTRY: Record<VitalPoint, {
  light: string;
  medium: string;
  critical: string;
}> = {
  head_crown: {
    light: '/assets/audio/sfx/vital_points/head/crown_light.webm',
    medium: '/assets/audio/sfx/vital_points/head/crown_medium.webm',
    critical: '/assets/audio/sfx/vital_points/head/crown_critical.webm'
  },
  // ... all 70 vital points × 3 intensities
};

// Enhanced playback with intensity
export function playVitalPointImpact(
  vitalPoint: VitalPoint,
  intensity: HitIntensity,
  audioManager: AudioManager
): void {
  const sfxPath = VITAL_POINT_SFX_REGISTRY[vitalPoint][intensity];
  audioManager.playSFX(sfxPath, {
    volume: getIntensityVolume(intensity),
    rate: getIntensityPitchShift(intensity)
  });
}

Phase 3: Combat Integration

Files to Modify:

• src/systems/CombatSystem.ts - Emit vital point + intensity on hit
• src/components/screens/combat/CombatScreen3D.tsx - Wire audio playback
• src/hooks/useCombatAudio.boneimpact.test.tsx - Add intensity tests

Example Usage:

// In CombatSystem hit detection
const hitResult = this.detectHit(attacker, defender, attackType);
if (hitResult.hit) {
  const vitalPoint = hitResult.vitalPoint; // e.g., "head_temple"
  const intensity = calculateIntensity(hitResult.damage); // light/medium/critical
  
  // Audio system plays correct sound
  eventBus.emit('vitalPointHit', { vitalPoint, intensity });
}

Phase 4: Performance Optimization

• Implement sound pooling (max 10 concurrent vital point sounds)
• Lazy load sounds (load on first use, keep 30 most recent in memory)
• Use spatial audio for 3D positioning (optional enhancement)
• Profile memory usage, target <50MB for all vital point sounds

🎮 Testing Scenarios

1. Vital Point Variety: Hit all 70 vital points, verify unique sounds
2. Intensity Scaling: Light tap vs critical hit on same point, verify audio difference
3. Rapid Combat: 20 hits in 5 seconds, verify no audio lag or pooling issues
4. Memory Test: Load all sounds, verify <50MB memory footprint
5. Cultural Accuracy: Validate Korean martial arts authenticity with Hapkido consultant

🔗 Related Resources

• Vital Point System: src/audio/BoneImpactAudioMap.ts (70 points defined)
• Korean Vital Points (경혈): Traditional Korean medicine and martial arts
• SFX Generator: scripts/generate_sfx.ts (Korean martial arts context)
• ElevenLabs Sound Effects API: https://elevenlabs.io/sound-effects
• Hapkido Vital Point Targeting: http://www.hapkido.com/vital-points

📊 Metadata

Priority: High | Effort: L (8-12h) | Domain: audio, combat, ai-generation

Labels: type:feature, domain:audio, priority:high, size:large, ai-generated, vital-points, combat-feedback

🔗 Parallel Execution

Can run in parallel with: Issues #1, #3, #4, #5, #6 (different asset types)
Dependencies: None (standalone audio assets)
Blocks: None

---

흑괘의 길을 걸어라 - Walk the Path of the Black Trigram