---

name: Audio Mixing & Mastering tier: 3 load_policy: task-specific description: Mix voice, binaural, and SFX stems into final master version: 1.0.0 parent_skill: production-operations

Audio Mixing Skill

The Mix Is Where Magic Happens

This skill handles combining voice, binaural beats, and sound effects into a cohesive, hypnotic audio experience.

---

Purpose

Mix multiple audio stems at correct levels and apply hypnotic post-processing for production masters.

---

Standard Stem Levels

CRITICAL: Always use these exact levels.

---

Canonical Mix Command

Copyffmpeg -y \
  -i sessions/{session}/output/voice_enhanced.wav \
  -i sessions/{session}/output/binaural_dynamic.wav \
  -i sessions/{session}/output/sfx_track.wav \
  -filter_complex \
    "[0:a]volume=-6dB[voice]; \
     [1:a]volume=-6dB[bin]; \
     [2:a]volume=0dB[sfx]; \
     [voice][bin][sfx]amix=inputs=3:duration=longest:normalize=0[mixed]" \
  -map "[mixed]" \
  -acodec pcm_s16le \
  sessions/{session}/output/session_mixed.wav

Important: Use normalize=0 to prevent unpredictable level changes.

---

Input Stems

Never use voice.mp3 - always use the enhanced version.

---

Hypnotic Post-Processing (MANDATORY)

After mixing, apply psychoacoustic mastering:

Copypython3 scripts/core/hypnotic_post_process.py --session sessions/{session}/

Triple-Layer Hypnotic Presence

Additional Enhancements

---

Output Files

---

Two-Stem Mix (No SFX)

When no SFX track is needed:

Copyffmpeg -y \
  -i sessions/{session}/output/voice_enhanced.wav \
  -i sessions/{session}/output/binaural_dynamic.wav \
  -filter_complex \
    "[0:a]volume=-6dB[voice]; \
     [1:a]volume=-6dB[bin]; \
     [voice][bin]amix=inputs=2:duration=longest:normalize=0[mixed]" \
  -map "[mixed]" \
  -acodec pcm_s16le \
  sessions/{session}/output/session_mixed.wav

---

Level Verification

After mixing, check levels:

Copy# Check peak level (should be < 0 dB)
ffmpeg -i session_mixed.wav -af "volumedetect" -f null /dev/null 2>&1 | grep max_volume

# Check LUFS (target: -14 LUFS)
ffmpeg -i session_mixed.wav -af "loudnorm=print_format=json" -f null /dev/null 2>&1

Target Levels

---

Binaural Beat Standards

Typical Journey Curve

Copy0:00  - Alpha (10 Hz) - Induction
5:00  - Theta (7 Hz)  - Deepening
15:00 - Deep Theta (4 Hz) - Journey core
25:00 - Theta (7 Hz)  - Integration
28:00 - Alpha (10 Hz) - Emergence

---

Troubleshooting

---

Quality Checklist

Before video assembly:

• {session}_MASTER.mp3 exists
• No clipping or distortion
• Binaural beats audible but not intrusive
• Voice clear and present
• SFX properly timed (if used)
• LUFS within target range
• Duration matches expected

---

Integration with Pipeline

Before (dependencies):

• Voice synthesis complete (voice_enhanced.wav)
• Binaural generated (binaural_dynamic.wav)
• SFX track if needed (sfx_track.wav)

After (next steps):

• Video assembly
• YouTube packaging

---

Related Resources

• Skill: tier3-production/voice-synthesis/ (input)
• Skill: tier3-production/video-assembly/ (next step)
• Serena Memory: audio_production_methodology
• Script: scripts/core/hypnotic_post_process.py