Home / Ops / The Locked Demo

Affirmology Demo Audio - Baseline Specification

Updated Jun 13, 2026 · Affirmology_DemoAudio_Baseline_v1.md

Affirmology Demo Audio - Baseline Specification

This is the locked baseline for the Sacred Audio demo. The Jeff Parker original demo is the reference point. Every personalized demo we render (Laura, Steven, Sol, Colin, Alex Carper, etc.) is matched to this spec. Do not deviate without explicit approval from Jeff.

Reference recording

The Jeff Parker original demo audio is the gold standard. Match its: - Voice prosody and warmth - Music bed presence (felt but not competing) - Length around 6 to 7 minutes - Voice-anchored mix (music ends WHEN voice ends, no 30+ second tail of music alone) - Loudness perceived as cinematic but not fatiguing

Voice settings

• Provider: ElevenLabs
• Voice: Charlotte
• Model: eleven_turbo_v2_5
• Speed: auto-tune (the auto-tune speed selector we built picks the right pace per script length and tone)
• Polyvagal warm zone: 200 to 400 Hz vagal range preferred for the script's emotional core

Music settings

• Bed track: music/Heavenly Circuit.mp3 (lives in affirmology-agent repo)
• Voice offset: 8.0 seconds (voice enters 8 seconds after music begins)
• Mix: voice-anchored. Music duration is sized to match voice + outro tail, not the other way around.

Mix and mastering

• FFmpeg sidechain compression so music ducks under voice naturally
• Final loudness target: -16 LUFS (integrated)
• True peak ceiling: -1.0 dBTP
• Dynamic range: leave breathable, not slammed

File naming

{FirstName}_{LastName}_Affirmology_Demo_v{N}.mp3

Examples: - Jeff_Parker_Affirmology_Demo_v1.mp3 - Laura_Nielsen_Affirmology_Demo_v2.mp3

Increment v{N} for every re-render, even small ones. This kills the iTunes/Finder stale-cache problem we solved earlier.

CLI command that produces a baseline-spec demo

PYTHONPATH=src python -m affirmology.cli \
  --name "First Last" \
  --date YYYY-MM-DD \
  --time HH:MM \
  --location "City, State/Country" \
  --audio \
  --music "music/Heavenly Circuit.mp3" \
  --voice-offset 8.0 \
  --out out/demo_{job_id}

This is the exact command embedded in /Users/jeffreyparker/CLAUDE/AFFIRMOLOGY/affirmology-demo-site/server/api.py lines 109 through 119. That file is the canonical reference for the production demo backend.

Sacred Audio Report PDF (visual companion to the audio)

Every demo audio is accompanied by a Sacred Audio Report PDF. Standards: - Cover page with name, archetype line, birth data - Astrology summary (Sun, Moon, Ascendant, dominant aspects) - Gene Keys profile (first 4 gates with shadow/gift/siddhi notes) - Human Design summary (type, authority, profile, defined centers) - Numerology core numbers - Why these elements were chosen for the audio script - Closing reflection

The Josh / Jeff Parker version is the visual gold standard. Match its layout, typography, and tone.

Things we have explicitly solved (don't relitigate)

1. The 39-second music tail problem (voice-anchored mix solved it)
2. iTunes/Finder caching a stale mp3 (versioned filenames solved it)
3. Voice ending before music ended in earlier renders (voice-offset 8.0 + voice-anchored mix solved it)
4. Auto-tune speed selection (built and shipping)
5. Sacred Audio Report PDF agent (built and shipping)

Things still open

• Audio mastering chain integration (the deeper master we researched, see Affirmology_AudioMastering_Production_Research_v1.md, has not been integrated into audio_mix.py yet)
• Bricasti M7 impulse response reverb for voice (researched, not integrated)
• Cousto Cosmic Octave personal pitch derivation (researched, not integrated)

These are upgrades on top of the baseline. The current baseline is the floor, not the ceiling.

Writing rules to honor

• Never use em dashes anywhere
• "Subconscious Operating System" is the brand framing
• No Joe Dispenza trademark references (use "heart coherence breath" or "embodied gratitude induction")
• Sol = Soledad Gabriela Ballard (not Soul)