Home / Audio / Production and Craft

C3 Spec: Frequency / Sound Layer (tones under the Sacred Audio)

Updated Jun 21, 2026 · Affirmology_FrequencySoundLayer_C3_Spec_v1.md

C3 Spec: Frequency / Sound Layer (tones under the Sacred Audio)

Status: to build. Studio only. The locked demo never changes. Request: Jeff wants binaural / solfeggio / isochronic / tonal overlays under the audios, chosen by what the audio is for. Test & Change Queue item C3. Informs and is informed by: C26 (per-Structure bed_profile / pace) and the frequency_sound craft area (6 live entries) in the corpus.

What this is

A new optional layer in the mix that places a low, subtle tonal bed under the voice and music: a brainwave-entrainment tone matched to the state the audio is trying to produce. Delta under sleep, theta under journey and before-bed install, alpha under calm-but-aware morning work, beat-driven energy under activation. The tone sits well below the voice, supportive and barely conscious, never the focus.

The honesty rule (non-negotiable)

Entrainment tones are framed as supportive and traditional, not as medical or guaranteed-neuroscience claims. We never promise a clinical effect, never make a health claim, and in any user-facing or build-report copy we state plainly where the science is suggestive versus established. This is the same integrity rule we already apply to the "wave function" manifestation framing. The craft base entries that inform this layer carry that distinction; preserve it.

State to frequency map (the default profile table)

Driven by Structure.pace and Structure.bed_profile (added in C26), with a clean per-type default: - SLEEP / before-bed install: delta (roughly 1 to 4 Hz) for slow-wave consolidation, lowest and softest. - JOURNEY / deep meditative: theta (roughly 4 to 8 Hz), the encoding window for new self-narrative. - MORNING / calm-aware: alpha (roughly 8 to 12 Hz), relaxed but present. - ACTIVATION / gym: a low beta or rhythmic pulse for drive, paired with the driving bed. - Optional accent: a single solfeggio-style tonal center (e.g. a warm low tone) under the bed for color, kept subtle. Each value is a Studio-tunable default, not a hard constant, so we can iterate by ear.

Modalities to support

• BINAURAL beats (two slightly detuned tones, one per ear, the brain perceives the difference frequency). Requires true stereo and is headphone-dependent; flag in the build report when a render uses binaural so we know it needs headphones.
• ISOCHRONIC tones (a single tone pulsed on and off at the target frequency). Works on speakers, not headphone-dependent. Likely the safer default for broad delivery.
• SOLFEGGIO / tonal center (a sustained warm tone for character, not entrainment per se). Start with isochronic + a tonal center (robust on any device), add binaural as a headphone-flagged option.

Where it goes in the pipeline

All of it lives in affirmology-agent/.../audio_mix.py, as a new mix stage AFTER the voice and music bed are combined and BEFORE the final master: 1. Generate the tone bed programmatically (sine synthesis at the mapped frequency, isochronic gating or binaural detune as selected) for the render's duration. No external sample library needed for the core tones. 2. Mix it in at a low, fixed-headroom level well under the voice (a configurable gain floor so it never competes with words or trips QC). 3. Run the existing QC gate on the final mix unchanged. The tone layer must never introduce artifacts the QC would catch, and must never raise the voice-intelligibility floor.

Configuration

• Structure.bed_profile selects the frequency profile (reuse the C26 field).
• A Studio-only flag (e.g. --tones on/off plus an intensity) so a render can include or omit the layer and we can A/B by ear.
• Tone gain, modality, and the per-state frequency table are config, tunable without code changes where practical.

Guardrails (hard)

• The demo is locked. The demo render never includes the tone layer unless Jeff explicitly says "make this the new demo." Verify the demo path does not invoke the tone stage.
• QC gate runs unchanged on the final mix; a tone-bearing render that fails QC is rejected like any other.
• No health or medical claims anywhere. Honesty note travels into the build report.
• Additive and off by default for existing structures, so nothing changes silently.
• No em dashes.

Acceptance tests

1. A before-bed structure renders with a delta isochronic bed at the configured low gain, voice fully intelligible, QC pass.
2. A journey structure renders with a theta bed; an activation structure renders with the rhythmic/driving profile; both QC pass.
3. Tones off reproduces today's mix (voice + music only), proving the layer is cleanly optional.
4. The locked demo render is unchanged and never carries a tone bed.
5. The build report records: modality used, frequency, gain, headphone-required flag (if binaural), and the honesty note.

Open choices for Jeff (decide by ear later, not blockers)

• Isochronic-first (works on speakers) vs binaural-first (needs headphones) as the default. Recommendation: isochronic + tonal center as default, binaural as a flagged option.
• Whether to add a small curated library of recorded tonal pads later, or stay fully synthesized. Recommendation: start synthesized (zero licensing, fully controllable).

The exact paste for Claude Code (when ready to build)

Read Affirmology_PROJECT_STATE.md, Affirmology_FrequencySoundLayer_C3_Spec_v1.md, and the frequency_sound craft entries (search_craft area=frequency_sound).

Implement C3 the tone/frequency layer per the spec: add a synthesized tone-bed stage in audio_mix.py after voice+music and before the master, isochronic + tonal center first with binaural as a headphone-flagged option, driven by Structure.bed_profile with the state-to-frequency default table, low fixed-headroom gain, Studio-only --tones flag, off by default. Keep the QC gate unchanged and run it on the final mix. The locked demo must never carry a tone bed. Carry the honesty note into the build report. No em dashes.

Add the five acceptance tests from the spec. Show me diffs, test output, and three real renders to listen to (before-bed delta, journey theta, tones-off control) before deploying. Do not touch the demo.