CymaVoice-Science

The Science of CymaVoice™ — CymaTones
Technical Science Brief

Your Voice is the Most Accurate
Biometric on Earth

A deep-dive into the cymatics science, FFT technology, and clinical frequency mapping that powers CymaVoice™ — the most advanced vocal frequency scanner available without a clinical lab.

32,768FFT Resolution
1.35HzPer Bin Precision
4Harmonic Layers
30sFull Capture
40Chladni Patterns
01 · The Foundation

The Science Begins with
Cymatics & Resonance

Every organ, tissue and cell in the human body vibrates at a specific, measurable acoustic frequency. When a structure falls out of its natural resonance, biological function is compromised. Restore the frequency — restore the health.

"Every cell, tissue and organ in the body has its own resonant frequency. When a structure falls out of its natural resonance, disease follows."

Dr. Peter Guy Manners
Pioneer of Cymatics · Inventor of the Cyma 1000 Clinical Device · Creator of Commutation Frequencies

Dr. Manners developed the Cyma 1000 — the world's first clinical acoustic therapy device — after decades of research. He catalogued specific frequency combinations (commutations) for every major organ system, verified through clinical outcomes with thousands of patients.

CymaTones was built from Dr. Manners' original Cyma 1000 commutation database. Every track in our library corresponds to a specific organ system, mapped to his clinical frequency specifications. CymaVoice applies this framework in reverse — instead of delivering frequencies to the body, we analyse which frequencies are missing from your voice and map those deficiencies back to the organs they represent.

What is a Commutation? A commutation is Dr. Manners' term for a precise combination of 5 acoustic frequencies that together produce the resonant signature of a specific organ or tissue. Not a single note — a complex chord of frequencies that matches the biological resonance of that structure.

02 · The Mapping System

12 Notes · 12 Organs · One
Complete Meridian Map

In Traditional Chinese Medicine, every organ system vibrates at a specific musical note. Each chromatic note frequency resonates with and activates the meridian of its corresponding organ.

💧
Water
C · C#
Bladder
Kidney
🌿
Wood
D · D#
Gallbladder
Liver
🔥
Fire
F# · G · G#
Heart · SI
Triple Warmer
🌍
Earth
E · F
Stomach
Spleen
⚙️
Metal
A · A#
Lung
Large Intestine
Note Hz Organ Meridian Element Chladni (m,n) Stored Emotion
C 130.8 Hz Bladder BL Water [1,2] Fear
C# 138.6 Hz Kidney KD Water [1,3] Fear / Willpower
D 146.8 Hz Gallbladder GB Wood [2,3] Anger
D# 155.6 Hz Liver LV Wood [1,4] Frustration
E 164.8 Hz Stomach ST Earth [2,4] Worry
F 174.6 Hz Spleen SP Earth [3,4] Overthinking
F# 185.0 Hz Heart HT Fire [1,5] Joy / Anxiety
G 196.0 Hz Small Intestine SI Fire [2,5] Vulnerability
G# 207.7 Hz Triple Warmer TW Fire [3,5] Social Safety
A 220.0 Hz Lung LU Metal [1,6] Grief
A# 233.1 Hz Large Intestine LI Metal [2,6] Letting Go
B 246.9 Hz Pericardium PC Fire [4,5] Heart Protection
03 · The Scanner Technology

The Most Advanced Vocal
FFT Engine in a Browser

CymaVoice uses a professional-grade Fast Fourier Transform engine — the same mathematical process used in acoustic research laboratories — to decode the frequency composition of your voice in real time.

FFT Resolution
32,768
Maximum FFT Window Size
The highest FFT window available in the Web Audio API. At 44,100Hz sample rate, this gives ~1.35Hz per frequency bin — precisely isolating each chromatic note with ~10 bins of separation. No frequency bleed between organs.
Precision Mode
Float32
Full Floating-Point dB Values
getFloatFrequencyData() for full 32-bit floating point decibel precision. Most implementations use 8-bit byte data (0–255). Our approach captures every fraction of a decibel — critical for distinguishing subtle frequency deficiencies.
Dynamic Range
-100dB
Extended Voice Capture Range
minDecibels set to -100dB and maxDecibels to -10dB, tuned for the human voice dynamic range. Captures both whisper-quiet undertones and full chest resonance without clipping or missing subdued frequencies.
Smoothing
Zero
No Built-In Frame Averaging
smoothingTimeConstant = 0. We disable all built-in frame averaging and perform our own multi-dimensional accumulation across the full 30 seconds. Clinical-quality statistical analysis, not a smoothed approximation.
Why does resolution matter? At the default fftSize of 2048, each bin spans ~21.5Hz. The gap between C (130.8Hz) and C# (138.6Hz) is only 7.8Hz — adjacent organ notes would bleed into each other at octave 3. At 32,768, each bin spans 1.35Hz — we get 5+ clean bins between adjacent notes. This is the difference between seeing organ imbalances clearly and seeing noise.

4-Dimension Composite Score

No single measurement tells the complete story. CymaVoice tracks four independent dimensions across the full 30-second scan:

Composite Score Formula
Score(note) = Presence × 0.40 + Average × 0.30 + Peak × 0.20 + Harmonic × 0.10
40%
Presence Duration
What fraction of the 30-second recording was this note active above the detection threshold? The most clinically diagnostic dimension — persistence reveals constitutional patterns.
30%
Sustained Average
Mean amplitude across all frames. Shows constitutional energy level — baseline vitality, not just the best moment. Reveals chronic depletion vs acute imbalance.
20%
Peak Amplitude
Highest single-frame measurement across the scan. Captures the full range of the organ's potential — important for detecting dormant or suppressed frequencies.
10%
Harmonic Richness
Accumulated harmonic product spectrum score across all frames. Measures genuine tonal quality — a rich harmonic stack confirms genuine organ resonance, not surface noise.
04 · Harmonic Intelligence

Harmonic Product Spectrum —
Confirming True Frequency

A simple FFT cannot tell you whether a detected frequency is a true fundamental tone or merely an overtone harmonic of a different fundamental. The Harmonic Product Spectrum solves this — used in professional pitch detection and clinical acoustic analysis.

HPS Composite Score Formula
S(note) = F₀ × 0.70 + (H₂ × 0.40 + H₃ × 0.25 + H₄ × 0.10) × 0.30
F₀Normalised amplitude at the fundamental frequency bin (±¼ semitone window)
H₂Amplitude at the 2nd harmonic (2× fundamental Hz) — strongest overtone
H₃Amplitude at the 3rd harmonic (3× fundamental Hz) — confirms rich tone
H₄Amplitude at the 4th harmonic (4× fundamental Hz) — validates resonance
Why this matters clinically: Without HPS, a strong overtone of the Liver frequency (D#) could falsely appear as a Heart frequency (F#). HPS cross-references all four harmonic layers to confirm the true fundamental, eliminating false positives that would lead to incorrect organ assessments.
05 · The Scan Protocol

30 Seconds Designed to
Capture Everything

The guided scan is not random. Each phase is acoustically designed to activate specific frequency bands — ensuring all 12 chromatic notes have the opportunity to be captured. Up to 3 scans per day.

0 – 2s
Breath — Baseline Calibration
One deep breath before recording sets the baseline noise floor and allows the analyser to calibrate. Silence captures ambient room noise subtracted from subsequent analysis.
2 – 10s
🎵 Sustained Hum — Chest & Skull Resonance
Closed-lip humming activates the lowest fundamental frequencies through bone and tissue — exciting C, C#, D, D# (Water and Wood elements: Kidney, Bladder, Liver, Gallbladder). Sustained 8 seconds for strong presence scores.
Activates: C · C# · D · D#
10 – 18s
🌊 Vowels — Formant Activation
Each vowel shape creates a distinct resonant cavity, exciting different formant frequency bands. AAAA activates mid frequencies, EEEE opens upper-mid bands, OOOO drops into low-mid resonance.
Activates: E · F · F# · G · G#
18 – 25s
🎶 Pitch Glide — Full Chromatic Sweep
A slow glide from deepest chest note to highest comfortable pitch and back sweeps through all 12 chromatic notes in one continuous motion. The FFT captures each note as the voice passes through it.
Activates: All 12 notes across full range
25 – 30s
🗣️ Speech — Harmonic Speech Pattern
Speaking your name and counting backwards activates real speech frequencies — consonants, diphthongs and natural prosody — that humming cannot capture. Activates A and A# (Lung, Large Intestine — Metal element) through consonant bursts.
Activates: A · A# · B
06 · From Voice to Report

The Complete Analysis Pipeline

From microphone input to your personalised wellness report — every step of what happens to your vocal data.

01
Microphone Capture via Web Audio API
Your device microphone is accessed via getUserMedia() with audio-only constraints. The stream is connected to an AudioContext at the device's native sample rate (typically 44,100Hz or 48,000Hz). No audio is stored, transmitted or recorded — all processing happens locally in your browser.
navigator.mediaDevices.getUserMedia({audio:true}) → MediaStreamSource
02
AnalyserNode Configuration
The audio stream is routed through a Web Audio API AnalyserNode configured for maximum precision. fftSize 32,768 provides ~1.35Hz/bin resolution. Float32 data gives full dB precision. Zero smoothing — every frame is independent.
fftSize:32768 · smoothing:0 · minDb:-100 · maxDb:-10 · Float32Array
03
Real-Time Harmonic Product Spectrum
Every animation frame (~60fps), the full frequency spectrum is read. For each of 12 chromatic notes, we measure the fundamental frequency bin and its 2nd, 3rd and 4th harmonic bins. The HPS composite score confirms true fundamentals vs harmonic artefacts.
∀ note: HPS = F₀×0.7 + (H₂×0.4 + H₃×0.25 + H₄×0.1)×0.3
04
30-Second Multi-Dimensional Accumulation
Across the full 30 seconds (~1,800 frames), four independent statistics are accumulated per note: running sum (average), peak value, presence counter (frames above threshold), and harmonic richness score.
fftAccum · fftPeak · fftPresence · fftHarmonic → 48 data points total
05
Weighted Composite Spectrum
The four dimensions are combined into a single composite score per note using clinical weighting: Presence 40%, Average 30%, Peak 20%, Harmonic Richness 10%. This produces a 12-value rich spectrum reflecting genuine constitutional organ frequency patterns.
presence×0.40 + avg×0.30 + peak×0.20 + harmonic×0.10 → 12-value spectrum
06
Instant Local Analysis
The composite spectrum is immediately processed through our built-in clinical analysis engine. TCM element scoring, organ vitality assessment, chakra mapping, emotional pattern identification and a 13-track CymaTones protocol are all generated locally — in under 4 seconds, with zero dependency on external services.
buildLocalResult() → TCM patterns · Element scores · Protocol → onResult() in ~4s
07
AI Enhancement via Claude API
In parallel, the rich spectrum with full clinical context is sent to Claude claude-sonnet-4-20250514 for deep clinical interpretation. The AI receives your exact frequency percentages — e.g. "D# (155.6Hz) 18% [SEVERE DEFICIT] → Liver / LV · Wood element" — and performs a full TCM differential diagnosis. If it responds, your results silently upgrade.
tryAPIEnhancement() → claude-sonnet-4-20250514 → silent result upgrade if successful
08
Report Generation
Your results populate across 7 tabs: VoiceMap™ radar, 12-note spectrum, Chladni geometry patterns, organ meridian vitality, chakra resonance assessment, 13-track sound protocol, and Personal and Clinical wellness reports — all rendered from your frequency data.
VoiceMap · Spectrum · Chladni · Organs · Chakras · Protocol · PDF Reports
07 · Sacred Geometry & Live Cymatics

Chladni Figures — The Geometry
Your Sound Creates in Matter

When you vibrate a plate covered with sand at a specific frequency, the sand migrates to the nodal lines and forms geometric patterns. These are Chladni figures. Every frequency creates a different, reproducible, unique geometric shape. CymaVoice renders these in two contexts — your scan results, and the Live Cymatics visualiser during protocol playback.

The Scan Result Chladni Grid

During and after your scan, each of the 12 organ notes is rendered as its own Chladni pattern. The brightness of each pattern corresponds to the strength of that frequency in your voice — a dim pattern means that organ frequency is deficient; a bright high-contrast pattern means it is strongly present. The pattern equation uses sine-based mode functions:

Scan Grid — Chladni Pattern Algorithm
val = |sin(m·u + φ) · sin(n·v) − sin(n·u) · sin(m·v + φ)|
brightness = 1 − min(1, val × 3)
φ = amplitude × π
m, nHarmonic mode numbers — unique per chromatic note, mapped by frequency order
u, vSpatial coordinates mapped across the canvas (-1 to +1)
φPhase angle — scales with your measured frequency amplitude

The Live Cymatics Visualiser

When you tap Visuals during protocol playback, a full-screen live Chladni sand particle simulation launches — driven by the actual audio coming from the player in real time.

Particle Count
4,000
Sand Particles
4,000 individual particles distributed uniformly across the canvas. Each particle follows real Chladni physics — pulled toward node lines by the gradient of the standing wave equation.
Pattern Library
40
Unique Geometries
40 distinct (m,n) mode pairs ordered by frequency using Chladni's law: f ∝ √(m²+n²). Simple patterns at low Hz, ultra-complex patterns at high Hz — matching the physical reality of resonance.
Detection
Live
Real-Time FFT Pattern Selection
Web Audio API reads the dominant frequency from the playing track every frame. That Hz is mapped to the closest Chladni mode using √(m²+n²) — the pattern shown IS the frequency you are hearing.
Response
60fps
RMS-Driven Particle Physics
Particle movement intensity scales directly with the RMS amplitude of the audio. Beats scatter particles outward; silence lets them settle into precise node line geometry. Slow pulses = slow movement. Fast beats = rapid scatter and snap.
Live Visualiser — Paul Bourke Chladni Equation
f(x,y) = cos(m·π·x/R) · cos(n·π·y/R) − cos(n·π·x/R) · cos(m·π·y/R)
m, nMode numbers selected by live dominant frequency: rank = Hz / 55, mapped to closest PATTERNS[i]
x, yParticle position normalised to canvas radius R (-1 to +1)
f=0Node lines — where sand settles. Particles are pulled toward these zero-crossing surfaces

The 40-Pattern Frequency Library

All patterns ordered by f ∝ √(m²+n²) — lower frequency = simpler geometry, higher frequency = more complex. This matches the physical behaviour of real Chladni plates.

# m n Freq Rank √(m²+n²) Node Lines Geometry
1 1 2 2.24 2 Simple cross
2 1 3 3.16 3 3-spoke star
3 2 3 3.61 4 4-petal flower
4 1 4 4.12 4 Grid with diagonals
5 2 4 4.47 5 5-section radial
6 3 4 5.00 6 6-section complex
7 1 5 5.10 5 Fine 5-spoke
8 2 5 5.39 6 Lobe pattern
9 3 5 5.83 7 Complex 7-fold
10 1 6 6.08 6 Fine 6-spoke
· · · continuing through 40 unique geometries to [2,11] at rank 11.18 · · ·
Pattern = Frequency: The geometry you see during protocol playback is not decorative — it is mathematically determined by the dominant frequency the analyser detects in the audio at that exact moment. When the track changes to a new organ frequency, the pattern morphs smoothly to the corresponding Chladni mode. You are watching sound organise matter in real time.

Ready to Scan? Your Vocal Blueprint Is
Waiting to Be Decoded

30 seconds. The most advanced vocal frequency analysis outside a clinical lab. Free to scan — $3 for your full protocol and reports.

🎙️ Start My Free Voice Scan →

CymaVoice™ by CymaTones · Grounded in cymatics research