Can the human voice help us identify a jaw-harp tuning that will feel naturally compatible? Not by looking for a single “magic note”, but by combining two layers: the fundamental frequency of a relaxed voice and the way the mouth actually shapes the instrument’s sound through different vowels.
We did not want to leave that question at the level of impression. For one player, we collected 224 recordings of the voice, calibration plucks and jaw harps played in the mouth using the vowel shapes U–O–A–E–I. The result was striking: A#1, almost exactly one octave below the centre of the relaxed voice, also produced the most balanced formant match of every tuning tested.
This supports the octave hypothesis, but it does not prove a universal rule.
Yes: this experiment did test the link with mouth resonance across different vowels. A#1 did not perform well only because of its fundamental-frequency ratio. It also achieved the highest overall F1/F2 match and was the only tuning to retain a usable partial score across all five vowel shapes. That is a strong convergence of two independently evaluated layers, not universal proof for every player.
This is not a universal A#1. It is a personal tuning range.
The most important conclusion is not “A#1 is the right jaw harp”. A#1 is the result for this one player. The more interesting question is whether each person has a tuning region in which a jaw harp’s harmonic series meets their vocal system more naturally.
The voice combines a source and a filter. The vocal folds generate the fundamental frequency F0; its momentary value depends mainly on their length, mass and tension. The pharynx, mouth, tongue, jaw and lips form a changing resonant filter expressed through formants. These two layers are neither identical nor rigidly locked together, but they develop and operate within the same individual anatomy and interact during phonation.
Each player may have a personal candidate range for jaw-harp tuning. Roughly half of their stable relaxed F0 provides an octave-based harmonic foundation; their own U–O–A–E–I formant map then tests whether the instrument is actually compatible. Our result supports this hypothesis in one person. Measurements from more players are needed to show how often it holds and how wide the useful range may be.
What we measured
The study had three layers: first the relaxed voice, then the instrument on its own, and finally the instrument inside the changing resonant space of the mouth. The third layer is what tests whether a promising frequency relationship also survives articulation.
The main voice set contained three repetitions of each of the five vowels; two earlier recordings were pilot attempts. Calibration covered 14 physical instruments with three plucks each. The in-mouth set contained 11 separate tuning profiles, each recorded with five vowels and three repetitions: 11 × 5 × 3 = 165 recordings.
Interpretation note: several physical instruments with the same tuning share one in-mouth profile in this dataset. The results therefore cannot support claims about differences between individual pieces or makers within the same tuning.
For English readers: U–O–A–E–I refers to five pure vowel shapes, not English letter names or diphthongs. They function here as repeatable mouth positions for formant filtering.
Why we began with half the voice frequency
The fundamental frequency, F0, determines the perceived pitch of a voiced sound. Across the five vowels, the measured centres ranged from 114.84 to 119.12 Hz; the median of the main set was 115.36 Hz.
The difference was approximately −24 cents from an exact octave ratio, roughly one quarter of a semitone. Acoustically, this is plausible: the second harmonic of the lower jaw harp reaches the region of the voice’s fundamental, while the higher harmonics continue through a range in which the vocal tract is accustomed to working.
That relationship is not enough by itself. The mouth is not a passive resonator. Movements of the tongue, lips and jaw emphasise different parts of the harmonic series. We therefore treated the octave relationship only as a way to select a candidate range, which then had to pass a second test: formants across U–O–A–E–I.
Think of the jaw harp as providing a harmonic ladder at approximately 58, 116, 174 and 232 Hz. Its second harmonic meets the relaxed voice around 115 Hz. This identifies a starting region, not a rigid prescription.
The second test: mouth resonance across U–O–A–E–I
A vowel is created mainly by changing the resonances of the vocal tract, not by changing the fundamental frequency. The two strongest resonant bands are called F1 and F2. When the resonant peaks of the voice and jaw harp move closer together under the same mouth shape, the instrument is making better use of a region that this player can produce naturally.
For the voice, we reduced the harmonic “teeth” using cepstral smoothing and estimated the spectral envelope. For the jaw harp, we worked with peaks in a lightly smoothed spectrum. We then compared the F1 and F2 differences for each vowel and averaged those similarities across all five vowels.
| Vowel | Voice F1 / F2 | A#1 F1 / F2 | F1 / F2 difference | Partial score |
|---|---|---|---|---|
| U | 394 / 644 Hz | 351 / 818 Hz | +43 / −174 Hz | 69 |
| O | 553 / 803 Hz | 406 / 875 Hz | +147 / −72 Hz | 54 |
| A | 618 / 900 Hz | 580 / 818 Hz | +38 / +82 Hz | 88 |
| E | 547 / 1552 Hz | 528 / 1049 Hz | +19 / +503 Hz | 49 |
| I | 220 / 2111 Hz | 295 / 1907 Hz | −75 / +204 Hz | 52 |
The strongest match appeared at the open A. The second formant diverged more for the front vowels E and I, but the overall profile remained more balanced than for any other tuning. That breadth across five mouth shapes matters more than one exceptionally good point.
Why C2 can look good in one score and fail in another
An earlier exploratory assessment also rated C2 highly. That is not an error or a contradiction. The two scores answer different questions.
Exploratory composite score
Rewards pronounced differences between vowels, F1/F2 range, energy and harmonic distribution. It can indicate that an instrument is colourful and responds clearly to articulation.
Direct formant score
Compares the F1/F2 position of a specific voice and tuning under the same vowel shape. It asks about personal resonant proximity.
C2 may therefore be excellent for clear articulation, teaching or ensemble playing, yet it did not reproduce this player’s own formant map as evenly as A#1. Its resonant peaks moved far away from the voice at O, A and I.
What players’ experience adds
Alongside the measurements, a practical pattern is beginning to recur in feedback: some players describe tunings around half of their relaxed voice centre as particularly comfortable. This is useful support for the working hypothesis because it comes from actual playing rather than a graph alone.
This feedback is not yet a blinded or statistically evaluated trial. Measurement conditions, experience, instrument construction and expectations were not standardised across all participants. We can say that these reports are consistent with the hypothesis. We cannot yet say that they prove it.
The voice is not a point. It moves around a personal orbit.
A single average can suggest that we are looking for the arithmetic midpoint between a person’s lowest and highest notes. Repeated morning measurements track something else: the range to which the voice repeatedly returns under calm, comparable conditions. In Dynamic Coherence terminology, we call this a personal orbit — not a fixed pitch or an ideal performance, but a reference region of return.
In observations so far, this region does not sit in the middle of the full available vocal range. It lies lower, closer to the relaxed morning voice. F0 can move in either direction from that region. A higher value often accompanies greater activation, effort, stress or excitement. A lower value may appear before the morning voice has warmed up, during fatigue or with some vocal and health changes. The number alone is not a diagnosis: time of day, hydration, measurement technique, vocal load and many other factors also matter.
Lower than usual
A possible sign of reduced activation, an unwarmed morning voice, fatigue or change in the vocal system. Meaning comes from the individual trend and context.
Personal orbit
A repeatedly available relaxed range. It is neither one exact frequency nor a compulsory target, but an individual reference region of return.
Higher than usual
Often greater activation, effort, focus, enthusiasm or excitation. Here too, one number cannot explain the cause.
Why use the Morning humming check-in?
The Morning humming check-in repeats a short voice measurement under similar conditions and saves several pitch snapshots together with context about sleep, load, mood and the state of the voice. A single day can mislead; a series can show whether a more stable relaxed range is actually emerging and how the voice moves away from it in different states.
This is a second, independent line of enquiry. The formant experiment asks whether a particular jaw harp meets the personal resonances of the mouth. The Morning humming check-in asks whether a repeatable personal voice reference exists at all. Neither method currently proves the other, but if both point towards the same individual range, the hypothesis gains a substantially firmer foundation.
A jaw harp tuned in relation to a person’s vocal range might do more than feel pleasant. Its stable fundamental and harmonic series could act as an external resonant support around which breathing, the mouth, hearing and the vocal system can reorganise more easily.
In other words: a well-chosen jaw harp might help a player return towards their own orbit. That is a direction we want to measure, not a proven therapeutic effect.
Find your voice reference first
One measurement shows a momentary state. Repeating the Morning humming check-in begins to reveal your own relaxed range and its changes.
Start the Morning humming check-in Read the measurement methodThe practical result is not one exact note but a sensible candidate range around half of the relaxed voice centre. The nearest available tuning and its neighbours should then be compared by ear, by the sensation in the jaw, by how readily the instrument responds and by whether its tone colour can be shaped without forcing.
Treat “voice ÷ 2” as the centre of a small group of candidates. Compare the closest tuning and its neighbours by sound and feel. The method narrows the search; it does not declare one absolute winner.
What the result shows — and what it does not
What we can say
- in this case, the octave candidate was supported by an independent formant analysis;
- A#1 produced the highest and most balanced direct match among the measured tunings;
- five vowel shapes add more information than F0 alone;
- different metrics can honestly recommend different instruments for different purposes.
What does not follow
- A#1 is this player’s result, not a shared note for everyone; the broader hypothesis concerns each person’s individual tuning range;
- a score of 62.2 is neither 62.2% certainty nor a measure of instrument quality;
- formant proximity alone does not measure comfort, loudness or playing technique;
- one case study cannot replace measurements from many people.
How the hypothesis should be tested next
The next step is not a more complicated graph but a better experiment. The same tunings should be tested by people with different voice ranges. Instruments should be presented in random order without revealing the expected match. Alongside F0 and formants, participants should rate comfort, pleasantness, control of the vowel shapes and whether they want to play the instrument again. Repeating the test on another day would show whether the result survives a change of momentary state.
Only then can we estimate how wide a practical recommendation range should be, how strongly maker and construction influence the result, and whether subjective preference systematically follows the formant score.
A practical conclusion for choosing a jaw harp
The most honest conclusion today is this: a jaw harp roughly one octave below the relaxed voice may provide a suitable harmonic foundation, but its practical usefulness must still be confirmed by mouth resonance across different vowels and, ultimately, by the player.
In our measurements, both layers converged on A#1. That is more than an impression and less than a general law. It is a measurable, repeatable way to reduce a large field of instruments to a smaller and more meaningful group of candidates.
Start with the voice. Decide with your ears.
Measure your relaxed voice centre, explore the surrounding tunings and treat the result as the beginning of listening — not an algorithmic verdict.
Find your jaw-harp tuning from your voice Compare other ways to chooseFrequently asked questions
Should a jaw harp be tuned to exactly half my voice frequency?
Not necessarily. Half of the relaxed voice centre is a practical starting point, not a prescription. Mouth resonance, playing technique, instrument construction and personal sensation still determine whether the result is useful.
Does this prove that A#1 is the best tuning for everyone?
No. A#1 was the result for one particular player. The broader hypothesis is that each person may have a different candidate range derived from their relaxed voice and tested through their own U–O–A–E–I resonances.
What are F1 and F2?
They are prominent resonant bands of the vocal tract. Their positions change with the tongue, jaw and lips and help create differences between vowels. Here we compared how close those bands were in the voice and in a jaw harp played with the same mouth shape.
Why did C2 perform well in another assessment?
Because that assessment asked a different question. The exploratory composite score rewarded distinct vowel articulation and the distribution of spectral energy. The direct formant score asks how closely a particular voice and jaw harp align in F1 and F2 under the same vowel.
What does “personal voice orbit” mean?
It is a working term for the relaxed range to which a person’s voice repeatedly returns. It is not one exact frequency, a diagnosis or a compulsory target. The Morning humming check-in helps distinguish this personal reference from a one-off fluctuation.
Method notes and sources
The formant score uses a smoothly declining similarity based on the F1 and F2 distances, with working widths of 80 Hz for F1 and 150 Hz for F2, then averages both bands across the five vowels. It is an internal comparative metric developed for this measurement. Pitch detection can make octave errors on strongly harmonic sounds, so suspicious results were checked against the nominal tuning and the spectrum.
- Besnainou, Frelat & Mamou-Mani: experimental study and acoustic model of the jaw harp — Acoustics ’08 proceedings (PDF).
- Peterson & Barney: classic work on measuring the fundamental frequency and formants of vowels — Control Methods Used in a Study of the Vowels (PDF).
- Skarnitzl & Volín: reference formant values for Czech vowels — Institute of Phonetics, Charles University.
- Titze: physical account of interaction between the voice source and resonant filter — Nonlinear source-filter coupling in phonation.
- Garrett & Healey: changes in voice F0 at different times of day and substantial individual variation — An acoustic analysis of fluctuations in the voices of normal adult speakers.
- Systematic review of the relationship between acute stress and voice fundamental frequency — The Fundamental Frequency of Voice as a Potential Stress Biomarker.