The vocal tract is modelled as a chain of N = 20 cylindrical tube segments running from the glottis to the lips, each with cross-sectional area Ai(x) set by the constriction's position and degree, and by lip rounding. Sound propagates through each segment as a lossy transmission line:
Z0(x) = ρc / A(x) (characteristic impedance)
γ = α + jω/c (propagation constant, ω=2πf)
Zin = Z0 · (ZL + Z0·tanh(γl)) / (Z0 + ZL·tanh(γl))
Chaining this impedance transform segment-by-segment from the lips (small radiation load ZL) back to the glottis gives the total input impedance Zin(f). Because the glottis behaves as a high-impedance volume-velocity source when the vocal folds are closed, the tract's resonances — the formants — are exactly the frequencies where |Zin(f)| peaks. The simulator sweeps f from 150–4000 Hz, solves the chain at every step, and picks the first three peaks as F1, F2, F3.
- Dragging the dot toward the centerline narrows the constriction (lower "tongue height" value) — this lowers F1 (F1 tracks tract openness at the constriction).
- Dragging the dot left/right slides the constriction from the pharynx toward the lips — a front constriction raises F2, a back one lowers it (F2 tracks tongue backness).
- Lip rounding shrinks the area of the last two segments, adding effective length and lowering all formants — this is why /u/ sounds "darker" than /i/.
- Tract length rescales the whole tube; since resonant wavelengths scale with tube length, every formant shifts by 1/L — a child's ~13 cm tract sounds noticeably higher-pitched in timbre than an adult male's ~19 cm one, independent of pitch (F0).
The chart plots the live (F2, F1) point against the five cardinal vowels' textbook adult-male formant values inside the classic IPA vowel quadrilateral (/i/ /u/ close corners, /ɑ/ open corner) and reports the nearest one by weighted distance — the same F1/F2 space linguists use to describe vowel quality.