Presets
Playback
String
Layers
Model check
The science
Model
Transverse displacement y(x,t) of a stiff, lossy string with fixed ends: mu y_tt = T y_xx - EI y_xxxx - 2 mu sigma y_t. It is solved on 100 intervals with an explicit second-order finite-difference scheme; the time step is set from the highest grid frequency (dt < 2/omega_max) so the scheme stays stable.
What you see
- Ideal string: modes
f_n = n f1,f1 = sqrt(T/mu) / 2L. - Bending stiffness raises the partials:
f_n = n f1 sqrt(1 + B n^2),B = pi^2 EI / (T L^2)(piano-like inharmonicity; the orange curve in the third graph). - A triangular pluck at
pexcites modenwith amplitude2h sin(n pi p) / (n^2 pi^2 p (1-p)): plucking at the middle kills even modes, near the bridge brightens the tone.
Simplifications
One polarisation, small amplitude (linear), the bridge and nut are rigid so nothing is radiated: the real guitar body, string coupling and frequency-dependent damping are not modelled. The loss rate is one number for all modes. Displacement is drawn exaggerated (slider), real amplitudes are millimetres. The sound is the pickup signal of this same model, not a sample.
Checks (run on load)
(1) period from time stepping vs f1; (2) total energy of a lossless stiff string stays constant; (3) the numerical pluck projection matches the closed-form Fourier series. Artwork (guitar top, plectrum) was rendered in Blender (Cycles).