An ultrasound transducer converts electrical energy into acoustic waves. Its internal structure — piezoelectric elements, matching layers, backing materials, geometry — determines the wavefield it produces. Modeling that wavefield accurately requires knowing the transducer's construction in detail. But transducers are commercial products with proprietary designs. You don't know what's inside. Burchner et al. (arXiv: 2603.24415) solve this by working backward: reconstruct an effective transducer model from the wavefield it produces.
The method is distributed source inversion. Measure the wavefield in an aluminum test sample using laser vibrometry. Then solve the inverse problem: what spatio-temporal source distribution would produce the observed field? The reconstructed source captures phase and amplitude variations across the transducer aperture — the effective radiation pattern — without any knowledge of the transducer's internal mechanics.
The effective model reproduces experimental directivity patterns accurately. More importantly, when used in downstream applications — reverse time migration, full-waveform inversion — the accurate transducer model significantly improves imaging quality. The transducer characterization bottleneck, which limited these techniques, is removed by inverting the problem.
The through-claim: you can reconstruct what something is from what it does. The transducer's internal structure is inaccessible, but its external behavior fully constrains an effective model. The reconstruction doesn't recover the true piezoelectric layout — it recovers a mathematical object that produces the same wavefield. The effective model is not the transducer; it's an equivalent source that makes the same sound. The distinction between “what it is” and “what it does” dissolves when only the output matters.
Burchner, Schmid, Rank, Kollmannsberger & Fichtner, 2603.24415. Ultrasound / inverse problems / transducer modeling / full-waveform inversion / distributed source.