Characterization of Piezoelectric Speakers in Smartphones for Audio Signal Processing

In applications involving portable devices, piezoelectric speakers present different advantages over traditional electrodynamic speakers, including a thinner profile and higher efficiency. At the same time, their use is somewhat hindered by well-known drawbacks, such as poor bass reproduction and high harmonic distortion. In this paper, we present an approach for the characterization of the acoustic behavior of smartphones mounting piezoelectric transducers underneath the screen. Firstly, we describe how to carry out measurements to identify the nonlinear response of said complex multiphysics systems. Then, as an example of application, we show how the acquired data can be used to train a neural network to obtain a lightweight digital model of the device under test. These are two crucial steps toward the development of linearization algorithms or, more generally, algorithms for optimizing the device acoustic response.