Synthesizing bird song.

In this work we present an electronic syrinx: an analogical integrator of the equations describing a model for sound production by oscine birds. The model depends on time varying parameters with clear biological interpretation: the air sac pressure and the tension of ventral syringeal muscles. We test the hypothesis that these physiological parameters can be reconstructed from the song. In order to do so, we built two transducers. The input for these transducers is an acoustic signal. The first transducer generates an electric signal that we use to reconstruct the bronchial pressure. The second transducer allows us to reconstruct the syringeal tension (in both cases, for the time intervals where phonation takes place). By driving the electronic syrinx with the output of the transducers we generate synthetic song. Important qualitative features of the acoustic input signal are reproduced by the synthetic song. These devices are especially useful to carry out altered feedback experiences, and applications as biomimetic resources are discussed.