Music Is Quasi-Rhythmic Too: Comparing Bandwidth in Speech and Music Acoustic Modulation Spectra.

Speech and music both unfold over time. However, these dynamics are perceived quite differently: music as rhythmic, and speech as quasi-rhythmic. Here, we aim to identify whether this distinction between rhythm and quasi-rhythm is sourced from the raw acoustic waveform by analyzing the modulation spectrum of the acoustic envelope. We analyzed several large corpora on three scales: the coarsest scale containing recordings of each corpus, the intermediate scale of individual speakers/songs, and the finest scale of individual sentences or short musical segments. We confirm previous findings that speech and music modulation spectra differ in their center frequency, which reflects how quickly the sound envelope fluctuates, but find that the modulation bandwidth, which captures how periodic the speech/music envelope was, is comparable for music and speech. Furthermore, the bandwidth is largely preserved across the three scales, indicating that the corpus-level temporal irregularity is dominated by the irregularity within a few seconds of speech/music recordings. These results demonstrate that the perceived rhythmicity of music may not directly reflect the temporal periodicity present in its acoustic structure.