Analysis of Ultrasonic Components in Voices of Chosen Bird Species

Ultrasonic components of bird voices are rarely analyzed in bioacoustical papers. The main reason behind this situation is an equipment used for digital recording of bird voices in which: 1) sampling frequencies do not usually exceed 48 kHz and 2) microphones that limits bandwidth of recorded bird voices are also commonly used. The experiments described in this paper were carried out with 96 kHz sampling frequency and high quality broadband microphones. Moreover, because the commonly used 1-bit sigma-delta A/D converters introduce some distortion to the signal – they particularly affect its high frequency components. In the research we propose using the SAR converter instead. Preliminary automatic recognition research comparing performance of SAR and 1-bit sigma-delta A/D is also presented in the paper. Investigation concerned analysis of several voices of passerine birds e.g. Parus major, Passer montanus in ultrasonic range which unveiled existence of strong ultrasonic tonal frequency components and formants in the analyzed bird voices. Methodology that allowed us to record bird voices in an ultrasound range was described. Additionally, the general conclusion outlining the best bird voices recording strategy has been drawn. Biological meaning of the ultrasound components in bird voices still remains an open research question.