Bioacoustic Research

The Animal Sound Archive at the Museum für Naturkunde Berlin is one of the oldest and largest collections of animal voices worldwide. It was founded in 1951 for scientific research in the field of behavioural biology and comprises now 120.000 sound recordings. The major part of the old analogue tapes has been already transformed in digital format. Digital recording and processing techniques open up new fields of application.

This open-reel tape recorder was used for the first audio recordings in the 1950th. Weighting 40 kg and depending from a power plug the device hardly could be used for recordings of animals in the wild.

Through acoustic analyses animal vocalizations can be objectively described in the frequency and time domain. Even old sound recordings, like the oldest wildlife recording in the archive, have a good quality and are suitable for acoustic analysis. This recording of a tawny owl was made on October 30th 1951. Only the running noise of the tape recorder is a bit annoying.

Bird vocalisations very often have a very complex acoustic structure like the song of a chaffinch.

But there also appear monotonous songs as the buzzing sounds of a Savi’s warbler.

The song of the nocturnal Spotted Crake, consisting of whipping notes, is very easy to recognize.

The sound recordings stored in the Animal Sound Archive provide us with an excellent reference material, enabling us to apply modern methods of acoustic signal analysis and detect automatically bird species within a very complex soundscape. We focus on application scenarios, where an acoustic approach has significant advantages over the acquisition by a human observer.

In Mecklenburg, numerous former bogs are being restored. Where 10 years ago were meadows covered the landscape, today a large wetland provides habitat for many rare marsh and aquatic birds.

In the wetland area calibrated long-term acoustic recordings were created using multi-channel recorders.

A solar-powered autonomous unit enables automatic recording of ambient noise.

Many bird species are only active when the sun has set. The natural nocturnal soundscape is very complex.

It is a challenge to reliably detect individual calls in this cacophony.

Methods of acoustic pattern recognition enable to reliably detect individual calls of Spotted Crakes (SpCr) in the spectrogram. As template clean calls were used from the Animal Sound Archive were used. The marks were set automatically by the program (Avisoft SASLabPro, © Avisoft). The automated analysis allows investigate even several days of continuous recordings.