Audio Streaming and radio shows
Report: Justin Bennett
Between November 2024 and July 2025, we have been streaming sound every week from a different location in the Binckhorst. The apparatus is housed in a small plastic waterproof box. Two omnidirectional microphones are situated on each side of the box, approximating the position of ears on a human head. This gives a very realistic stereo rendering of the soundscape when listened to on headphones. These microphones are amplified and digitised with an audio interface connected to an small computer (Raspberry Pi Zero) which uses a mobile modem with a Data SIM to connect to a remote server. We used the server hosted by the organisation Locus Sonus, based in France. A powerbank provided electricity for at least 24 hours. Listeners connected with the stream by opening the URL web address in a browser. We printed flyers with a QR code that takes smartphone users directly to the stream when it is live.
The idea to use a streaming box came from listening to broadcast streams of environmental sound from the organisations Locus Sonus and Radio Aporee / Radio Earth. Locus Sonus have a map interface where one can choose to listen to live streams spread across the globe. Radio Earth on the other hand announce live streams on a Telegram channel, where listeners can comment on what they are hearing and interact with the people running the stream. The way people listened and engaged online with the broadcast sounds was very interesting to me. The fact that it is "live" seems to encourage a different listening - you are not listening to a recording, you can't rewind or skip, if you stop listening you might miss something and you perceive the soundscape evolving in real time. People leave the stream on while they're doing something else, listening over a long period to subtle changes. Although this remote "acousmatic" listening and the particular microphones used tends to abstract the sounds somewhat, the feeling of experiencing something live is quite different to that of listening to a recording. Because the streaming box is small and unnoticable and it's left in the same spot for a long time, small animals and birds come extremely close to it. This perspective is, again, very different from that of a sound recordist who is actually on site, whose presence always affects the environment.
Some ways we have used streaming:
- Weekly streamings of 24 hours, Thursday to Friday
- We tried to spread the locations over the whole area of the Binckhorst. We had to hide the box well as there is no way of securing it and we avoided places where it is possible to eavesdrop on people, although this is not always possible to predict, particularly at night.
- It was framed as a "live broadcast" which was promoted on social media and also spread once a month on a local web radio: Bug Radio.
- A game - with a group of students guessing where the streaming box is placed. Interestingly some students were sometimes able to pinpoint the places accurately - using a map to trace possible sources of the sounds that they were hearing.
- Live comments from listeners (from all over the place) mostly who don't know the area. This mostly happened on the radio.earth telegram channel, but also via instagram.
Some preliminary results:
You can hear how monotonous some places are - the constant traffic noise from the Binckhorstlaan or the motorway filtered and resonated through spaces. Others, shielded from the roads by buildings or simply distance, are much more dynamic. Also green spaces / water places are more dynamic in terms of sound with much less background noise and more wildlife. Bird identification software counted up to 22 species of birds in some locations, in others there was just one or two species. Even unappealing-looking scrubby wasteland seems to have more action / biodiversity than some planned gardens. I would like to think that audiodiversity = biodiversity but of course there are many more factors at play - constant loud sounds like busy roads inhibit interaction between humans and between non-humans but if there were voices or bird calls, they might be inaudible on the stream - buried in the noise.
As well as running the continuous live stream, I and/or my students generally tried to record sections of the stream periodically (with a self-developed a software application). For instance, we recorded 3 minutes every hour - sometimes more often at times of rapid change like sunrise and sunset or peak times for commuter traffic. Although sampling the sound once an hour seems a rough approach, we collectdc data about the soundscape and how it evolves over time that would be impossible to do manually, on-site.
Listening and looking at these recordings in chronological order gives a sense of how the sound changes on a longer timescale, for instance the volume of the background noise floor can be clearly seen on a waveform plot or spectogram. Although we were not always able to record for the full 24 hours in each location, we were able to build up a picture of how sound changes over time AND how it is distributed over the area.
