Drafted by Shane Andres at WN’s Documentation-palooza 2k23
The audio team is focused on putting together hardware and software in order to collect sound data from hives. Our goal this year was to use a raspberry pi (rpi) to record the sounds off of multiple microphones simultaneously, so that we can put them in a test hive and get lots of data.
As of now, we are focusing on putting many microphones in a hive. This would allow us to try and triangulate the queen bee, which is useful for beekeepers as finding the queen bee is normally a long and tedious process. We have been thinking about working together with the camera team down the road to develop this.
Having microphones in the hives also allows us to try and understand what different sounds mean in terms of hive activity. For example, when a hive is swarming (half of the bees leave to create a new hive), a certain frequency is produced which can help beekeepers know about this in advance.
Status of Project
We are currently using a rpi 3 with many (identical) usb microphones. While using one microphone is simple, using multiple microphones has proven to be much more difficult. Each usb microphone has its own sound card, making it very difficult to record all of them at once. We have created a software solution which links the microphones together into one audio device in linux, (this new audio device created from two microphones is stored in the rpi’s asound.conf file). Since each microphone has its own sound card with its own clock, there is the possibility that over time, each microphone’s recordings will drift from each other. We have not tested this as of yet. Also, since the microphones we are using are identical, the rpi is having difficulty differentiating them, and wherever the rpi is booted, each microphone is mapped to a different port.
After talking to Professor Schmitz and a former member of the team (Jimmy He), we learned that the first test hive created by Wagglenet actually used a mac, which enabled them to avoid this problem as the mac could distinguish physical usb ports from one another, meaning that each microphone was always mapped to the same software port. However, we have not been able to get our hands on their code as of yet, which is something we think would be very useful. Professor Schmitz is unsure of where the old mac is, and says it may be in the Openlab somewhere.
As of now, the people most able to answer questions about the project are probably myself (Shane Andres), Richard Jia, Professor Schmitz, and Jimmy He. I am familiar now with the problems associated with trying to use multiple sound cards on linux. Richard has been working on the data processing side of things, and knows about the algorithms and techniques involved with that. Jimmy He worked on all aspects of the project when he was a student here, and he knows a great deal about the entire project.
What do I need to know to join the team?
Familiarity with linux and the command line, as well as basic coding experience is definitely desired. In the future, we plan on coding in python for higher level audio analysis.
Where is our stuff?
All of our team’s stuff is in a clear plastic tub with a white lid underneath the Wagglenet desk. It is filled with microphones and a speaker. There should also be an old observation hive next to the Wagglenet desk, which may be a good reference for microphone placement.