I resumed work recording my stomach by trying to create a better connector for my microphone to my stethoscope. I bought some kind of brass pipe coupler and taped it to the mic. On the other end, I retrofitted my old connector to fit into the brass pipe, allowing the stethoscope tube to be threaded through while blocking out some of the outside noise. This seemed a little better but not good enough.
Next I tried a lavalier mic inside the stethoscope tube. The mic’s receiver would be fed into my iphone using a mini audio interface. But alas, the mic wasn’t able to pick up th subtleties of sound.
Finally, I tried a piezo mic. It was better than I expected but still not sensitive enough.
I asked Eric Rosenthal for help and he pointed me to a graphene based substance that can be homemade and is very sensitive to vibrations. This could be turned into a mic, but the materials were still too expensive, and it felt too late to begin creating a custom rig to utilize this analog input on a microcontroller and log it successfully.
Throughout all this, I stopped wearing my apple watch to record my heart rate. At first I just forgot, but then I decided the data was too boring and even I couldn’t see a clear narrative in the information.
So as I was playing around with the lavalier mic, I wondered how much I talk in a day… My voice felt like a compelling output to monitor. So I first tried mic’ing my throat literally… this didnt really work with the lavalier.
It sorta worked with the piezo mic but it felt creep to tape a mic to my throat. So instead, I used my iphone’s headset to start recording.
The process felt awkward at first; it was weird knowing I, and my surroundings, were being recorded [weirdly enough, almost no one commented on its presence]. But after listening back to it, I loved reconstructing my memory through sound. I recorded as often and as long as I could, logging over 19 hours of recording.
The next step would be to do a number of different analyses on the audio itself. Voice Activity Detection seems to be a readily available technology that could allow me to calculate how much I spoke on my recordings, and maybe let me see how much I talk relative to the other voices present. I could have used a piezo to simply measure the vibrations on my throat, which would’ve allow me to avoid the post processing work to figure out how much I talk…. and with more time I would have considered that option. But I’m enjoying the experience of archiving my life in sound.