A little sabbatical

So, this project has been a put on a shelf for a while, but what’s happened in 4 years?

Several things–most notably we’ve moved out further into the suburbs, and the loud sounds seem to have mostly abated. Whereas at the old place, I was pretty much guaranteed a loud sound every few days, but here I’m limited to a few clusters of booms around the 4th of July and New Year’s Eve, with very little else in the way of sample data. I don’t think my new neighbors would appreciate my setting off fireworks just to have sample data. I like to tell myself that the sounds at the old place were fireworks, but the reality was more likely a mixture of gunshots and fireworks.

Second, I’ve just lost steam on this. The problems with getting the timing right and other things like that have convinced me that maybe I need to re-think the microphone design. I’m thinking of using a dedicated GPS module on each microphone unit to establish both location and time more accurately, but that pushes up the cost considerably. GNSS/GPS receivers can be had for $30 or $35 dollars on convenient breakout boards, but that nearly doubles the cost per microphone. At that point maybe I should consider using a Raspberry Pi Zero 2W or similar, though the lack of analog inputs is problematic.

Goals and Anti-Goals

I want to clarify what I’m trying to accomplish, and what I’m specifically not trying to accomplish. Commercial versions of this technology do a sophisticated analysis of the sounds they detect, so that they can do things like distinguish between fireworks and gunshots using shock wave profiles and such.

At least for this first iteration, I’m not looking to do anything nearly that complex. I just want to be able to pinpoint any loud sound in a reasonable range. Future iterations may do something as subtle as attempting to disentangle overlapping sounds, or deal with long continuous sounds (like a running motor, jackhammer, etc.), but those are stretch goals.