My project is a wearable, a vest with ~1400 “pearl” LEDs embedded in it.
The last version of the project had its microphone mounted to the control box which normally sat in an inside pocket. I didn’t find it worked very well—with the jacket sitting motionless on the back of a chair in the living room “listening” to my home stereo system everything was fine, but in a festival situation when it was brushing the inside of the pocket and dealing with large amounts of crowd noise and a 70kW PA I could not get anything like a useful signal.
Would mounting a dynamic Lavalier mic on my collar and plugging it into the PixelBlaze expansion board give me a meaningfully better result?
I’ve used this mic with built-in AGC and preamp on other projects, and it performs well. It outputs a line-level signal, so you could plug it right into the sensor board line-in, but you would also need to run a 3.3V connection from the sensor board as well. The AGC allows you to go between your living room and a giant dance-floor and the audio gain will automatically adjust as needed. Crowd noise might still be a problem if it’s really louder than the sound system, but it’s worth a try.
AH! That was where I was confused—it’s a line-in socket, not a microphone input.
Does the sensor board’s built-in mic do all the right stuff vis. amplification and auto-gain? It’s going to be inside a case of some kind so I might as well leave it in a pocket, it’s not going to be rubbing against anything.
The sensor board does not perform auto-gain, but is done in most patterns instead. Most of the included and shared sound patterns have a PI controller that does auto gain control, which may benefit from some tuning for your environment.
I’ve also been pondering sticking a full-size controller and a sensor board into a little container with an amplified mic to act as a sync leader not actually connected to any pixels, but it’d make me nervous to drop $100 in parts (after the battery and various breakouts needed for charging and amplification) next to a speaker at a party. I’d love to rig a seeeduino xiao to one of these breakouts with a little battery and ship the data to a single node or the sync leader over udp.
Edit: Ben just pointed me at the fact that the sensor board code is open source, I think I’m gonna build this. It’s a great chance to try Arduino dev in rust.
I’ve done sound in musical theater on and off for many years now, and have run a lot of lavalier mics in a lot of strange situations.
Here are some things I’ve learned that might be useful when building your own setup:
anything rubbing against the cable (not just the mic) makes a lot of noise. You need to tape or sew it so it stays stationary relative to the costume, or whatever its attached to.
use a windscreen. Even a small foam screen helps kill stray low frequency transients. (For noisy/windy environments, a big fuzzy one (a small piece of fake fur will work in a pinch) is even better.
if you’re building, consider adding a fairly steep high pass filter at around 80hz. It’ll reduce rumble from the crowd, and you’ll still get the big bass hits from the music.
also, chopping off the highs above 4k or so is reasonable for dance music. We’re trying to extract enough information from the music to make interesting lights, not record it with perfect fidelity!
