RTSP server with atom M5 stack

Hey all,

Looking for some insight if poss.
I’ve got a couple AtomS3’s laying around and have an istance of GitHub - tphakala/birdnet-go: Realtime BirdNET soundscape analyzer running on my server (addon : hassio-addons/birdnet-go at master · alexbelgium/hassio-addons · GitHub) amazing work from the guys doing the app and the one who’s ported it to HA.

As birdnet can use rtsp, and it so happens that I have the M5stack pdms i2s microphone, I was wondering if it’s poss to use esphome + microphone componant and spin up a rtsp server when I can point birdnet at?
There are other ways (micro rtsp server etc…) but i’m a fan of esphome and like the yaml approch.


Wow, thank you for drawing this to my attention. We live somewhere that may make this very fun.

Not sure on the rtsp, or whether we can fake a alsa solution.

I have been trying to figure out esphome audio all afternoon, as well as rtsp and other stuff, have up and then looked at the beta notes for 2024.5. I noticed this, which may (or may not) help us. Send/receive Voice Assistant audio via ESPHome native API by jesserockz · Pull Request #114800 · home-assistant/core · GitHub

Maybe it light be possible to trick the esphome into streaming via the camera stream :ESP32 Camera Web Server Component — ESPHome
I’ve seen people repurposing the rtsp lib on arduino for audio.