Skip to content

Latest commit

 

History

History
91 lines (66 loc) · 3.24 KB

README.md

File metadata and controls

91 lines (66 loc) · 3.24 KB
Raw

Intercom for multiple Raspberry Pi in the same network

Purpose

I want to be able to talk to the people in other rooms just by pressing a button. It's a fun weekend project. Might be useful for the cases where several people in the family work from home sitting in different rooms with closed doors. My experience shows that pressing an intercom button on a solderless breadboard is much more fun than shouting!

Setup

  • 2 (or more) Raspberry Pis with some kind of microphone and speaker connected. Both should work with ALSA. Basically, arecord and aplay commands should work. For arecord, you can pass the specific device name (-D option, see src/constants.ts). For aplay, the code expects the default device to work (adding -D option in a similar way won't be hard if you need it).

  • A button connected to GPIO pin 40 (GPIO21). Acts as push-to-talk button. The pin number is configurable in src/constants.ts.

  • All Pis should be in the same network since the code relies on mDNS to work. Avahi daemon should be on (it's on by default in Raspbian).

  • The code uses rpio Node.js package. Follow their requirements to setup your Raspberry Pis.

  • The ./chime.wav mentioned in src/constants.ts is played when a button is pressed, and when a message is received. It's not included, but you can make one for yourself, e.g. by taking one of the Freedesktop OGG files and converting it:
    ffmpeg -i /usr/share/sounds/freedesktop/stereo/message-new-instant.oga -t 0.6 chime.wav

Usage

On all Raspberry Pis connected to the same network:

$ npm install
$ npm start

Enjoy your intercom!

Implementation details

Written in TypeScript for no specific reason (Python might be more popular choice for Raspberry Pi programs, but I really like Node.js event loop in cases like this one where we need to pipe the data from arecord to an HTTP connection and then pipe it to aplay on the other side).

Note: there is no error handling by design. We log and ignore all HTTP errors. If anything else is wrong, we crash. Run the program from a shell script:

#!/bin/sh

export PATH=/path/to/nodejs/bin:$PATH
while :; do
  npm start
done

This script can be put in crontab @reboot (redirecting output into a log file). In this particular case, with Raspberry Pi, GPIO button, and mDNS, it's probably better to crash and restart (and reannounce ourselves) in any weird case rather than try to recover from an error.

Logs (printed by console.log here and there) are helpful to debug what's going on. Also, curl http://localhost:4242/debug shows what the given instance knows about itself and its neighbors.

Tests

Not at this time. At this point, I can either make a bunch of change detector tests, or spend a week properly injecting all dependencies (GPIO, mDNS, HTTP) which just does not make any sense for a fun project because it's not fun. Contributions are appreciated :)

Copyright

I work for Google, so the code has Google copyright. But see the following disclaimer:

Disclaimer

This is not an official Google project.