experimental/partial mumble voip client implementation targeted at embedded systems
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.



Rabarber is an experimental/partial mumble voip client implementation targeted at embedded systems. Specifically, the C.H.I.P. Pro is targeted, and it's starting to show in certain parts of the codebase, so far in lradc.c/h. It's also the first project where I try programming with alsa, openssl, or protobuf for that matter.

You'll need additional files to make it compile:

  • config_core.c and config_core.h from generic_config
  • mumble_pb.c and mumble_pb.h which are generated from the mumble protobuf protocol definition with protobuf-c
  • sll_meta.h from sll_meta
  • varint.c and varint.h from varint-c
    These are to be either copied or symlinked into src/

The reason for this is because they are either fetched or generated from other codebases, as is the case with the protobuf protocol files, or because I saw them as isolated enough in scope and functionality, as well as versatile enough that I might want to use them in other projects, that I made a generic non-rabarber-specific implementation.

Documentation/Understanding the Code

You'll probably want a cursory glance at what the mumble protocol looks like before delving into this. Unfortunately the code is pretty scarcely commented and this is unlikely to change, but there are two graphviz dot files in the doc/ folder which together with the code should help explain the dataflow from alsa to network, and network to alsa respectively. Running make documentation, or just the default target, should turn the dot files into pdfs.

Current Issues

  • It's not finished yet
  • Frequent ALSA underrun error messages, while not ideal, the program works for the intended purpose.
  • Only a minimally viable part of the mumble protocol is so far supported, for connection+tx+rx, e.g. no udp.
  • When receiving long continuous transmissions there is an increasing timeshift that becomes noticeable after >1 min. I do not currently know which part of the transmission chain that is responsible.