#mmb - fork This is a fork from the older original to make it more useful to my purposes.
##Initial commits:
- Get rid of pyMedia, it's outdated as all get out.
- Switch input to a FIFO buffer (feed from mpg123)
- Add arguments/commandline-ability
##Building
Looks like to get this going you must do two things:
###celt
Go in to the celt folder and run
python ./setup.py build
and copy the build *.so file to the main directory (you have to dig down in to the build folder).
###libcryptstate
Go in to the libcryptstate folder and run
make (simple, duh)
and copy the *.so file to the main directory.
From there it should all work.
###Other Dependencies
There are other dependencies required but I don't remember all of them. One is probably libcelt which can be had here:
git clone git://git.xiph.org/celt.git or git clone http://git.xiph.org/celt.git
Others will show up in stack traces as errors (protobuf) and you'll just need to add those. I'll do a proper writeup sometime soon.
- cython
- libprotobuf-dev and python-protobuf
- libcelt
##Running
python ./main.py
There's help in there so read that, it'll be kept more up to date. For a sample line:
mpg123 -r 48000 -s -m "./somefile.mp3" | python ./main.py -s 127.0.0.1 -u username -p password
###Piping and mpg123
mpg123 has just the right features to get us the PCM audio stream we need. It can also handle http streams:
mpg123 -r 48000 -s -m "http://relay.radioreference.com:80/138151917"
That will play the Boston Police radio feed from RadioReference.
- We do need
-r 48000because the celt encoder in our python script demands it (I've tried changing it, doesn't work). - We do need
-sto get the stdout output. - We might not need
-m. That mixes us down to mono instead of stereo. Try without, if the audio is "slow" use it.
##TODO
I want to get threshold detection going on the audio stream so we're not constantly broadcasting (useful in the case of voice streams). DONE.
##NOTES The threshold feature is hard coded to ">1000" which seems to work well enough. Tweak as you wish.