One of the greatest features of GNU/Linux is that it shares the Unix amaziness of having lots of small tools which are easily combined together via stdin/stdout. Today we are going to explore the possibilities of playing MP3 files with these little tools.
The simplest command to produce sounds is aplay. Prior to ALSA there was a device called /dev/dsp to which you could redirect any file and it would “play” it. Now aplay acts similarly. Try executing this:
cat /path/to/wav/file | aplay
You can also send to aplay something that is formally not music at all. Consider an example:
cat /bin/bash | aplay # Sounds a bit like dubstep at the beginning!
As you can see aplay will play any binary stream you pass it.
This tools is a free (and less buggy) version of mpg123 which is a dead simple MP3 player. First of all you can provide it with a list of files to play:
mpg321 one.mp3 two.mp3
You can also supply the call with a playlist (which is a simple list of files to play):
find *.mp3 > 1.pls # Will scan the current dir for MP3 files and write their names to 1.pls mpg321 -@ 1.pls
mpg321 also supports playing from the stdin and writing output to the stdout.
Right now you have a working MP3 player that you can pause (by pressing Ctrl-Z in the terminal. To bring it back type fg %1), play next track (by pressing Ctrl-C once. Pressing Ctrl-C twice in a rapid succession will stop the playback).
We hadn’t seen anything cool so far. OK, a poor man’s “player” controllable from shell. Can you do with it anything at all? Well, perhaps there are things you can. For example, streaming the songs directly to another computer via ssh!
cat /path/to/mp3 | mpg321 -w - - | ssh firstname.lastname@example.org aplay
-w - option to mpg321 means that it would output the wav file to the stdout. The second - tells it that it should take output from stdin. The output from mpg321 is then passed via ssh to the aplay which would play the stream on the remote machine.
Nothing stops you from translating a series of tracks. Actually MP3 files are excellently concatenable with mere cat:
cat *.mp3 | mpg321 -w - - | ssh email@example.com aplay
This kind of “translation” has its weakness – you can’t stop it from the target machine unless putting down sshd or sound system (Did I say weakness?:)). To achieve some better feeling of an Internet radio you can use netcat:
cat *.mp3 | mpg321 -w - - | nc -l -p 9999
From the other machine just type the following:
nc first.machine.ip.address 9999 | aplay
Voila! You have an Internet music radio going directly to your audio device. How more awesome can this be? Well, there are certain limitations of the approach: it doesn’t support multiple connections and doesn’t restart listening the port after the client disconnects. These issues are fixable but are probably close to rewriting Icecast:).
Here’s the command to make our Radio “Netcat” a bit more spicy:
find -name "*.mp3" | sort -R | sed -E 's/(.+)/"\1"/' | xargs cat | mpg321 -w - - | nc -L -p 9999
If you execute it from the root of your music folder it will find all MP3 files, shuffle them, wrap names into quotes (this is done to fight a xargs issue) and stream it to our 9999 port.
The final neat feature
If you can redirect anything to aplay why can’t you send it… your own voice! It is as simple as this:
arecord | ssh firstname.lastname@example.org aplay
arecord will record the input from the microphone and send it via pipe directly to the sound output of the target computer. Execute the same command from the second computer (with first computer’s IP address respectively) and you’ll have a super-secured overly-hipster voice connection (still with a couple of seconds lag:)). Who needs that crappy Skype anyway?
I covered a little bit of examples of what you can do provided a certain knowledge of Linux tools and a certain amount of stupidity. Hope this proves useful to somebody and remember – I was using this before it was cool!