Redirect jukebox output to fifo pipe (snapcast)

[n0rdik]

Hi!! I want to use subsonic with snapcast multi-room streamer (https://github.com/badaix/snapcast). Snapcast uses a named pipe allocated in "/tmp/snapfifo" which uses to distribute the audio over the network to the clients. I want to pipe the output of the jukebox to this pipe. I've modified the server, adding the line '-Djavax.sound.sampled.SourceDataLine=com.sun.media.sound.DirectAudioDeviceProvider' \ becouse is the only way for me to make jukebox mode working (ubuntu 16.10 linux machine with pulseaudio). I think there may be a way to use the pipe as a SourceDataLine, but I can't find how. Can someone help me? Thanks in advance.

[trumanpacody]

You can edit ~/.asoundrc for subsonic user to reflect the following:

Code: Select all

pcm.multi {
    type route;
    slave.pcm {
        type multi;
        hint.description "MULTI OUT";
        slaves.a.pcm "output";
        slaves.b.pcm "rate48000Hz";
        slaves.a.channels 2;
        slaves.b.channels 2;
        bindings.0.slave a;
        bindings.0.channel 0;
        bindings.1.slave a;
        bindings.1.channel 1;
        bindings.2.slave b;
        bindings.2.channel 0;
        bindings.3.slave b;
        bindings.3.channel 1;
    }

    ttable.0.0 1;
    ttable.1.1 1;
    ttable.0.2 1;
    ttable.1.3 1;
}

pcm.!default {
        type plug
        slave.pcm "multi"
}

pcm.output {
        type hw
        card 0
}

pcm.rate48000Hz {
        type rate
        slave {
                pcm writeFile # Direct to the plugin which will write to a file
                format S16_LE
                rate 48000
        }
}

pcm.writeFile {
        type file
        slave.pcm null
        file "/tmp/snapfifo" # can test with /tmp/testsound.wav
        format "raw" #test with "wav"
}

Then direct snapserver to that fifopipe in the server settings. This works but you won't be able to use the same card to output the snapclient stream locally. The local soundcard will output the non-delayed stream.

It would be ideal if we could bypass alsa entirely and output directly to a fifopipe. I tried using transcoding settings to send audio to fifo w/ ffmpeg instead of stdout, but this approach didn't seem to work.

Is there anyone that knows java that might be able to chime in here? Given that subsonic has gone closed source and there seems to be little to no activity for feature reqs / help I'd suggest also posting at madsonic or airsonic.

Edit: this may have been you:

https://stackoverflow.com/questions/45753484/create-java-method-for-javax-sound-sampled-sourcedataline-to-map-it-with-a-unix

Edit2: There's also this but I don't know how exactly to apply it to this scenario:

https://cdimascio.wordpress.com/2014/01 ... with-java/

[n0rdik]

Yeah thanks, it works. But I don't know why it worked now, because I've tried this solution before and gave no chance to work. Maybe the last system update. Btw, I implemented the general "ALSA" solutión, just creating the default plug as the snapcast web says:

Code: Select all

pcm.!default {
   type plug
   slave.pcm rate48000Hz
}

pcm.rate48000Hz {
   type rate
   slave {
      pcm writeFile # Direct to the plugin which will write to a file
      format S16_LE
      rate 48000
   }
}

pcm.writeFile {
   type file
   slave.pcm null
   file "/tmp/snapfifo"
   format "raw"
}


To achieve this only with subsonic, I've created the user subsonic (previously I used my own user for launching the server) and made "/var/subsonic" as the home for this user, then created there de .asoundrc file with that content.
But I think this solution is not so flexible and complicates a little: Subsonic -> ALSA -> fifo pipe -> ALSA. My question is still there, just for general culture, if there is a method (maybe programming it in java) to connect a java audio app that writes to a SourceDataLine directly to a named pipe, resulting in Subsonic -> fifo pipe -> whatever. Maybe inside subsonic code, like with SONOS, but now is closed. I think I will make a request for this. Sorry for my english and thanks for all.