Audio not working - gstreamer-audio_1 errors

[guythnick]

I have setup your project on a raspberry pi via the docker install method. Video seems to be working fine, but I have no audio. Using a cheap USB mic that shows as card1 in arecord -l.

When I run
docker-compose logs | grep gstreamer-audio

I get repeats of this output:

gstreamer-audio_1  | Setting pipeline to NULL ...
gstreamer-audio_1  | Freeing pipeline ...
gstreamer-audio_1  | Additional debug info:
gstreamer-audio_1  | pulsesrc.c(1015): gst_pulsesrc_open (): /GstPipeline:pipeline0/GstPulseSrc:pulsesrc0
gstreamer-audio_1  | mkdir: cannot create directory ‘/var/run/dbus’: File exists
gstreamer-audio_1  | dbus-daemon[16]: Failed to start message bus: The pid file "/run/dbus/pid" exists, if the message bus is not running, remove this file
gstreamer-audio_1  | E: [pulseaudio] main.c: Daemon startup failed.
gstreamer-audio_1  | Connection failure: Connection refused
gstreamer-audio_1  | pa_context_connect() failed: Connection refused
gstreamer-audio_1  | Connection failure: Connection refused
gstreamer-audio_1  | pa_context_connect() failed: Connection refused
gstreamer-audio_1  | Setting pipeline to PAUSED ...
gstreamer-audio_1  | ERROR: Pipeline doesn't want to pause.

[alogoc]

May I ask which mic you are using? Having the exact same problem, I used same as the authors mic, which is USB and used an adapter to convert it to mini usb to plug it in as I am running this on pi zero. Ordered another mic now..I'll post the results once I get to try it.

[guythnick]

I actually was using the same cheap microphone originally recommended. Today I received a AmazonBasics Portable USB Condenser Microphone, and that is working with Pulse as-is. So, I think it is an issue with those cheap microphones.

[kotarokun]

I facing the same problem. Is it an problem of Docker? or of microphone?

[J4nsen]

The audio device (if I remember correctly hw1) is hardcoded in the docker containers. If a HDMI cable was plugged in, my cheap usb mic got the identifier hw2, and since HDMI does not provide audio-input, the container fails to start. So please make sure that hw1 actually is the microphone.

[nine86]

So at first I thought my audio issues were related to the device being hw2 vs hw1 but as suggested by @J4nsen once the HDMI cable is removed that is no longer a problem. However I get the following errors when I run:
docker-compose logs | grep gstreamer-audio

I get:

gstreamer-audio_1  | Command not found: /audio_entry.sh
gstreamer-audio_1  | Command not found: /audio_entry.sh
gstreamer-audio_1  | Command not found: /audio_entry.sh
gstreamer-audio_1  | Command not found: /audio_entry.sh

When I run: docker-compose ps
I get:

            Name                          Command                 State      Ports
----------------------------------------------------------------------------------
fruitnanny_fruitnanny_1        /usr/bin/entry.sh node ser ...   Up                
fruitnanny_gstreamer-audio_1   /usr/bin/entry.sh /audio_e ...   Restarting        
fruitnanny_gstreamer-video_1   /usr/bin/entry.sh gst-laun ...   Up                
fruitnanny_janus_1             /usr/bin/entry.sh /bin/sh  ...   Up                
fruitnanny_nginx_1             /docker-entrypoint.sh ngin ...   Up                

Any help would be much appreciated

[J4nsen]

Did you change anything in the Dockefile related to the audio container? Is the image still ivadim/fruitnanny-gstreamer?

[nine86]

Did you change anything in the Dockefile related to the audio container? Is the image still ivadim/fruitnanny-gstreamer?

Do you mean the Dockerfile in opt/fruitnanny/docker/gstreamer? That was unedited. However I did make a change to /opt/fruitnanny docker-compose.yml. I changed ivadim/fruitnanny-gstreamer container to klemmie117/fruitnanny-gstreamer-fix as that was suggested to fix the video streaming issues.

[J4nsen]

You only have to change the image to klemmie117/fruitnanny-gstreamer-fix for the video container. The audio container can be left as is.

[nine86]

You only have to change the image to klemmie117/fruitnanny-gstreamer-fix for the video container. The audio container can be left as is.

Thanks I'm going to try that now and update here. I figured as much when you asked if I made a change to the container.

[nine86]

@J4nsen Doesn't seem to work. The error log is as below:

gstreamer-audio_1  | I: [pulseaudio] main.c: Daemon startup successful.
gstreamer-audio_1  | Failure: Module initialization failed
gstreamer-audio_1  | Failure: Module initialization failed
gstreamer-audio_1  | Setting pipeline to PAUSED ...
gstreamer-audio_1  | Pipeline is live and does not need PREROLL ...
gstreamer-audio_1  | Setting pipeline to PLAYING ...
gstreamer-audio_1  | New clock: GstPulseSrcClock
gstreamer-audio_1  | ERROR: from element /GstPipeline:pipeline0/GstPulseSrc:pulsesrc0: Failed to connect stream: No such entity
gstreamer-audio_1  | Additional debug info:
gstreamer-audio_1  | pulsesrc.c(1622): gst_pulsesrc_prepare (): /GstPipeline:pipeline0/GstPulseSrc:pulsesrc0
gstreamer-audio_1  | Execution ended after 0:00:00.003790619
gstreamer-audio_1  | Setting pipeline to PAUSED ...
gstreamer-audio_1  | Setting pipeline to READY ...
gstreamer-audio_1  | Setting pipeline to NULL ...
gstreamer-audio_1  | Freeing pipeline ...
gstreamer-audio_1  | mkdir: cannot create directory ‘/var/run/dbus’: File exists
gstreamer-audio_1  | dbus-daemon[18]: Failed to start message bus: The pid file "/run/dbus/pid" exists, if the message bus is not running, remove this file
gstreamer-audio_1  | Process 22 died: No such process; trying to remove PID file. (/run/avahi-daemon//pid)
gstreamer-audio_1  | E: [pulseaudio] main.c: Daemon startup failed.
gstreamer-audio_1  | Connection failure: Connection refused
gstreamer-audio_1  | pa_context_connect() failed: Connection refused
gstreamer-audio_1  | Connection failure: Connection refused
gstreamer-audio_1  | pa_context_connect() failed: Connection refused
gstreamer-audio_1  | Setting pipeline to PAUSED ...
gstreamer-audio_1  | ERROR: Pipeline doesn't want to pause.
gstreamer-audio_1  | ERROR: from element /GstPipeline:pipeline0/GstPulseSrc:pulsesrc0: Failed to connect: Connection refused
gstreamer-audio_1  | Additional debug info:
gstreamer-audio_1  | pulsesrc.c(1015): gst_pulsesrc_open (): /GstPipeline:pipeline0/GstPulseSrc:pulsesrc0
gstreamer-audio_1  | Setting pipeline to NULL ...
gstreamer-audio_1  | Freeing pipeline ...

When I run: docker-compose ps everything was fine for the first 2 mins then after that I get the restarting status.
I get:

            Name                          Command                 State      Ports
----------------------------------------------------------------------------------
fruitnanny_fruitnanny_1        /usr/bin/entry.sh node ser ...   Up                
fruitnanny_gstreamer-audio_1   /usr/bin/entry.sh /audio_e ...   Restarting        
fruitnanny_gstreamer-video_1   /usr/bin/entry.sh gst-laun ...   Up                
fruitnanny_janus_1             /usr/bin/entry.sh /bin/sh  ...   Up                
fruitnanny_nginx_1             /docker-entrypoint.sh ngin ...   Up  

[J4nsen]

Please make sure that hw:1 is actually your microphone. You can do this on the docker-host, e.g., by running alsamixer or aplay -l.

[nine86]

I run arecord -l and I get the following:

**** List of CAPTURE Hardware Devices ****
card 1: Device [USB PnP Sound Device], device 0: USB Audio [USB Audio]
  Subdevices: 1/1
  Subdevice #0: subdevice #0

[J4nsen]

This looks ok. Sorry, I dont have my fruitnanny here anymore, so I can only give some advice.

Next step would be running the container on its own and execute the steps in audio_entry.sh and audio_stream.sh (https://github.com/ivadim/fruitnanny/tree/master/docker/gstreamer) manually.

[nine86]

This looks ok. Sorry, I dont have my fruitnanny here anymore, so I can only give some advice.

Next step would be running the container on its own and execute the steps in audio_entry.sh and audio_stream.sh (https://github.com/ivadim/fruitnanny/tree/master/docker/gstreamer) manually.

Hey I really appreciate any advice at all. Thank you so very much!!! Trying to complete this project in time for when my baby is born. Haha. It's been a slog since I'm so unfamiliar with the software. Sorry to ask, but any chance you could elaborate on the steps to do the above?

[J4nsen]

Make sure all other fruitnanny containers are stopped.

Execute $ docker images, use the ID to run a bash in the container with $ docker run -i -t idofcontainer /bin/bash.

In the container run:

$ chgrp audio -- /dev/snd/*
$ mkdir /var/run/dbus
$ dbus-daemon --system --fork
$ avahi-daemon -D

$ pulseaudio -D -v --log-target=stderr
$ pactl load-module module-alsa-source device=hw:1
$ pactl load-module module-echo-cancel source_name=microphone aec_method=webrtc 'aec_args="analog_gain_control=0 digital_gain_control=0"'
$ gst-launch-1.0 -v pulsesrc device=microphone ! audioconvert ! audioresample ! opusenc ! rtpopuspay ! queue max-size-bytes=0 max-size-buffers=0 ! udpsink host=127.0.0.1 port=5002

Check after every step if it worked. I suppose it should fail around pulseaudio/pactl. You could then run pulseaudio in the foreground with $ pulseaudio -vvv & and continue again with the rest of the commands.