voctomix

Voctomix

name:
Voctomix
project-owner:
Mazdermind
git-url:
https://github.com/voc/voctomix
project-status:
running

Voctomix ist ein Softwaremischer der es uns erlaubt auf Konferenzen HD-Recording und -Streaming anbieten zu können. Voctomix wird aktuell von mazdermind fürs VOC entwickelt und ist gst-switch's kleine Schwester. Sie lebt im Git unter git@c3voc.de:voctomix und read-only auf Github.

The installation on VOC machines (encoder cubes and mixer notebooks) is managed through Ansible. The Ansible deployment creates the following things on the encoder-cubes:

  • Source of last release in /opt/voctomix/release
  • Config file for voctocore in /opt/voctomix/voctocore-config.ini
  • Scripts to source/sink/playout video- content into/out of the voctocore in /opt/voctomix/scripts
  • Systemd units in /etc/systemd/system/*.service
  • A system status script in /usr/bin/voctomix-status

Ansible always resets the setup to a default configuration, where all SDI-ports are used as camera INs, no playout is active and all supporting scripts are enabled. Ansible generated the following Systemd units for Voctomix:

Systemd-Unit Description
music-source.service Source Music from /opt/voc/share/pause-music/ into the pause loop
pause-source.service Source pause loop from /opt/voc/share/pause.ts
program-to-[NAME]-playout.service Such a unit is generated for each Decklink output device present at the time of the ansible-run. Playout video- and audio from the program output (=what is being recorded; before the stream-blanker) to the Decklink output
program-to-framebuffer-playout.service Playout video from the program output (=what is being recorded; before the stream-blanker) to the HDMI/DVI/VGA output on the mainboard
recording-sink.service Record the program output to segmented .ts-files in /video
streaming-hd-sink.service Encode and stream the stream-blanker output to
rtmp://127.0.0.1:1935/stream/s{{ room_number }}_native_hd
streaming-sd-sink.service Encode and stream the stream-blanker output to
rtmp://127.0.0.1:1935/stream/s{{ room_number }}_native_sd
stream-to-[NAME]-playout.service Such a unit is generated for each Decklink-Output-Device present at the time of the ansible-run. Playout video and audio from the stream output (=what is being streamed; after the stream-blanker) to the Decklink output
stream-to-framebuffer-playout.service Playout video from the stream output (=what is being streamed; after the stream-blanker) to the HDMI/DVI/VGA output on the mainboard
voctocore.service The Voctocore

All supporting units are WantedBy=voctocore.service as well as the request Requires=voctocore.service and After=voctocore.service. This means:

  • The supporting units only start up when the core has successfully started up first
  • When enabled, the supporting units automatically start up when the core is up
  • The supporting units go down when the core goes down

In practice this means, that to restart the whole chain, you can do:

sudo systemctl restart voctocore.service

Stop recording at night and schedule re-start in the morning:

for i in 1 2 3 4 5 6 41 42; do echo "10.73.${i}.3"; ssh 10.73.${i}.3 'sudo systemctl stop recording-sink.service'; done
for i in 1 2 3 4 5 6 41 42; do echo "10.73.${i}.3"; ssh 10.73.${i}.3 'echo "sudo systemctl restart voctocore.service" | at 09:00'; done

Similar to stop or start everything you can use

sudo systemctl stop voctocore.service

and

sudo systemctl start voctocore.service

To see which units are actually started, you can call

voctomix-status

To Reconfigure a Decklink-Duo Port as Playout, in order for it to survive a restart/reboot, you can do:

sudo systemctl disable decklink-source-decklink-sdi-2.service                                                                                                     sudo systemctl stop decklink-source-decklink-sdi-2.service                                                                                                        

sudo systemctl enable stream-to-decklink-sdi-2-playout.service                                                                                                   
sudo systemctl start stream-to-decklink-sdi-2-playout.service  

If you leave out the enable/disable the change will be temporary and restarting the voctocore unit or the encoder cube will reset the change.

Similar to enable playout to the Framebuffer, you can do:

sudo systemctl start stream-to-framebuffer-playout.service                                                                                                        sudo systemctl enable stream-to-framebuffer-playout.service

Leaving out the enable/disable the change will also here be temporary.

Playout to the framebuffer can always be interrupted by funny tty0 printouts at any time :) Also think about whether you want to playout the program output (before the stream-blanker) or the streaming output (after the stream-blanker).

To see the output of the various units, you can use a command like this:

journalctl -au decklink-source-decklink-sdi-2.service  

Alternative version:

journalctl -a -f -u  grabber-source.service

Some useful control commands when you only have a shell (see also example-scripts/control-server: For more details see https://github.com/voc/voctomix/blob/voctopanel/voctocore/lib/commands.py

echo set_audio cam1 | nc -q0 localhost 9999
echo set_audio cam2 | nc -q0 localhost 9999

echo set_composite_mode fullscreen | nc -q0 localhost 9999
echo set_composite_mode picture_in_picture | nc -q0 localhost 9999
echo set_composite_mode side_by_side_equal | nc -q0 localhost 9999
echo set_composite_mode side_by_side_preview | nc -q0 localhost 9999

echo set_stream_blank pause | nc -q0 localhost 9999
echo set_stream_live | nc -q0 localhost 9999

echo set_video_a cam1 | nc -q0 localhost 9999
echo set_video_a cam2 | nc -q0 localhost 9999
echo set_video_a grabber | nc -q0 localhost 9999

echo set_video_b cam1 | nc -q0 localhost 9999
echo set_video_b cam2 | nc -q0 localhost 9999
echo set_video_b grabber | nc -q0 localhost 9999

here you can see the commands together with replays from the core

=> set_video_a cam1
<= video_status cam1 cam2

=> set_composite_mode fullscreen 
<= composite_mode fullscreen
<= video_status cam1 cam2

=> set_composite_mode side_by_side_equal
<= composite_mode side_by_side_equal
<= video_status cam1 cam2

=> get_config
<= server_config {"stream-blanker": {"sources": "pause,nostream", "enabled": "true"}, "DEFAULT": {}, "side-by-side-equal": {}, "fullscreen": {}, "side-by-side-preview": {}, "mix": {"sources": "cam1,cam2,grabber", "videocaps": "video/x-raw,format=I420,width=1920,height=1080,framerate=25/1,pixel-aspect-ratio=1/1", "audiocaps": "audio/x-raw,format=S16LE,channels=2,layout=interleaved,rate=48000"}, "previews": {"deinterlace": "false", "enabled": "false"}, "output-buffers": {"mix_out": "10000"}, "picture-in-picture": {}}

=> get_video
<= video_status cam1 cam2

=> get_composite_mode
<= composite_mode side_by_side_equal
  • voctomix.txt
  • Last modified: 2020/12/06 22:18
  • by andi