loopback on ladspa-sink: huge, persistent delay introduced by VT switch
Submitted by Roland Pallai
Assigned to pul..@..op.org
Link to original bug (#104408)
Description
I use the following "chain" for VoIP audio processing:
VoIP apps: higher tone
load-module module-ladspa-sink sink_name=spk-voip-eq-out plugin=mbeq_1197 label=mbeq control=-30.0,-19.1,-18.6,-18.6,-18.6,-10.0,-8.0,-6.5,1.5,1.5,1.5,8.5,10.6,10.6,10.6
VoIP apps: joint volume control
load-module module-null-sink sink_name=spk-voip-vol-out sink_properties='device.description="Speaker VoIP Volume Out"' load-module module-loopback source=spk-voip-vol-out.monitor sink=spk-voip-eq-out latency_msec=4
The problem is a huge, persistent, 100% reproducible delay that appears on the loopback after VT switch. The delay could be 5 seconds or even more, followable on volume meters of pavucontrol. No problem if loopback's sink is other loopback or hardware device instead of ladspa-sink OR user is a member of the audio
group thus access is not suspended on VT switch. Need to restart pulseaudio to get rid of the introduced delay. Temporary moving of loopback's sink to other loopback or hardware device eliminates the delay but it returns after restore. Pulseaudio does not log anything that worth to mention.
I have nothing special in my PA config but I can share as is if necessary. Fedora 27, pulseaudio-11.1-7.fc27.x86_64