audiolatency does not work well with encoders/decoders and queues
Submitted by Daniel F
Link to original bug (#796352)
Description
New audiolatency element has some problem if both encoders/decoders and queue are used.
This works as expected:
gst-launch-1.0 audiolatency name=a print-latency=true ! audioconvert ! audioresample ! opusenc ! opusdec ! audioconvert ! audioresample ! a.
After inserting queue between encoder and decoder it stops printing results periodically:
st-launch-1.0 audiolatency name=a print-latency=true ! audioconvert ! audioresample ! opusenc ! queue ! opusdec ! audioconvert ! audioresample ! a.
However when I tried to use queue without encoder/decoder, it started working again:
gst-launch-1.0 audiolatency name=a print-latency=true ! audioconvert ! audioresample ! queue ! audioconvert ! audioresample ! a.
In second case audiolatency most probably does not collect measurements at all. I have more complex pipeline in my app, which sends and receives RTP streams. I tried to use audiolatency there with print-latency=true, and also it does not work. I also periodically print values of last and avg latency, and I always get zeroes there.
Version: 1.14.0