Autoaudiosink ignores the buffer-time and latency-time properties
On both windows and Linux the autoaudiosink ignores the buffer-time and latency-time properties. As reported in the documentation the autoaudiosink is not a subclass of baseaudiosink.
To test you can use the audiolatency module to compare the output of these pipelines:
gst-launch-1.0 pulsesrc device=auto_null.monitor ! audiolatency print-latency=true ! autoaudiosink buffer-time=20000 latency-time=10000
Setting pipeline to PAUSED ...
Pipeline is live and does not need PREROLL ...
Setting pipeline to PLAYING ...
New clock: GstPulseSrcClock
Redistribute latency...
Redistribute latency...
last latency: 135ms, running average: 135ms
last latency: 122ms, running average: 129ms
last latency: 122ms, running average: 126ms
last latency: 122ms, running average: 125ms
last latency: 133ms, running average: 127ms
...gst-launch-1.0 pulsesrc device=auto_null.monitor ! audiolatency print-latency=true ! pulsesink buffer-time=20000 latency-time=10000
Setting pipeline to PAUSED ...
Pipeline is live and does not need PREROLL ...
Setting pipeline to PLAYING ...
New clock: GstPulseSrcClock
Redistribute latency...
Redistribute latency...
last latency: 153ms, running average: 153ms
last latency: 30ms, running average: 92ms
last latency: 27ms, running average: 70ms
last latency: 23ms, running average: 58ms
last latency: 31ms, running average: 53ms
...