gst-plugins-good: Revert "Pulsesink: Allow chunks up to bufsize instead of segsize"
pl bossart
bossart.nospam at gmail.com
Fri Apr 8 13:53:59 PDT 2011
> Hi,
> can you elaborate a bit on this? Strictly speaking, it's not true that the
> property is ignored without the revert, since segsize/latency-time is still
> passed to PA's minreq attribute. Since that is the minimum amount of free
> space that must be available before pulse calls us for more data, that
> should still have a huge effect (if not, there is something else wrong in
> pulsesink).
> I haven't measured the wakeup behavior and power consumption of pulsesink in
> a while, so I'm wondering if you have some up-to-date empirical data that
> prompted you to this revert :)
Agree with Rene. The notion of 'latency-time' doesn't make sense here
anyway, this parameter only corresponds to the buffer size send to
pulseaudio and has nothing to do with latency. With the latency
adjustment flag there is no direct relationship between latency-time
and buffer-time.
-Pierre
More information about the gstreamer-devel
mailing list