On Mon, 2019-12-02 at 08:49 -0700, Angus Ainslie wrote: > > So this was a weird one. The modem and codec are fixed at a 48K sample > rate ( the codec can do other rates but the clocks are currently set for > 48k). On the devkit we added an alternate rate of 44.1K to deal with > call audio. Pulse audio was using the alternate rate instead of the > default rate and it was setting up the codec for 44.1K and one of the > rates got set to a sample rate of 47999 Hz > > https://source.puri.sm/Librem5/calls/uploads/c6430c76b6a79aae827291e55305d3ec/after_call.sink-inputs > > In one of the tests I also saw a rate of 43KHz but I don't seem to have > captured any debug for that one. > > Dropping the alternate rate caused pulse audio to use 48K which allowed > the sound to flow through the loopbacks and the latency disappeared at > the same time. > > https://source.puri.sm/angus.ainslie/librem5-base/commit/450976adfacf25da79b71b6dd24a5f8643da0873#dd3464357c7325c6a3167bc6243c1a62926ba401 > > Thanks for your help. Great that you got it behaving! Regarding the weird rates: it's normal that the loopback sink input has a rate that is slightly different from the normal rate. It's part of the logic that adapts to clock drift between the source and sink. If the rate changes a lot, however, that can be an indication that the hardware isn't working properly (it's producing or consuming audio too fast or too slow). -- Tanu https://www.patreon.com/tanuk https://liberapay.com/tanuk _______________________________________________ pulseaudio-discuss mailing list pulseaudio-discuss@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/pulseaudio-discuss
