4 Dec
2023
4 Dec
'23
10:57 p.m.
On Mon, Dec 04, 2023 at 01:22:12PM -0800, Len Ovens wrote:
If Jack is syncing from the output device and has an
xrun on the input...
you record that xrun. Depending on where your recording is going, demo to
show the band what it sounds like may not matter but song to distribute you
want zero xruns.
It's some time since I looked at the code, but AFAIK Jack's ALSA backend
will wait for both input and output to be ready to provide/accept a
period of samples.
Then there is a difference between Jack1 and Jack2:
Jack1 will read the input at the start of a period, then call all
clients, and when all are done write to the output.
Jack2 (without -S) will read the input and write the output of the
previous period at the start a period, then run all clients.
With -S it behaves as Jack1.
So there are 4 cases to consider:
Jack1 or Jack2
Input is slower or output is slower.
At this time of day (22:55) I'm not going to torture my brain
to find out what would happen in each of these cases. But it's
certainly an interesting exercise, maybe it will appear in
my dreams tonight.
Anyway if the two devices don't have coherent sample rates,
just use zita-a2j/j2a. Of course you could use the input as
master when recording, and the output when mixing. But in
practice it wouldn't matter at all - there is no loss of
'quality' when the resampling is done the right way.
Ciao,
--
FA