2 Sep
2015
2 Sep
'15
9:05 p.m.
I've run across a phenomena that puzzles me.
It appears that each time I start the jack server on an usb card,
jack_iodelay will report a different "extra loopback latency". I've
seen this phenomena on a behringer x32, and just verified it on a rme
babyface.
For instance with jack1 at 64/2 on the babyface, I come up with values
of 232,233,234, and 266. With jack2 I saw 231, 233, 248, 318, 335. A
run with jack1 at 2048/3 came up with 2589, 2590, 2633, 2634, and 2667.
I also tested with jack_delay, which also comes up with different
values when I restart the server. Note that on consequent invocations
without restarting the server the values stay the same for the 2
utilities.
Maybe not a huge problem and I suppose I'll set the latency
compensation somewhere in the middle of the results, but still I'm
curious of the mechanism behind this phenomena.
FWIW, I don't think I've ever seen this on my pci cards, but I'll run
some tests when I get to the hardware and find some extra time.
--
Joakim
2 Sep
2 Sep
9:17 p.m.
On 02/09/15 21:04, Joakim Hernberg wrote:
Maybe not a huge problem and I suppose I'll set
the latency
compensation somewhere in the middle of the results, but still I'm
curious of the mechanism behind this phenomena.
USB devices add some extra buffering, check out this thread:
http://jack-audio.10948.n7.nabble.com/Differences-in-latency-between-USB-an…
Regards,
L
9:35 p.m.
On Wed, 2 Sep 2015 21:17:43 +0200
Luis Garrido <luis(a)vagar.org> wrote:
On 02/09/15 21:04, Joakim Hernberg wrote:
Well yes, I'm aware of that and it was this extra latency that I wanted
to ascertain. What puzzles me is that each time the jack server is
invoked, it will need a different amount of latency compensation even
when no other parameters have changed.
I'll read that thread and it's links again, but IIRC they don't hold
the answer to my query.
--
Joakim
Maybe not a huge problem and I suppose I'll
set the latency
compensation somewhere in the middle of the results, but still I'm
curious of the mechanism behind this phenomena.
USB devices add some extra buffering, check out this thread:
http://jack-audio.10948.n7.nabble.com/Differences-in-latency-between-USB-an…
9:52 p.m.
On 02/09/15 21:35, Joakim Hernberg wrote:
I'll read that thread and it's links again,
but IIRC they don't hold
the answer to my query.
Ah, ok, sorry, I didn't pay enough attention while reading your post and
got tripped by the USB vs PCI comment at the end.
Please disregard, as you say that thread is not really relevant to your
question.
Regards,
L
11:51 p.m.
On 09/02/15 21:35, Joakim Hernberg wrote:
Well yes, I'm aware of that and it was this extra
latency that I
wanted to ascertain. What puzzles me is that each time the jack server
is invoked, it will need a different amount of latency compensation
even when no other parameters have changed.
Didn't Fons mention a similar problem lately? (think two weeks ago).
He's currently travelling, but maybe you want to check out his patches
(yet to be posted).
Note that Fons' patches are only about client execution order in jackd1.
If you also get different values in jackd2, then it's either unrelated
or needs fixing as well.
Cheers
3 Sep
3 Sep
11:42 p.m.
On Wed, 2 Sep 2015 23:51:36 +0200
Adrian Knoth <adi(a)drcomp.erfurt.thur.de> wrote:
On 09/02/15 21:35, Joakim Hernberg wrote:
IIRC the problem Fons uncovered had to do with the effective latency
changing by one period when clients connected/disconnected. This
doesn't appear to be related. And AFAIK it only happens with USB
devices.
--
Joakim
Well yes, I'm aware of that and it was this
extra latency that I
wanted to ascertain. What puzzles me is that each time the jack
server is invoked, it will need a different amount of latency
compensation even when no other parameters have changed.
Didn't Fons mention a similar problem lately? (think two weeks ago).
He's currently travelling, but maybe you want to check out his patches
(yet to be posted).
Note that Fons' patches are only about client execution order in
jackd1. If you also get different values in jackd2, then it's either
unrelated or needs fixing as well. 
17 Sep
17 Sep
9:19 p.m.
On Wed, Sep 02, 2015 at 11:51:36PM +0200, Adrian Knoth wrote:
Note that Fons' patches are only about client
execution order in jackd1.
If you also get different values in jackd2, then it's either unrelated
or needs fixing as well.
The OP's problem seems unrelated to the issue I reported.
Ciao,
--
FA
A world of exhaustive, reliable metadata would be an utopia.
It's also a pipe-dream, founded on self-delusion, nerd hubris
and hysterically inflated market opportunities. (Cory Doctorow)
3167
days inactive
3182
days old
linux-audio-dev@lists.linuxaudio.org
6 comments
4 participants
participants (4)
-
Adrian Knoth -
Fons Adriaensen -
Joakim Hernberg -
Luis Garrido