[LAU] Estimating JACK latency

[Robin Gareus]

On 05/03/2012 10:43 AM, Joel Roth wrote:
> On Thu, May 03, 2012 at 10:21:16AM +0200, Robin Gareus wrote:
>> Capture latency in above message is a wrong. It was reported a while
>> back: check jack-devel archive Dec 2010:
> 
> Thanks
> 
>> with jack2, there's also the difference of sync (jackd --sync .. -d alsa
>> ...) vs. async mode. The latter adds 1 period of output latency.
> 
> I'll need to read about this. --sync is listed in --help,
> but not in the man page.

http://www.grame.fr/~letz/jackdmp.html

> I hope the difference shows up in get_jack_port_latency_range()!

It does. The output latency is one period longer in async mode.

>> see also:
>>   http://wiki.linuxaudio.org/wiki/jack_latency_tests
>  
> And again. This is very helpful.
> 

This wiki-page is a bit outdated (it's from before the new latency API
was introduced) but item (2) under
http://wiki.linuxaudio.org/wiki/jack_latency_tests#interpretation_and_analysis
describes the same bug that you stumbled over.

(1) was actually solved: The additional latency for USB devices comes
from ALSA's USB driver implementation. Clemens explained the details in
the same thread on jack-devel "Re: Differences in latency between USB
and internal audio interfaces"  December 9, 2010.

ciao,
robin