[LAD] Strange inconsistency in jack frame time

Fons Adriaensen fons at linuxaudio.org
Sun Jun 22 23:51:26 UTC 2014

On Sun, Jun 22, 2014 at 06:16:04PM -0400, Paul Davis wrote:
> On Sun, Jun 22, 2014 at 5:38 PM, Fons Adriaensen <fons at linuxaudio.org>
> wrote:
> >
> >
> > So it seems that the frame time is not a reliable method to
> > determine how many frames have been lost.
> >
> I don't think that anything in the API documentation suggests that it is.
> xruns, whether caused by clients or kernel scheduling, essentially reset
> the clock. nothing in jack1 (and i believe nothing in jack2) attempts to
> maintain clock consistency (monotonicity or linearity) across xruns.

Clear. But then the question remains how Jack computes the number 
of frames that get added to frame_time when it resumes running cycles.

As you say the only thing that can be relied on is the microseconds
time, since its source is external to Jack. So the obvious solution
is to use 

  delta_frames =  delta_time * sample_rate

I don't think you can find anything simpler than that.

The current value is not even the nearest multiple of a period in
some cases (and why should it be a multiple of a period at all).


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)

More information about the Linux-audio-dev mailing list