You are here

QJackCtl xruns indicator

QJackCtl shows 2 numbers for xruns (as I understand it, I could be completely wrong) like this : N (M) . My dim understanding from seeing an email post of yours, is that the left one, N, is an approximation obtained by searching some Jack log file or something, and the one in (M) is from Jack directly (not sure) and is more accurate. In any case, I observe that M > N, often by a large amount, which fits in with the approximation theory. However, if N is really some approximation, and (M) is the true value, why not remove N from the display? It is confusing, especially to new users, to have both numbers displayed when only the right one matters. On the other hand if N and (M) are measures of 2 different things it would be nice to know what they are. Thanks. QJackCtl is a very useful application, I use it all the time.

rncbc's picture

qjackctl is a pretty old piece of ancient software :) those numbers are just there since ages ago.

and you're correct, N is the number of xrus as detected from scraping the messages log and it only applies to the alsa backend (the most usual perhaps), while M is the number of xruns actually reported through jack api (ie. counts the number of xrun callbacks issue by the jack subsystem).

i guess new users have to live with all these qjackctl idiosyncrasies though.

don'rt worry, be happy :)

Thanks for clarifying the xrun display. Maybe the tooltip could be changed to something like:
XRUN count (L:Alsa R:Jack) or XRUN count: Alsa (Jack). In my early Linux audio days I used to speculate that one number was perhaps errors on buffer reads, and the other writes. I knew that one did not want xruns but I was confused as to which number was more important and why they could be different. Since I use firewire which has ffado as the backend rather that alsa, I suppose that could explain why I sometimes get many more of the M than N. Anyway, its just nice to know what is going on.

Add new comment