On Fri, Feb 19, 2010 at 01:48:24PM +0100, Fons Adriaensen wrote:
good point. ok. so this still leaves us with the problem of the phase
relation in case of non-power of 2 period sizes.
(i am not sure if this patch slipped into jack1 accidentally or on
purpose, but the alsa driver doesnt enforce power of 2 anymore)
but that kind of thing can be fixed easily.
jack_nframes_t jack_cv_get_phase( jack_client_t *client );
i am basically more in favour of timestamped events.
but your argumentation in this thread is pretty convincing.
and timestamped events come with a galore of other problems,
which i think are harder to tackle in a jack context.
- multiple interpolation schemes
- how to get the current value when port is connected, and didnt see an
so what do people think ?
shall we just add CV port type ?
(this would sort of rule out timestamped control events, since there
should be only one control port type)
not if the period is not a power of 2 :(
but maybe we should just go back to enforcing that again.
thanks for explanation :)
Linux-audio-dev mailing list