9 Sep
2004
9 Sep
'04
6:45 p.m.
On Thu, Sep 09, 2004 at 10:41:34PM +0300, Kai Vehmanen wrote:
Hello,
to avoid too heavy cross-posting, I dropped jackit-devel from the
cc-list...
oops. <blush> sorry about that everyone. :(
On Thu, 9 Sep 2004, Eric Dantan Rzewnicki wrote:
Oh, good. Thanks for clarifying that. I have been wondering why a null
object would have trouble keeping up with making zeros. I'm sorry to
have misunderstood rtnull's purpose. :(
So it would seem that my audio glitches are indeed something in
ecasound, analogueOsc or somewhere in their interaction.
That's good news on the kernel side of things. Thanks Kai!
Lee, I'll still take a look at the stuff you mentioned earlier, just to
be sure. But, it looks like I don't have a kernel problem after all. So
this story turns from a possible trouble report to a positive audio user
experience report.
seeing aren't reported by jackd at all. I
only see them from the
ecasound instance that is hosting the analogueOsc LADSPA plugins and
only when it exits. The rtnull input object reports them when this
ecasound instance exits. Every ecasound chain has to have an input and
an output. Since I just want the audio from the plugin, which in
ecasound's world is an effect attached to a chain between an input and
an output, the rtnull input object serves as a place holder virtual
audio device/file/thing.
Maybe these xruns are red herring? Kai?
Definitely! You should use
"null" instead of "rtnull" if you have any
JACK, ALSA, or other real-time objects in the same setup.
Rtnull, or realtime null, consumes or produces data at real-time speed
(unlike plain "null" that does the same as fast as it can). This is useful
feature in some circumstances, for example when writing audio to a pipe
and from there with netcat to network. Without rtnull, you'd overflow the
network interface (and the receiving end). But if you have both rtnull and jack/alsa devices in
the same setup, you
will get xruns, as rtnull is driven by the system clock while jack/alsa
are driven by the soundcard clock. Sooner or later the clocks will drift
enough and you'll start to get constant xruns.
I was ready to take a look at the ecasound code tonight to see if I
could understand what the null-rtnull difference was ... I'm not sure I
could have figured it out on my own, but now I know. Thanks again.
btw, I'm still really having fun with the 8-oscillator thing. A bit of
it's output came up in one of my automated random mixes yesterday. It's
a sound resource I've been missing in my pool for a while and it's
really good to have that gap filled.
-Eric Rz.