I get short clicks and plops on the start of recording
where there should be perfect silence and lots of
artefacts after about 80 measures. Timemachine is able
to record a clean take from the same source, an Om kick
Logged In: YES
...so it's not all in my head, I have had suspicions about
Diving right in.
What jack settings do you use?
If you listen to the file muse recorded with another
player does the crackling persist? I'm asking because it
may be a playback issue as the prefetch functionality has
If the file indeed has clicks, could I check a sample of
it? Not sure where that will lead, but you never know..
Logged In: YES
jackd -R -P80 -t10000 -dalsa -dhw:0 -r48000 -p128 -n2
I used MusE, Sweep and mplayer to listen to it.
Sample attached. First the Artefacts on silence, then a cut
from the later part.
BTW, I had yet another kind of recording trouble: a delay in
what had been recorded. Plus not only wrong wavform
graphics, but parts of previous recordings at the start of
new ones. Didn't see anything of this with 0.7.x versions.
Gosh those are audible.... not what I expected at all..
Hmm, how is that possible
About the other issue you mention, when you understand
more about it please submit it as a separate bug-report.
On recording, MusE reports:
Prefetch: NO BUFFER
FIFO 0x89858b8 overrun... 1024
WaveTrack::getData(1045903, 1, 128): fifo overrun
First several of the first, than lots of the other with some
Prefetch messages inbetween. (I looked past this mesagges
initialy, because i'm used to muse reporting something all
the time, even if it just works, sorry.)
Attaching a package with the MusE file and Om patches from
my ongoing project. You need latest Om CVS (currently frozen
and working well). Uncheck Run before loading root.om. The
kick patch has been the most 'reliable' regarding broken
recording (worked with timemachine, though, so I think
neither om nor Jack are broken).
Tapas had similar issues when testing with this package.
I know testing with this is a bit much to ask, and no bad
feelings if you don't ;) (I can often be found in #lad,
Project with Om patches that trigger recording problems
I'm trying ;) Checking in on irc tonight if time permits.
So I found out it's the same with 0.7.1. I know I recorded
stuff from Om with the 0.7 series without problems, before,
Either there is something specific about the signals my new
Om patches generate that bothers MusE, but not Jack itself
and Timemachine (I deem that quite unlikely), or it doesn't
like my current Jack version, 0.100.7 (or any lib that also
plays a role in recording.)
Ok, I was suspecting that.
Did you always run jack with 128 frames?
Could you try running jack with bigger buffers? E.g. 1024
instead of 128.
With 512 the issue is still there, but to a lesser amount,
it seems. Everything above 256 makes for bad midi timing,
1024 ruins it, so while this might help to get at the
problem, there's no workaround here.
Oh, with 1024 I get a very interesting result. Attaching
recording with MusE, 1024 frames, faded out after the totaly
off start (it continous with the right sound and timing,
just with crackling). Plus for comparsion, what I got with
Timemachine (on 128 frames, though).
Kick recorded with MusE, 1024 frames
Kick recorded with Timemachine, 128 frames
A fix in CVS that hopefully takes care of the clicks and
Also note bug: 1469994, which explains a lot of strange
distortion happening when small jack buffers are used.
Workaround: disconnect the audio-group so there's only one
output from the track.
Problem has disappeared (0.8.1a)
Missed to close this one earlier.