8 Apr
2006
8 Apr
'06
3:22 p.m.
Lee Revell wrote:
On Fri, 2006-04-07 at 16:58 -0400, Bill Allen wrote:
I found that if I increase the number of periods to 3 that the xruns on
startup don't happen. Thus, using:
/usr/bin/jackd -R -dalsa -dhw:0 -r44100 -p512 -n3
I was able to bring up qsynth and attach vkeybd and play some notes no
problem. But, when I launched jack rack and loaded a rack with a couple
of the CAPS amp/cabinet simulation plugins, all hell broke loose. Even
with a buffersize of 2048 and 4 periods (-p2048 -n4), loading the rack
caused way too many hard xruns.
On 2.6.15-18-686, I'm running jack with:
/usr/bin/jackd -R -dalsa -dhw:0 -r44100 -p512 -n2
with no problems at all.
I'm trying to think of what could have regressed in the newer kernel
builds that would explain this behaviour?
Bill
I'll try to do as you suggest. As I
understand it (which could be not
at all) Dapper doesn't support realtime-lsm in any of those kernel
builds. It uses Rlimits-Aware PAM. I've added myself to the audio
group and as I mentioned, jack works great in 2.6.15-18, but in *-19,
and *-20 it's miserable.
OK that's a kernel bug of some kind - can you please report it to
Ubuntu? There should not be too many changes between those point
releases.
Lee