On Tue, Feb 02, 2016 at 12:17:08AM +1100, Patrick Shirkey wrote:
Maybe it is the difference in the way we are using
JACK over here.
Particularly with netjack but also with a variety of older systems. In
some cases we don't have pulseaudio or D-bus running. in total we have
more consistent results with JACK1 for our use case.
I'm not saying that JACK2 cannot be made to perform as well as JACK1 in
the same circumstances and it could also be that proper A/B testing
reveals a user bias instead of real world performance differences.
However we generally do not have the time or inclination to do such tests
as it usually takes a considerable amount of effort to get a fully working
system and trying to replicate the results with JACK2 is not our main
motivation.
The point is that JACK1 is still very useful so unless we are going to
spend a lot of time/effort to track down the differences between JACK1 and
JACK2 it is useful to keep JACK1 alive for specific use cases.
Assuming that we have a few people who are willing to keep contributing to
the JACK1 codebase then it shouldn't be a problem for JACK1 to track JACK2
or vice versa.
However if no one wants to continue maintaining JACK1 then we need to
spend some time to find and isolate any differences in performance between
the two.
I agree that JACK1 is still very useful. I'm mainly using JACK2 at the moment
because I need to run multiple clients with the load spread over several CPU
cores. I used JACK1 for many years before my present use case, and I'd use it
again on a simpler setup.
As a user it's a bit frustrating that there are other differences between
JACK1 and 2 apart from UP/SMP and dbus capability. There's often some
other trade off to consider when deciding which one to use.
John