[ubuntu-studio-devel] continued as per request, from IRC
Len Ovens
len at ovenwerks.net
Thu Aug 20 20:49:27 UTC 2015
On Thu, 20 Aug 2015, Mike Holstein wrote:
> 15:48 < zequence> holstein: I really urge you to put your thoughts down and write
> an email instead
> 15:48 < holstein> well, its fashionable to not like ubuntu.. and, thats
> something larger than ubuntustudio.. but, when folks go to #ardour, for
> example, and the major piece of advice is "whatever you do,
> dont use ubuntustudio", i would like to think about why
That is not really true, "don't use Ubuntu" Yes I see that... and probably
with good cause. It is possible to get good results with Unity, easy to
get bad results. Certainly Studio sometimes just gets lumped in with
Ubuntu. And when suggesting a distro made for Audio, generally kxstudio or
avlinux are the two mentioned. However, I have heard UbuntuStudio
recommended sometimes as well. (especially lately as kxstudio has had some
issues related to KDE)
Studio has some good stuff:
- xfce
- a good set of applications
- audio and RT allready works
On the other side:
- LTS releases with sometimes the buggiest release of some
required audio utilities.
- LTS releases mean that by the time the next one comes out
the old one is hopelessly behind. Kubuntu may have
the best way of dealing with this by trying to make
each release LTS-able. Anything based on debian, tends
to be release based.
- It is not easy to update an LTS, the policys for adding a new
version for anything besides bugs is not an easy road to
take.
It takes a lot of work to keep an LTS current and we just haven't been
able to do that. Both kx and av add the latest versions to their repos
within days (minutes sometimes)... they can do so because they own the
repos and manage them.
We could set up an upgrade repo ppa, but I do not know if that is what
Ubuntu is all about. Ubuntu flavours are meant to use the Ubuntu repos.
Directions we could go that remain Ubuntu-ish but still make a good distro
for audio:
remove module-udev-detect from pulseaudio and run jackd as the only back
end. So jackdbus would start at session start and pulse would use either
jack or dummy as it's only backends.
Create a udev utility that replaces module-udev-detect for PA with
something that adds a plugged in audio IF to jack on the fly. The user in
-controls would be asked or allowed to determine if the new device became
the jack maser device or if it was added via zita-a2j/j2a. If the
(probably USB) new device was to be master, the internal would then get
added via zita-a2j/j2a.
These two things alone would make Studio unique in the Linux audio world
and would solve more than 50% of support requests both in ubuntuStudio and
in other places like #Ardour.
Make performance mode default with the option when battery operation is
detected to goto a slower speed or ondemand. (in general a slower
_constant_ speed is better for low latency)
Note on performance mode: I have found that performance mode runs cooler
at high CPU use than ondemand. Ondemand is good for mostly idle use.
Allow sw update stuff to be turned off while doing audio intensive stuff
(stop cron works for me).
Any place I have mentioned starting jack should include a2jmidid, using
a2j_control seems to be more reliable for me than using a2jmidid directly.
Note that this whole topic is audio only and does not address other
workflows in Studio. It happens to be what I know :) Also, I have not
mentioned the tweaks we already do for audio which should remain.
--
Len Ovens
www.ovenwerks.net
More information about the ubuntu-studio-devel
mailing list