[linux-audio-dev] Re: What Parts of Linux Audio Simply Work Great?

[Christoph Eckert]

> Now given the two following options:
>
> A. Extend Jack to accomodate all the diverse needs of the
>    desktop developers, and request them to use it for all
>    applications,

this woud IMHO be the best solution, but it will not happen or 
at least last very long until

* all distros will have JACK running per default
* all developers of any audio program have rewritten their 
code.

> B. Provide a layer on top of both Jack and ALSA (as for
>    example the jackified artsd), and recommend that for
>    desktop apps,

This is a solution we already can provide. The problem is not 
that an active audio user cannot learn that he needs to kill 
the desktop server, to start JACK and then to optionally 
restart the desktop sondserver on top of JACK.

The problem is that so many apps use so many different ways to 
output audio: arts, esd and gstreamer, ALSA or OSS direct 
access.

> then for me it's clear that I would prefer B. It would
> provide a solution for at least three types of users:
>
> - desktop audio only: run the server on top of ALSA and
>   don't bother with Jack.
>
> - audio production workstation (no 'desktop
> entertainment'): just run Jack as we already do.
>
> - those in between: run the server on top of Jack.

I agree. But what still is needed is a common desktop sound 
server which will be part of each distro. arts is dead, esd 
will also be dropped and be replaced by anything like 
polypaudio or gstreamer.

What's missing is a commonly accepted soundserver which isn't 
bound to a particular desktop, and each audio developer 
should use this (still missing) common soundserver, 
regardless if you're working shell only (as julien does), on 
FVWM, IceWM or anything else.

So what's missing in my opinion is that the community needs to 
agree about one common soundserver for all desktops and then 
any audio application should use it.

If the community will not choose JACK then another one needs 
to be elected or created. Any audio developer should use this 
soundserver. Direct access to the card should become a nono 
on multitasking multiuser computers.

We as professionals simply could ignore this because most of 
us kill all soundservers, fire up JACK and we're happy, even 
using xmms to play some oggs via JACK.

OTOH I'd really like to see that the audio experience of a 
desktop user improves.

In my opinion it is not important if this default soundserver 
will be JACK, another one running on top of JACK or directly 
on top of ALSA or OSS. It is only important that the 
community agrees about one solution and starts using it.

What have we to do with this?

Less. Except for the fact that we have enough knowledge to 
design an optimal solution. What I want to achieve is not 
that we're working on this issue by writing the code. None or 
less of us are interested in such things. But we could create 
a recommendation.


Thanks & best regards


    ce