Listkeepers - Joern Nettingsmeier - Linux Audio Dev
Mstation: How did you come to be the listkeeper for the Linux Audio Dev mailing list?
Joern Nettingsmeier: well, it sort of happened. :)
sometime in 1999, i got a computer again after dropping out of computing for a few years. i put linux on it and looked around for people doing audio with it. naturally, i came across Dave Phillips' site (sound.condorow.net), and i decided, what the heck, write him an email. to my surprise he responded, with a very cordial letter, and invited me to join the linux-audio-dev list, which at that time was maintained by Alexandre Burton at the university of montreal.
that's how i made first contact with the free software community.
the list was a great place to be (even though i was a total newbie) and i learned a lot of useful tricks from the "tribal elders". over time, it turned out that alexandre was no longer able to spend time on list maintenance, and some people were having problems that required admin help. i was looking for ways to contribute something, and since my coding sucked (and still does), i wrote to alexandre and offered to take over as list admin instead. he gratefully accepted and went on to other things, and i moved the lists over to columbia university, new york, where douglas irving repetto had offered space and bandwidth.
with the help of Paul Winkler, i put a list homepage together and saw to it that the list and its community was made known to a wider public. the addition of a user and announce list followed rather quickly.
In putting a setup together for yourself, did you have a particular starting point - such as a certain type of music you wanted to be able to do or record? Or is it more of a testbed for all kinds of different Linux audio apps?
at the beginning of my studies, i had briefly been into computer music (of the csound and clm type), and i wanted to run the software i knew from the studio (http://icem.folkwang-hochschule.de), where they had sgi boxen running irix. most of that software was also available on linux, and the ia32 hardware was a whole lot more affordable than the mips stuff. later i shifted my attention towards natural sounds and live music, and i used the machine as a 2-track tape with editing. for the last couple of years, i have mostly done mixes and masters of classical chamber music, choirs and small jazz bands, and produced a couple of demos for friends and myself.
more recently, i have begun to use it for teaching acoustics and music production in my theory class in a local music school.
What is the Linux Audio Dev list about?
dunno. it depends on what people post there ;) i'm the guy who empties the spam bucket twice a week, but there my influence ends. i have tried to describe the most frequent topics on the list page to give new subscribers a rough idea (see http://linuxaudiodev.org), but it's already a little out of date. the focus is constantly shifting, as is the community itself. still, there seems to be enough of a consensus about the topics that most subscribers praise the high s/n ratio.
for me, the most important aspects of the list are the development of critical infrastructure such as JACK and LADSPA, and the exchange of ideas between projects. nowadays it is customary for each project to have its own mailing list(s) (especially since sourceforge appeared on the scene), and LAD plays an important role in keeping people in touch with the community as a whole.
and of course the LAD list has been the starting point of many community events (the linuxtag and sounds expo trade show booths, the developer congresses in karlsruhe/germany, the linux audio miniconf in adelaide/australia, and numerous invitations to presentations and tutorials at LUG meetings and other events).
there was an interesting if a little unproductive flamewar recently, which brought up the question what the linux audio community actually is. i found that i don't really know. it was pointed out that the LAD crowd is by no means representative of the entire linux audio community, which is certainly true. it's just that many interesting things in linux audio today bear the LAD tag because that's where they came from.
some people perceive a taboo on discussing fundamental issues of freedom on the list and feel the gist is too pragmatic. others note the absence of many commercial players and feel the environment is harsh to anything not GPL'ed. some fear that the list gets more crowded and noisy, others view the LAD community as a somewhat elusive and "self-selecting" group. all i can say is that it's the people on the list who define this by their contributions. the list is a means, not an end.
judging from those varied views my feeling is that we got it Just Right (tm) :)
Could you tell us about the hardware you have now?
it's still my first computer! i bought an asus p2b-ds more than 5 years ago. it's a dual slot-1 board with integrated u2w scsi. in the beginning i ran it with a single 350 mhz p2, and then i upgraded to dual 600 mhz p3 with half a gig of ram. as to storage, i still have my first 4g scsi disk for the system, a soft-raid-0 array of two 9g sca disks for audio data, and a 180g ide drive to dump the projects when they are finished.
the machine has a cheap ensoniq 1371 sound card i use for analysing signals from my mixing desk, and as an outboard fx loop (did you ever patch freqtweak into an analog mixing desk aux? - great!). a few years back, i bought a soundblaster live platinum with the break-out box. this is my main card. but when i do important stuff, i tend to use some dat recorder's converters and go via the spdif input, because the sblive plain sucks for serious use :(
...What are its drawbacks?
the worst are the card's implicit resampling of all input signals (even when you use the digital ins), and the weird input stages - the recorded material sounds slightly compressed, i'm suspecting there is some kind of automatic gain control going on. ok for pop, but quite a nuisance when you do classical stuff or jazz. and then there's the *very* arcane signal routing. you can do everything, but i still have a hard time understanding how after 4 years of use.
but that's always a problem under linux. if the card's routing is very idiosyncratic, you end up having problems when you have to use generic alsa mixing tools. but it has improved a lot lately, thanks to jaroslav's and takashi's great work!
otoh, the card has very nice features and the break-out box is extremely handy. price/performance ratio was ok when i bought it, but it is simply not a professional card....
i was going to upgrade to a new machine with a hammerfall this spring (i'm dreaming about a dual opteron), but we are moving houses at the moment, so all further investments are on hold ;)
in the analog department, i have an eight track half-inch machine (tascam tsr-8), and a quarter-inch four-track (teac x1000r), which regrettably have fallen into disrepair, but i can't afford new heads at the moment. the center of my setup is an old soundtracs 32-8-2 analog console. i love the eq, and i still do most of my mixing in the analog domain - i need that hands-on feeling. pushing mice around does not quite cut it for me, nor does the sound of many digital mixers. i have a used peavey studiomix controller, but i haven't yet got it to work with ardour.
Ah yes. Aside of anything else, once you get a few tracks going, there's a limit to how many simultaneous mouse click events you can have happening! Does ardour "officially" work with any of the control surfaces?
yes, ardour can bind to arbitrary MIDI controllers, so if your control surface sends them, you are fine. when you ctrl-right-click (iirc) on any widget, ardour waits for the next midi controller message to come in. just wiggle a knob on your midi controller, and presto: it is bound to the widget. very cool.
the only thing to watch out for is that your controller must emit sane, standard control messages. the studiomix produces non-registered parameter numbers, which are multi-byte and not recognized by ardour. it's possible to write a simple mapper for such devices, but i haven't found the time to do it yet, and that part of the MIDI spec is a frightening mess.
... the software you mostly use? Have you tried out any of the audio addons/distros?
i've started with a SuSE distro (5.1 iirc), and i still use one. SuSE have become one of the main supporters of audio under linux (they have more or less the entire ALSA team on their payroll), and their distro has a very cool range of audio packages by default, although i rarely use them (most of my software is built from cvs, which is inevitable when you want to be in close contact with the developer community). but suse keep their libraries very up-to-date, so compiling is very easy - most of the dependencies are already there.
my main application is ardour. i rarely use its editing facilities, but it is the hub of my JACK setup. when i play around, i like freqtweak, zynaddsubfx and hydrogen. steve harris' meterbridge is very important for me when i master (especially the stereo scope and correlation meter), and almost all the stuff i do goes through jamin before it's burned to cd. and i use the swh and lately the tap plugin collections, of course.
i haven't tried any of the specialized audio distros yet, out of laziness. i'm using the apt-get based yellowdog linux on a powerbook (and loving it), and i'm sure demudi and the planet are great.
What kernel do you think is best right now for audio use?
still 2.4 with ll, or the stock suse kernel, which is also very good for multimedia. i've not been able to permanently move to 2.6 at home, because of trouble with the soft-raid i'm using. 2.6 performed ok, but the measurements with latencytest were not quite as good. mostly better on average, but the worst case wasn't, which is pretty much all that counts.
The coming of JACK, a while back, was one of those great leaps forward. How do you use it?
for everything, daily. especially to amaze the guts out of windows users :-D
my machine is old, so there's not too much i can do at the same time, but watching an automated mixdown in ardour through jamin streamed live over the net via ices2/icecast and listening to it with a stream player is about as cool as you can get :)
What do you think of the general state of Linux Audio now?
it has reached critical mass, and i feel it's really going through the roof! the coming of LADSPA and JACK left sliced bread sobbing in a corner, i can do almost everything that i want with linux, and with every new plugin and jack client, the possibilities multiply.
we need to improve some more in usability and lower the entrance barrier, but a lot has been done already, and the amount of good documentation that has sprung up lately is impressive (alsa.opensrc.org, patrick shirkey's site with the howtos and quicktoots, and even some features in printed magazines such as c't, sos and keyboards).
contrary to what most professional windows and mac audio users request, i'm not at all for tighter integration of components and the creation of an all-in-one-place type of application.
for instance, many people dream of sequencing in ardour. i'm sure paul davis can pull it off, but i'm not so sure it's a good idea.
leave this job to charlie steinberg, if he ever gets the linux thing :). monolithic apps are really one of the strengths of the closed, commercial development model.
otoh, JACK has shown what separation of tasks combined with interoperability can achieve in an open-source community. most free monolithic apps are too inert to ever take off, but the piecemeal addition of cool new JACK clients has amounted to the most versatile audio framework on the market IMHO.
i'd rather see interoperability improved further, and have standalone apps for audio and MIDI with an even better common transport control and a coupling of editing features via JACK. and of course we need a good mechanism for session management that's powerful and easy and fun to implement for the project authors.
for me, JACK is the unix way of doing audio.
and if the linux and macosx audio infrastructures continue to converge towards jack, world domination will be a rather boring routine job in a few years' time. ;)
Do you think there's anything missing?
yes, user education.
with linux going mainstream, there will be millions of new users who don't have this open-source acculturation and insight that you get when you've been with a free software community for a while.
we need to welcome them, and also make sure they understand what this is all about, that you get the software for free and can pay back in bug reports, by being nice to people on mailing lists, or even by writing your own code.
we need to show them that the huge, integrated, monolithic behemoths that most people are used to (ms word, logic, etc.) are inherently open-source-unfriendly, and that to request them is a lack of understanding of the driving force of free software.
when people criticize the lack of features of LADSPA, we need to show them that a more luxurious API would not have taken off, leaving them with impressive blueprints but no software. we must educate people not only to cheer about spiffy GUIs that appeal to users, but to also appreciate features that ease the live of programmers, and that if they take on a programmer's attitude, they can profit from those as well.
in my opinion, the first thing a new user of open-source software must learn is how the software came to be in the first place, how the community works, and how to make use of it and contribute. no amount of GUI integration and no installation wizard is ever going to work around this necessity.
One thing that's interesting about user education is that it's not just a Linux thing. In any of the more complex hardware/software setups, you need to spend time on learning. That's always been the case. What seems to be new is the existence of users who think they shouldn't have to make any effort.
and in the linux specific case, there is the home user who has been happily using an illegal copy of logic or cubase on windows, and then for some reason switches to linux.
since the user never paid for software anyway, s/he does not perceive the financial benefit of free software. and not seeing this "gift" also keeps this user unaware of the potential and the social benefits of the software. all these are motivating factors which help the user over the initial steep part of the learning curve, and the radically different style of work that linux requires.
so what s/he will do is to complain about inferior user interfaces, whine a little on some mailing list and switch back eventually.
With Linux audio it used to be true that the docs were so sparse that you had to figure things out by experimentation but those days are pretty much past now.
yeah, and that's good news. initially, i was afraid that with a lower entry barrier, a lot of not-so-nice, lazy folks with a demanding attitude would swamp the culture, those who used to be kept out by frustration before :) but this has turned out to be rare.
on the contrary, it seems that community support via forums, mailing lists and contributed documentation scales excellently. s/n on audio lists is still very good, most questions are answered or at least discussed, and we have reached a critical mass of knowledgeable users who can help out others with basic stuff. this frees the programmers to only answer the harder questions and otherwise get on with their work.
Bookmark:post to Delicious Digg Reddit Facebook StumbleUpon
Recent on Mstation: music: Vivian Girls, America's Cup, music: Too Young to Fall..., music: Pains of Being Pure At Heart, Berlin Lakes, music: Atarah Valentine, Travel - Copenhagen, House in the Desert