Vol. 28 Issue 4 Reviews
Linux Audio Conference 2004

Zentrum fur Kunst und Medientechnology, Karlsruhe, Germany 29 April-2 May 2004.

Reviewed by Dave Phillips
Findlay, Ohio, USA

The 2nd annual Linux Audio Conference took place in Karlsruhe, Germany, from April 29 through May 2, 2004, at the Zentrum fur Kunst und Medientechnology (ZKM). More than 30 presentations were given, along with numerous impromptu workshops, four concerts, a dance, and a final roundtable discussion. Presentation topics included issues regarding Linux audio software development, design details of synthesis and composition programs, software instrument design, programming techniques for spatial audio, digital audio signal processing, sound recording and production, and graphic interface design details. In addition to this focus on software, Linux-based synthesis hardware was presented by the makers of the Lionstracs Mediastation X76 and the Hartmann Neuron Synthesizer.

Given the number of presentations, I am unable to relate much more than a superficial account of the conference, so I will limit myself here to giving brief reports on the events I attended. The organizers found it necessary to schedule parallel presentations, so it was not possible to attend every topic and demonstration.

The quality and depth of the presentations naturally varied, and the overall tone of discourse was technical without becoming obscure. Presenters included citizens from many European nations, but all presentations were delivered in English, a testament to the intelligence and sophistication of the participants.

I must confess that I found it difficult to attend some presentations simply because I was meeting so many friends and having so many interesting conversations. Often I looked at the clock only to find that I had missed a topic completely while engaged in a most absorbing impromptu dialog, and I suspect I'm not the only person who had that "problem."

Day 1
The kick-off address, delivered by this author, summarized the various activities of the Linux audio development community since last year's conference (which was also hosted by ZKM). My speech was followed by presentations by Jaroslav Kysela and Takashi Iwai. Both developers are core Advanced Linux Sound Architecture (ALSA) programmers, and their presentations naturally focused on the advances and future of the ALSA sound system (www.alsa-project.org). ALSA is now the default sound system for the Linux kernel, so Mr. Kysela provided an overview of the project's history in the past year. Mr. Iwai's presentation focused on user-level troubleshooting techniques, demonstrating many useful tools and utilities.

Paul DavisPaul Davis is a widely respected Linux audio developer, both for his outstanding contributions to the software base as the chief architect of the Ardour (www.ardour.org) and JACK (jackit.sourceforge.net/) projects, and for his generosity and civility toward his colleagues. His first presentation gave us a look at his recent FreeST (fst) project, a system for seamlessly integrating VST/VSTi plug-ins into the Linux audio software universe (see Figure 1). VST/VSTi plugins are fundamental to the Windows/Macintosh audio software worlds, and their incorporation into Linux is a significant achievement. Incidentally, Mr. Davis's project builds upon work begun by Kjetil Matheussen at the Norsk Nettverk for Teknologi, Akustikk og Musikk (NoTAM) in Norway (www.notam02.no/). The fst project is an expansion of that work by Mr. Davis and Torben Hohn, author of the gAlan audio synthesis system (galan.sourceforge.net/).

Developer Bob Ham revealed plans for his Linux Audio Session Handler (LASH), a system for saving and restoring the states of, and connections between, any number of LASH-aware audio applications (lash-audio-session-handler.org/). LASH is a greatly needed system: as Linux audio applications continue to board the JACK train, a means for saving and restoring their states becomes most valuable. Restoring the connections for a few applications is not particularly burdensome, but as more applications are used a system such as LASH becomes a necessity. LASH is still in its infancy, and interested developers are urged to contact Mr. Ham via the Linux Audio Developers mailing list (www.linuxdj.com/audio/lad/).

Fernando Pablo Lopez-Lezcano is perhaps best known to the Linux audio software community as the developer and maintainer of the Planet CCRMA packages (ccrma.stanford.edu/planetccrma/software/), a set of RedHat Package Managers (RPMs) designed to provide a simpler entry point for new users interested in learning about Linux sound and music software. However, he is also a respected composer and teacher, and his conference presentation focused on his use of the Lisp programming language in the Common Music and Common Lisp Music composition and synthesis environments. This presentation was successful enough to warrant an unscheduled workshop, a good indicator of interest in the subject and Mr. Lopez-Lezcano's abilities as an enthusiastic instructor.

The first day's final presentation came from Thomas Grill, presenting his flext system (www.parasitaere-kapazitaeten.net/ext/flext/), a means whereby writers of external additions for the popular Max/MSP and Pd sound synthesis systems can write code-compatible versions of their extensions.

After the presentation series, the first day ended with the first of four scheduled concerts, presenting works by Michael Edwards, Ludger Brümmer, Fernando Lopez-Lezcano, and Orm Finnendahl. All pieces were composed with the use of free software tools, and each piece had its singular attractions. The overall style could perhaps be summarized as "non-beat-oriented electroacoustic music," and while it may not be to everyone's liking, I'm quite fond of such music and I greatly enjoyed this concert. Incidentally, all the conference concerts took place in ZKM's Kubus, a marvelous hall designed for performance and recording.

Day 2
The second day's schedule began with two presentations by Pd gurus Orm Finnendahl and Frank Barknecht (see Web site www.pure-data.info for more information about Pd). At the same time, Martin Rumori presented his work on the foo sound synthesis language (foo.sourceforge.net/). After lunch, the presentations continued with topics from Ivica Bukvic, on his RTMix interactive multimedia performance software (meowing.ccm.uc.edu/~ico/index.html), Yann Orlarey, on the FAUST audio programming language (www.grame.fr/Research/list.php), Han-wen Nienhuys, on music notation with LilyPond (lilypond.org/web/), Matthias Nagorni, demonstrating his Alsa Modular Synthesizer (alsamodular.sourceforge.net/), and Fons Adriaensen, showing off his AEOLUS pipe organ synthesizer (users.skynet.be/solaris/linuxaudio/).

Other demonstrations and presentations included Orm Finnendahl's very personal use of Linux as a composer's workstation, Marije Baalman's use of wavefield synthesis as utilized in her sound installation at ZKM, Benno Sennoner's revelations regarding the LinuxSampler project (www.linuxsampler.org/), and a report from Christian Muhlethaler and Alexander Schuppisser on designing client software for the SuperCollider3 synthesis environment (supercollider.sourceforge.net/).

While these presentations took place, other conference participants were on hand for a public Linux installation fest. For users a little anxious about actually installing Linux on their computer there were also "live" CD-ROMs available from SuSE, the AGNULA/Demudi project (www.agnula.org), and the APODIO group (www.apo33.org/raccorps/article.php3?id_article=68). "Installfests" have become quite popular at Linux conferences, providing an excellent opportunity to see how the job is done and for posing questions to some of the outstanding talents in Linux development. The "live" Linux CD-ROM has also become a popular way to introduce interested users to Linux without touching the base hardware, i.e., nothing is necessarily installed to the user's hard-disk. The AGNULA/Demudi and APODIO live discs provide a wonderful means of introducing the world of Linux audio software to potential new users and are likely to become a common first experience for many converts.

Stephen Bernsee (a.k.a. Stephen Sprenger) and associates provided the day's final presentation with a demonstration of the Hartmann Neuron Synthesizer (www.hartmann-music.com/home/). The Neuron is a Linux-powered general-purpose synthesizer with a unique software engine, using a neural network to process and shape sounds based on the output from an audio analysis program. Analysis/resynthesis is itself a venerable method of creating sound with computers, but the Neuron is unique in its use of a neural network at the resynthesis stage. The sound of the synthesizer was quite good, and I would like to hear more from this interesting hardware.

The Neuron was not the only Linux-based hardware demonstrated at the conference. The Lionstracs group set up their Mediastation X76 (www.lionstracs.com) in a main hallway at ZKM for all to hear and play. I greatly enjoyed this machine; it has great sound and is filled with more musical amenities than I can count here. Dominico Colturato, Benno Sennoner, and the rest of the Lionstracs team have much to be proud of with the Mediastation, and I sincerely wish them great success in the world of hardware synthesizers.

Day 3
The third day of the conference began with Paul Davis's eagerly anticipated presentation on recent developments in his Ardour project (www.ardour.org). Ardour is a major project in the Linux audio software development community, an advanced hard-disk recording and editing system squarely aimed at the professional recording environment, and the strong attendance for this presentation clearly reflected its importance. Though plagued by crashes caused by some last-minute programming, the demonstration was enjoyable and illuminating. Ardour is a deep application, and Mr. Davis's presentations of his software always impress me. Invariably, I leave his demos saying to myself "I didn't know Ardour could do that!", and I suspect many people left this demonstration thinking the same thought.

Next, developers Steve Harris and Joern Nettingsmeier provided a two-hour "Audio Engineering in a Nutshell" presentation that enhanced and extended Mr. Harris's topic from last year's conference. Alas, I was unable to attend this presentation, opting instead for Stefan Kersten's informative status report on his work to provide a Graphic User Interface (GUI) for SuperCollider3 on Linux (supercollider.sourceforge.net/). The day continued with topics from Victor Lazzarini, on signal processing application developments at the National University of Ireland, François Déchelle and Patrice Tisserand, on audio networking at the Institut de Recherche et Coordination Acoustique Musique (IRCAM), Julien Claassen, on Linux sound and music software without GUIs, and Fons Adriaensen, on tools for precise audio measurement.

Again, I was unable to attend most of those presentations due to scheduled meetings with other interest groups. I did take the time for Steve Harris's excellent presentation of the JAMin mastering suite (jamin.sourceforge.net/). Mastering audio is one of those mysterious blends of science, engineering, and art, and Linux can now proudly claim an outstanding tool for this unique and necessary stage of the recording process. Mr. Harris's demonstration provided an overview of the mastering process and revealed much of the rationale behind JAMin's design.

Following a break for dinner and beer, it was music, music, and more music. The second scheduled concert was really more of a musical demonstration, with developers Julien Claaasen, Matthias Nagorni, Fons Adriaensen, and Thomas Grill showing off the musical capabilities of their software. This event was followed by yet another concert that continued the trends of the first night's concert, with compositions from Ramón González-Arroyo, Matthew Burtner, Torsten Anders, Ivica Ico Bukvic, Panayiotis Kokoras, and ZKM's own Ludger Brümmer.

For those participants whose musical desires were unfulfilled by three concerts, this day ended with electronica demonstrations and performances from Malte Steiner, Frank Barknecht, Eugene Kim, the ap group, Krzysztof Gawlas, Greg Kellum, and Diemo Schwarz. The evening ended with a dance party featuring music from DJ Lego, with students and other local citizens mingling with the conference population. It was a great opportunity for drinking, dancing, and casual conversation, and I hope the organizers plan another such evening for next year's meeting.

Day 4
Computer music stalwart Dave Topper started the last day of the conference with an introduction to his Graphical Audio Interface Application software (ftp://presto.music.virginia.edu/pub/gaia/), while at the same time Stéphane Letz reported on his work at GRAME, Centre National de Création Musicale, Lyon, France (www.grame.fr/) porting and integrating the JACK audio server to the Macintosh OSX operating system. Mr. Topper was then joined by Matthew Burtner for a discussion of their work on "recombinant spatialization for ecoacoustic immersive environments" (sic). Andrea Glorioso summarized and projected the work of the AGNULA project (www.agnula.org), then I presented my own topic on various aspects of writing and publishing documentation for Linux audio software, emphasizing issues surrounding user-level documentation and describing my experiences as a Linux journalist and audio software documentation writer. Following my presentation, Ivica Ico Bukvic discussed his experiences with, and recommendations for, introducing Linux as a viable alternative in the academic music curriculum.

The final scheduled event was a round-table discussion of the future of Linux audio software. Panel members included Fernando Lopez-Lezcano, Paul Davis, Takashi Iwai, Andrea Glorioso, Ivica Ico Bukvic, and myself. This discussion was a rather loose and casual conversation that served the purpose of wrapping up this wonderful conference, with subjects ranging from predictions for the future of Linux audio software to our personal plans upon arrival home. At the end of the discussion, plans for next year's conference were announced along with the happy news that LAC 2005 would be held again at ZKM.

A long last hurrah was held over pizza and beer at an excellent local restaurant, and once more we parted ways in the wee hours of the night, this time scattering to the corners of the world. I think it's fair to say that we all returned home with new resolve and inspiration, and I'm sure there will be many great things to come as a direct result of this year's conference.

Final Comments
The first Linux Audio Conference in 2003 was successful enough to acquire greater sponsorship for LAC 2004, with significant contribution from ZKM itself and from SuSE, manufacturer of the leading European Linux distribution.

Archived audio files, photographs, and other documentation regarding LAC 2004 are available at the conference Web site (www.linuxdj.com/audio/lad/eventszkm2004.php3). Special praise and thanks go to Joern Nettingsmeier for his work setting up and maintaining the live audio feed and Internet Relay Chat (IRC) channel throughout the conference.

Great thanks and appreciation also go to Goetz Dipper, Frank Neumann, Matthias Nagorni, Ludger Brümmer, and the administration and staff at ZKM for their technical assistance and personal attention to the many matters and arrangements surrounding a conference of this size and scope. Schedules were maintained smoothly, technical preparations were complete and ready for use, and help was always available to anyone needing directions or any other further assistance. ZKM is the perfect place for a Linux Audio Conference, and I am very happy that the Zentrum has offered its facilities for next year's conference. If you didn't make it there this year, you will have another chance to see Karlsruhe, ZKM, and the remarkable crew of Linux audio developers and users during LAC 2005.