Linux Audio Developers

Die Tagung der »Linux Audio Developers« soll in einer Reihe von Vorträgen aus erster Hand über den aktuellen Stand von Audio unter dem freien Betriebssystem Linux informieren. Nachdem sich die »Advanced Linux Sound Architecture [ALSA]« weitgehend stabilisiert hat, entwickelt sich die für Linux verfügbare Audio-Software rasant. Das Spektrum der Vorträge reicht von den technischen Grundlagen der Sound-Architektur [ALSA, JACK] bis hin zu Anwendungen aus den Bereichen Harddisk Recording, MIDI Sequencing, Software-Synthese und Notensatz. Im Rahmen eines Vortrags über LADSPA [Linux Audio Developers Plugin API] wird auch auf Methoden der digitalen Signalverarbeitung [DSP] eingegangen. Der Vortrag am Freitag um 20.00 Uhr [Übersicht »Linux und Sound« in deutscher Sprache] und die Session III am Samstag ab 17.00 Uhr wenden sich besonders an Anwender, denen softwaretechnische Aspekte weniger wichtig sind.
Zum Ausklang der Tagung findet eine »Linux Sound Night« statt. Hier wird gezeigt, wie man unter Linux Musik spielen und produzieren kann. Die Zuhörer sind dabei eingeladen, sich durch Fragen nach bestimmten Anwendungen am Ablauf zu beteiligen.

Die Vorträge werden, soweit nicht anders angegeben, in englischer Sprache gehalten.

 

Abstracts of Talks/Presentations

Paul Davis (Linux Audio Systems, Bala Cynwyd, Pennsylvania, USA)

»10 Things You Might Not Have Thought About When Writing An Audio Application«

After nearly 3 years of working on Ardour, a Digital Audio Workstation for Linux, a number of key ideas concerning the design of audio software have become more obvious. I will talk about a set of issues that I believe should be considered by all authors of audio application software. These will include:

• buffering strategies for hard disk recording and playback
• mutual exclusion/critical region design
• handling MIDI/MMC and GUI controls in the same program
• the importance of musical structures and concepts
• the central role of tempo
• positional and speed synchronization
• the difficulties of handling MIDI correctly

The talk will be focused on code- and design-level issues, with a particular emphasis on how things that seem like they should be simple or easy or both are often not.
 

François Déchelle (IRCAM, Paris, France)

»Various IRCAM free software: jMax and OpenMusic«

This talk will present two free software developed at IRCAM: jMax, a visual programming environment for real-time audio and multimedia, and OpenMusic, an object-oriented visual programming environment for music composition.

jMax recent developments includes JACK support, integration into LADSPA (i.e. the possibility to run a patch as a LADSPA plugin) and development of alternate GUI using Python/GTK.

OpenMusic is based on CommonLisp/CLOS. The port to Linux will be shown. Although being a work in progress, it already demonstrate the essential functionalities of the environment.
 

Frank van de Pol (Bergen op Zoom, Netherlands)

»Timing and Synchronisation for Sequencer Applications«

Providing proper timing accuracy to ensure sounds are made audible at the right time is a requirement for sequencer applications. Combination of digital audio and event based sequencing gives application developers new challenges. In my talk I'll explain the concepts of timing and synchronisation, including:

• understanding of time reference
• managing delay
• soft synths using call back API vs MIDI synths
• synchronisation with external sources
   

Andrea Glorioso (Centro Tempo Reale, Florence, Italy)

»AGNULA, A GNU/Linux Audio distribution«

AGNULA is an acronym for A GNU/Linux Audio distribution, an EC-funded project (IST-2001-34789) which will be active up to the end of March 2004, and will probably continue with other funds and contributions beyond that date. Important features of AGNULA are:

it is the first EC project to support Free Software exclusively (GNU/GPL licenses and compatibles)
it will provide two full-fledged GNU/Linux distributions dedicated to Audio and multimedia (one Debian-based, the other Red Hat-based)

Steve Harris (IAM – Intelligence, Agents and Multimedia Research Group. University of Southampton, Hampshire, UK)
»Digital Signal Processing and the LADSPA Audio Plugin Interface«

The basic operations that can be applied to an audio signal, the relationship between them and their effects will be presented in non mathematical terms. It will be shown how these basic operations can be combined to make classic effects. The talk will cover filtering, modulation, delay, phasing, chorus and frequency shifting.

The Linux Audio Developer's Simple Plugin API (LADSPA) will be introduced as tool to share digital signal processing algorithms among sound applications.
 

Takashi Iwai (SuSE Linux AG, Nuremberg, Germany)

»Ruminations on ALSA Drivers«

After a long chaotic time with everlasting API changes, ALSA (Advanced Linux Sound Architecture) is becoming more standard on the Linux world. Now it's time to stabilize toward version 1.0 and at the same time it's time to consider what we can do in future. The talk will cover the topics regarding to the ALSA drivers, including:

• the difference of hardwares; PCMCIA and USB
• comparison with other operating systems
• the memory and buffer management
• linux-kernel specific problems
   

Jaroslav Kysela (SuSE Linux AG, Ceske Budejovice, Czech Republic)

»ALSA – Always on the run«

The talk will touch some interesting mechanisms/implementation details used in the ALSA driver and library. I will also discuss the integration of ALSA drivers with the Linux 2.5 kernel tree, last ALSA code updates and future development (short time).

alsa-driver

• used mechanism to keep ALSA drivers synchronized with the Linux 2.5 kernel
• hotplug support (graceful disconnection of running hardware)
• the snd_device_t structure

alsa-lib

• hardware abstraction
• flexible configuration syntax
• implementing of runtime configuration for active handle
• mixing of multiple streams into one (mixing daemon)
   

Fernando Pablo Lopez-Lezcano (CCRMA/Stanford, California, USA)

»Soft landing on Planet CCRMA, a small Linux/Sound/Music world«

Planet CCRMA is a collection of rpm packages that you can add to a computer running RedHat 7.2, 7.3 or 8.0 to transform it into an audio workstation with a low-latency kernel, current ALSA audio drivers and a nice set of music, midi, audio and video applications. Planet CCRMA is easy to install and mantain as it uses apt (Advanced Package Tools) to manage the package collection. It can be installed and upgraded over the Internet from the Planet CCRMA apt repository or its mirrors, or from cdrom iso images. Planet CCRMA replicates the Linux environment we have been using for years here at CCRMA for our daily work in audio and computer music production and research.

This talk will focus on how Planet CCRMA came to be, how to install it, what applications are currently included, future growth potential and advantages and disadvantages of this approach to building a multimedia linux workstation.
 

Dave Phillips (Detroit, USA)

»Linux Audio Software Documentation: Problems & Solutions«

Documenting Linux audio software poses unique problems to the technical writer. The scope of an application's use, the assumed abilities of the targeted user base, the rate of change in the Linux audio software world, and selecting the best documentation format are only some of the problems faced by the documentation writer. Writing good documentation is usually a far more difficult task than it appears, requiring considerable time and energy dedicated to work that is likely to pay nothing at all, needs consistent updating even during its composition, and will still be out of date when it reaches public distribution. Resolving these difficulties calls for new approaches to an old problem. Group writing, 'docs-for-pay', 'docs-on-demand', hiring outside help, and other approaches will be discussed, along with a consideration of some examples of current approaches to documentation and their strengths and weaknesses.