Leonard Ritter

In this article I'm going to tell a bit about my past experience with song formats (advantages and problems), how it influenced my choices when extending the goals of Aldrin and what this will mean for the future of Aldrin, and collaborative music composition on the internet in general.

When I was about 13 years old, I got my hands on a cool program named Scream Tracker 3. It was providing abilities only known to be properties of expensive Akai Samplers, and even more: an integrated way to "script" entire songs using grid-like sheets.

I knew sequencers and had composed a few tracks on an Atari ST computer, which accessed an external Korg M1 Wavestation. Tracks had to be recorded on tape, and the process was complicated. There was not much of a chance I could show to anyone else what I did.

Thus, I was happy to find a program that (theoretically) not only allowed me to make music that sounded professional, because it was using top notch sampling technology, but also offered the ability to share my music. Scream Tracker saved all parts of a song - its patterns, arrangement and samples - into one file, a so-called Module, which could then be given to others and, if they used the (freely available) program or a player, listened to just as it was intended. They would not need any more expensive Hardware - besides a computer.

However, there were compatibility problems: Modules were not playing correctly with all players. The tracker program I was using by then, Impulse Tracker, allowed to use a quite advanced lowpass filtering technique that no player supported. Impulse Trackers code was not open, and back then not many had really much of an idea how to implement filtering properly. Sometimes, even sample offset effects were missing. It was not assured that everybody would be able to listen to my music as it was intended to sound.

Later on, the lossy audio formats we know all so well now emerged: Mp2 and Mp3 (and later on, Ogg Vorbis). Everybody could listen to them. Suddenly, I could pass around perfect copies of my music, as it was intended to sound. The module was dead.

Almost dead. When I collaborated with someone else, we needed each others original samples and pattern data to improve on the work we did. We were still passing around modules for pre-release work.

By then, I was using Jeskola Buzz, a tracker which used synthesizers and expensive sounding DSP effects to allow, for the first time, ways to (theoretically) compete with music that was popular in the clubs or on the charts.

But Buzz relied on countless numbers of external plugins, so-called "Machines". Collaboration became difficult: not only the modules had to be shared, also the machine libraries had to be put in the correct directories. Since there was no easy way to find out what kind of machine a module required, step by step addition of missing machines was necessary. That situation never changed. I ditched collaborative efforts several times because it was not feasible. There were simply too many machines, sometimes even the same in different versions, to keep track of.

Since a few years, my music is available through a Creative Commons license. Listeners are invited to share it and use it in their creative work. I also started to collaborate on ccMixter, a site dedicated to endless threads of people remixing each others CC-licensed content.

But how can someone remix an audio mixdown? Samples and notes can not be extracted without beyond-todays-sciences efforts. ccMixter solves it by making midi tracks, sample packs and vocal recordings the base of any track submitted. But from the resulting remix, it can only be guessed how and where these samples were used. If the artist introduced new samples, I could not get my hands on them.

These problems were inherent with audio mixdowns. Using the programming metaphor, you could say that samples, arrangement, sequences and tracks were the "source code", where the mixdown was the compiled program - hard, if not impossible to reverse engineer it in a way that would tell about the artists craftsmanship and sources.

When I started work on Aldrin, a successor to Jeskola Buzz, I remembered about the problems I had. Thus I decided that Aldrin would become a Monolith: plugins would always be tightly bundled to the host. This way, modules could be shared, listened to and edited quite instantly. I designed a new module format that allowed easier disassembling without prior knowledge of the format: simply structured binary formats, human readable configuration scripts and FLAC/Ogg Vorbis encoded samples packed in a Zip archive. The resulting format is called ZXM, "Zzub eXtensible Module" (I'm not married to this name, it is likely to change).

ZXM Modules have been designed for easy sharing between musicians and listeners. They could become a standard format supported by sites like ccMixter, allowing artists to build upon each others "source code", instead of decompiling and reassembling fragments of mixed music.

Therefore, to enable artists to collaborate thoroughly will become a main goal of Aldrin. There will be means to attach a Creative Commons license to your track within the program. There will be embedded access to the freesound project site, to up- and download new samples. And there will be support for a DSP scripting language that allows to write new plugins which can be distributed along with modules.

I hope that these features will be seen as opportunities for the online musician community. We could make the internet our studio, sit together although living hundreds of miles apart, and together compose the musical piece of musical pieces, a soundtrack that sums up the meaning of our very existence.