A Tribute to Trackers

Before the advent of MP3 compression technology mid–1990’s, it was prohibitively hard to share music that appeared in the charts, or was released on vinyl. 

However, that did not mean that there wasn’t a way to make music with computers. Within European networks of computer users used so called music trackers to compose and share digital music. My project for this first lab is a tribute to these music trackers.

The popular (professional) form of using computers to make music in the 1990’s was to use them as a MIDI sequencer, sending control data to instruments such as drumcomputers, sound modules and synthesizers, mix and record this output to a DAT tape. The ‘score’ that was played could be saved to a ‘midi’-file, that could be copied and passed around. However, because the output device that would play back the midi-file differed, it would sound different. As such, it was hard for someone making music to share his output with other computer users. Recordings could obviously be shared, but the hassle of recording output from a small project studio, meant that only a very small group of people was able to do this.

Digital Music in an era of small memory and limited CPU power (1987 - 1995)

However, there was another way of making music, and the specifics of this method of music composition had everything to do with a specific type of computer. The Commodore Amiga, released in 1985 (Maher, 2012). Where most computers had a CPU to handle all computing tasks, the Amiga had a different hardware setup, using numerous custom chips to offload specific tasks, such as digital sound playback (the PAULA chip), graphics (Agnus and Denise). This divide and conquer strategy, made it possible of the amiga to play back digital audio. Albeit with a limited bit rate (8 bits) and a low sample rate (aprox. 28khz). Compare this to a Compact Disc which in the late seventies was already using a 16bit bitrate, and had a 44.1khz sample rate. So while is wasn’t pristine digital audio; it was digital audio. 

However, other constraints also applied, such as available memory, had drive space and memory bandwidth. This limited the use of recorded digital audio to 15–20 second loops, that would get repetitive very quickly. It was game programmer and musician Karsten Obarski, who hit upon the idea of combining sample playback, with instructions as to how various samples should be replayed. Think of it, as combining the instruments (samples) with the score (notes) as to how these samples should be played by the computer in order to reproduce music. His software, ‘The Ultimate Soundtracker’ became the blueprint for so called ‘music trackers’, ‘all inclusive’ composition programs that allowed the computer to be used for creation and playback of digital music. 

The ‘nice’ side effect of this approach meant that a song, tracked on Amiga A, would sound exactly the same on Amiga B. And, due to its use of samples (instead of synthesised sounds such as the Commodore 64 SID chip), music would sound somewhat more realistic than a series of beeps. Innovation in tracker-software quickened when it made the leap to the PC platform. It’s ample CPU power, that grew rapidly thanks to Moore’s law, meant that trackers could be expanded from their original Amiga limits. Various programmers created a multitude of tracker variants, each trying to outdo the other on features. Initially this meant being able to play more samples simultaneously (as the amiga was mixing audio in hardware, it was limited to 4 channels). A 1993 486 could play up to 16–32 channels. Another innovation came with the tracker Fast Tracker 2, which made an important conceptual leap. Instead of samples, the musician had instruments at his availability, made up of multiple samples, that could be mapped across key ranges, and combined with envelopes that described volume and panning behaviour over time, which meant that it was easier to emulate the characteristics of ‘real’ instruments. Again, the realism in music took a step, allowing musicians to express themselves in a more rich manner. Tracker music, which initially had only been used in demos (see new media heroes) and games, became its own distinct scene, as people started sharing more and more of their music. Initially through floppies and CD-ROMS, but this behaviour was accelerated when the Internet made it’s way into the hands of consumers. 

The fact that a module would sound the same on any computer, regardless of hardware, meant that sharing your module with the world, was similar to publishing music on a cd. Except, it would be a digital file. And, another nice side-effect from the way that this form of digital music file worked (containing both samples and score) was that in essence, every song was ‘open-source’. Songs could be taken apart, techniques could be studied, in order to improve your own musical composition ability. 

Digital Music in an era of dial-up modems and limited bandwidth (1995 - 2000)

Modules, as songs made with tracker programmes were called, were perfect to share in the limited bandwidth environment of the dial up internet of the second half of the ‘90s. They were usually 512kb to 1mb in size (zipped). And could easily be made available from a webpage or FTP-server. Thru the internet musicians started to organise themselves into groups (usually gathering around a shared social environment, or style of music). These groups would in turn professionalize in so called ‘net-labels’ music labels that released their music not on physical media but thru the internet. 10 years before the introduction of Spotify, with its streaming music model, teens and early 20-somethings created a complete virtual music industry, with charts, music competitions and labels. If there was an early case for what distributed connectivity could do, tracker groups and its associated scene, is a subject worth of close study.

However, technological progress, (friendly) competition between musicians would also lead to the decline of the tracker scene. While the open-source nature of the tracker format, allowed people to study each other techniques, it also triggered a lot of copy-cat behaviour, with songs being created that sounded so similar to its original source, it became hard to distinguish oneself as an artist. Coupled with musicians wanting to incorporate more complex sounds into their work, meant that the introduction of the MP3 format played a large part in the decline of the tracker-scene. Musicians would render their songs to large CD quality audiofiles, that, thru MP3 compression, were only somewhat larger than modules (in the range of 2–4 mb per song). This destroyed the possibility for other musicians to study techniques, while at the same time allowing musicians to have more ‘professional’ distance from an audience. 

Interestingly, it amazing how many of the patterns and behaviour that is apparent in todays music industry mimic the pioneering work of tracker groups in the late 1990s. From digital composition, to distribution, to reproduction, many of today's phenomena were already observable.

Making a Track

Making a track in a tracker is somewhat like editing a spreadsheet. Within the lower half of the screen you find your score, which is organised in patterns each pattern contains 2 to 32 channels (in Fast Tracker). 

On the top left, patterns can be arranged into songs (by connecting different patterns together). 

On the right of the menu, the instrument area can be observed. Here it possible select instruments and adjust their properties.

Fast tracker 2 also contains a sample editor, that allows for basic audio sampling.

Each channel is made up of a row, this row in turn has 3 columns.

A typical row looks like this:

C–4..2.40.A0F

This is interpreted on playback as:

‘Play instrument 2, with the pitch of a C–4, with volume 40 (64 in hexadecimal) and apply effect A (volume ramp) with intensity 16 (in hexadecimal)’

For the NMNT lab, I created some short chip-tune style tracks. In essence, with these tracks I was paying tribute to a tribute :-), they can be heard in the video below:

link to video

The problem with 1994 technology in 2016.

Actually, managing to get an old laptop running (as long as the BIOS battery isn’t dead) isn’t that hard. If it boots, it boots. The harder problem (which ended up taking way more time than expected, was actually getting content (modules) from a 2016 macbook pro onto a 1995 toshiba laptop (with no networking, no USB etc). 

I assumed this would be simple; it turned out to be not the case.

My naive approach: 

• • do stuf on the mbp in a dos emulation environment,

  • burn stuff to a cd, copy cd content to toshiba laptop hd

  • run from HD.

Turns out it’s really hard to buy burnable CD’s nowadays, so that became a no-go. In the end the solution went somewhat as follows:

• • Find a bootable Windows Machine that has a) network access, b) a 3.5inch floppy drive that ISN’T BROKEN.

  • Put said windows machine in your network. Without it dying from Windows-XP rot.

  • Move files thru the network from the MBP to said windows machine.

  • Find 3.5 inch floppies (thank god for rewritable media).

  • Insert floppy into windows machine, open a dos environment and type something similar to: c:\> copy c:\nmnt\music\. a:\

  • take floppy

  • copy stuff from floppy to laptop

That turned out to actually work and allowed me to do my demo in class. I will immediately admit that the journey to get stuff running on that machine was way more involved than the stuff that was actually running.

References

Maher, J. (2012). The Future was Here. MIT Press.