Trackers from 1981 and 1983

It is common knowledge that Soundtracker or Soundmonitor were the first trackers. But it’s not necessarily true. In the 1960’s computer scientists and Fluxus artists made things that were rather tracker-like (here, here) and there are probably precursors from hundreds of years ago.

But talking computer software, we can have a look at for example Muzix81, released in Hungary in 1983. This was a CV sequencer for the tiny ZX81, and was apparently used for a few mainstream record releases. The manual reveals that Muzix81 has a similar layout to eg Soundmonitor, with a list of hexadecimal instructions. At the top left, “1. 0B” means duration 1, octave zero, note B. “+” means sharp and “*” means accent.

Muzix81 has many similarities with a tracker, but it uses explicit duration. The user sets the duration of each event instead of using a fixed time for each step (like a step sequencer). Most trackers don’t do this, but there is a strain of trackers with explicit duration that are sometimes called editors. (note that Muzix81 has a step sequencer for drums though)

Irrlicht Project recently told me about a similar program, released even earlier. Electric Duet for Apple II came out in 1981, and was made by Paul Lutus in USA. It uses the same basic principle as Muzix81: on each step the user sets duration and note.

What makes this tracker-like, as herr Irrlicht himself notes, is that it uses a vertical time line (not sure if Muzix81 does?) and that it has text notation rather than sheet music. These two factors taken together, means that it is as much a tracker as the myriad of C64-editors from the 1980’s and 1990’s are. But this is from 1981.

Following this terminology, we could say that Muzix81 is a hypertracker (as the musical output depends on the settings on the external hardware) and Electric Duet is more of a soundtracker, since it simply plays the data that the user has typed in. But I’m not sure if that makes sense.

If you know of other early tracker-like software, feel free to send us a fax!

Ancient Trackers

The roots of trackers is normally traced to the mid 1980s with Soundmonitor and Soundtracker. That was the genesis, The Root, the big bang. Boom! As if trackers appeared out of thin air. But where do we look for the precursors to what became known as trackers in the 1980s?

First of all, let’s define a tracker as a vertical step sequencer with alphanumerical symbols where positioning controls time but not pitch. They are either absolute (for soundtrackers) or referential (as in hypertrackers). Although there are other types of trackers, this covers the bulk of them (read more here and here).

Secondly, we need to distinguish between the materiality and the language of trackers.

Materially speaking, trackers work more like step sequencers than as piano rolls. Each event has a set duration in time, usually with a rather crude quantization. In piano roll sequencers the composer positions the stave to set the pitch and defines the duration, with a quantization that often has a higher resolution than trackers do.

More than 1000 years ago, persian inventors were automating the playback of music: ie, programming music. In the music box each note has a timeline where the composer inserts events, like in piano rolls. In the 20th century, player pianos built further on these ideas and became quite commercially successful. In these media there are individual timelines for each note. One timeline for C, one for D, and so on. Raymond Scott’s Circle Machine from the 1950s instead used one single timeline where the composer could set the pitch on each step. Many modern analogue sequencers use the same logic. In the 1980s, trackers brought this logic into a coded computer environment.

In a tracker, the composer can program each step to set pitch, timbre, volume or various kind of modulations. Right next to the note column there is an effects column. In Soundtrackers there are preset effects and the composer sets the intensity of the effect: 1XX for pitchbend up, 3XX to slide the note, 4XX for vibrato, AXX to fade the volume, and so on. In hypertrackers, the effects refer to a table of data elsewhere, so you can program your own effects.

This is what I mean with the language of trackers. Each step in a tracker sends instructions to the soundchip about what it should do. Tracker songs are like a list of instructions, and the palette is much bigger than in traditional notation. In the 1960s, Lejaren Hiller (almost the inventor of computer music) and a Japanese Fluxus-guy made musical instructions that looked similar to what tracker songs do today. Vertical text-mode lists of instructions to control the music, although not for computers.

In trackers (for PSG-soundchips that don’t use samples) each step can, in theory, contain any instruction that the soundchip or the computer can understand. Even non-musical instructions, like sync points to visuals (a popular trick in the demoscene). Trackers usually don’t offer this kind of freedom but the C64-tracker that I use, defMON, does something along these lines. While other trackers make a clear cut between what an instrument is and what a modulation is, defMON doesn’t discriminate. Everything is everything.

Sheet music are also a list of instructions, and could be called software. Or, with a gigantic stretch, even the more fashionable “algorithm”-term could be used. But imho, tracker music really is deserving of the software term, both materially and linguistically. A tracker song contains both the composition and the style of the performance. All the author genius, all the funk, all the style. It’s all in there. The instructions determine the performance almost completely. But it is not the performance. It is not a documentation of the performance, like recorded music is. The performance always happens in real-time as you play the file, and it will sound slightly different depending on the hardware you use.

Historically, economically and politically it has been – and still is – important to categorize music as either performance, recording or notation. Who wrote the song? Who owns the recording? Who gets the money for performing it?

I’m still fascinated by the materiality of tracker music for this reason. There is much more research to be made about the precursors to trackers, both as an interface and as a format. The history sketched above is just one suggestion, and I’m looking forward to the criticism! Be my guest to check out صفی الدین الارموی‎’s musical notation from 800 years ago based on numbers and letters, Braille music, ancient numbered musical notation in China, and klavarskribo which is a vertical notation sort of like Guitar Hero.

 

More Data – Less Knobs! About Tracker Music Remakes

Two old Amiga music legends have recently decided to re-make some of their olden goldies. Lizardking and Mantronix made melodic synth music, somewhere inbetween … Koto and Jean-Michel Jarre? German space disco meets Italo Disco? As cheesy as some of it is, it’s kind of hard not to like it. Atleast when you’ve been brainwashed by it when you were young. In the Amiga scene this style is known as doskpop.

After a new Facebook group popped up, with an amazing energy supplied by Slash who started it, plenty of old tracker celebrities have been showing up. Moby finished some old songs, for example.

Many of us have probably wondered what tracker musicians could’ve done with more powerful technologies. Some people seem pretty convinced that it would automatically be better, because there’s “more possibilities”. Let’s move back to the 80’s first, when Rob Hubbard made synth-versions of his songs before programming them on C64. Let’s look at One Man and His Droid. Listen to the first version here recorded with various synth gear (among them TR-606 and TB-303!). Then listen to the C64-version below.

The first version is much more rough than the second one. Understandably, of course, since the first one is a sketch. But if we look at other music that exists as both tracker music and “synth music” (in lack of a better word) it’s not unusual that this is the case. I’ve heard it in most game musicians who “re-master” their old songs to sound more modern, for example. The tracker versions are just more detailed, intricate, ornamental. Anal? Crazy? Yes, perhaps.

Check the original version of Act of Impulse by Tip & Mantronix, and then listen Mantronix’ 2012-version. The new version is better in many production-wise ways, like clarity and punch and bass and all that. But one thing is, atleast to me, better in the old version. The melodies. It’s a lot more expressive. In the original mod-file you can see all the effort of the composer with setting custom vibrato, glide and volume tweaks for individual parts, and even notes. That’s sort of a pain in the ass with modern DAWs, but in trackers you were encouraged to do so because otherwise it looked empty and “sounded lame”. The more the better. Aesthetic maximalism, as I’ve called it before.

I have a similar feeling with Lizardking’s re-makes of his own songs. It’s miles better than all the generic SID/MOD-remixes that are out there, but I still miss some of that tracker trickery that is so characteristic of tracker music. It needs to sound more “data” like we say over here in the Swedish scene. Less generic knobs, more detailed numbers. No more synth! hehehe

Ok, well, this might be the most grumpy-old-man post I’ve made so far, so I think it’s better to just stop here. But just to clarify: this is not meant to diss anything. I love the data-sound, yes, but most people obviously prefer knob music. My point is that the newer software sort of brings us away from these “hand-made” solutions that were more common before, and drives us into a different kind of sound. Because even these “unlimited” platforms have a sound to them, don’t they?

 

Tomorrow: Speedcomp 9

A while back I wrote a short history of hardcore chipmusic, where I outlined the proto-breakcore pioneers of Australia. As some people suggested I forgot to write about the current situation there. Which is that the lo-fi tracker hardcore is still alive and kicking. At Thematics Radio they do themed shows, and some of it is tracker hardcore with people like Hedonist, Xylocaine and Abortifacient being involved.

They also do competitions called Speedcomps. You get a bunch of samples and have 30 minutes to make a song with nothing but those samples. Here s a link to the Facebook-page of the previous one, on March 15. So yeah, now go find the link to the next one and have some fun!

Don´t forget the time difference, though. This takes place in Australia which is always one step ahead of the rest. So set the alarms, load your software and get ready to rumble! You should be able to find more info here, but I am not sure since this post was written in the past and sent into the future so I wouldn´t forget.

A Fluxus Tracker from 1961 (sort of)

In a previous post I suggested that a print-out by Lejaren Hiller as the first example of a tracker. Now I came across another interesting example, which appeared even earlier.

Music for Piano No. 7 was made by Toshi Ichiyanagi (Fluxus) in 1961.The instructions are arranged vertically instead of horizontally, just like in a tracker. And it could also be seen as a text-mode representation. While trackers normally only use the alphanumerical parts of the charset in its notation, this utilizes the other symbols of the charset. Just like I did in Remote Control Music Studio.

I think this is interesting as an odd proto-tracker. But the other isntances from this series are not as text-modey and seem to work with both horizontal and vertical organization. So, it’s not a spot-on example. But still. I wanted to have it here for future reference. So please enjoy responsibly.

A Tracker From the 1960s?

 

Lejaren Hiller was one of the first people to generate music with a computer. He was doing it already in the 1950s, just like for example Douglas Bolitho and Martin Klein (info).

The picture above though, shows something else. It’s a dot matrix print-out with instructions for how to operate the volume and EQ knobs on your hi-fi system while playing the record “Program (Knobs) for the Listener”, released in 1970.

While others would surely salivate over the random (?) numbers and the interaction/remixism that this presents, I’m more interested in seeing it as a tracker. A primitive tracker, but nevertheless:

• It’s a text-mode list of instructions that runs vertically.
• There are discrete steps fixed in time and all the instructions are locked to these steps, like a soundtracker.
• The instructions are not absolute, but relative to whatever sound is coming from “under the hood” like a hypertracker.
• It’s divided into tracks, and the tracks affect eachother just like they do on many old soundchips.

Sure, you could see this as an analogue step sequencer, combined with the ideas of John Cage (who Hiller worked with). It’s only the print out that makes it seem like a tracker. Makes sense. But then again, it is the level of interface that is the most defining part of trackers. Trackers could use analogue synthesis and generative features. They just never do. :–)

Btw – some people claim that Lejaren Hiller did the first computer music, but that is not true. In Australia and the UK people made computer compositions and audio as early as 1951. See here.

But could we say that this is the first example of a tracker interface? Yeah, of course we can. This is Chipflip, where dreams come true. So who’s up for the challenge of finding something older that looks like a tracker? I’m sure it exists, right?

Soundtrackers, Hypertrackers and Acidtrackers

tl;dr – There are two kinds of trackers: soundtrackers and hypertrackers. But it’s a combination of them that is showing the way forward. And perhaps the micro-efficient trackers are more useful than ever, with the popularity of handheld devices.

When I wrote my thesis I had some difficulties to cover the topic of trackers. Although they are old and popular programs, there’s not much scholarly research on them. I can’t remember anyone trying to categorize trackers properly, for example. If you know of any such attempts, please get in touch.

For my thesis, I ended up talking about soundtrackers and hypersequencers. They refer to two dominant families of chipmusic trackers. Soundtrackers use sampled sounds and have a user-friendly interface. Hypersequencers are more about synthetic sounds and efficiency.

I find these two categories quite useful for discussing trackers in general. But I have also found that talking about trackers as hypersequencers (originally from Phelps) doesn’t feel quite right. Instead, I suggest the term hypertracker.*

So:

Soundtrackers are similar to sheet music, because they display notes and effects next to eachother. You can see which note is played, and also its ornamentation (vibrato, arpeggio, etc). The song is arranged in patterns, and one pattern includes one bar of all the voices. That means that all voices are locked to the same tempo, and the same arrangement structure.

Hypertrackers use more of a code logic. If soundtrackers are like sheet music with absolute values, hypersequenced music is like code that executes instructions. The note C might play a completely different note, depending on what kind of code is next to it. It enables a wild and “generative” composing style. Voices can have different tempos and sounds can be connected to eachother in a modular fashion. Hypersequenced music requires few resources (in terms of RAM, ROM, CPU) and mostly use synthetic sounds. They are “hyper” because they are referential; a letter or number usually refers to something other than itself.

Personally I find soundtrackers very convenient to use. They are straight-forward, simple and direct. Hypertrackers on the other hand, are more versatile and offer more surprises. They have more character somehow, and can lead the music in directions that the composer wasn’t aware of. Hypertrackers offer a lot of control and yet, as a composer, you can choose to hand some of that control back to the software. In soundtrackers it’s more up to the composer to take command.

Plenty of chip software doesn’t fit into these two categories. LSDj is an interesting example, since it takes inspiration from both. Obviously Mr. Kotlinski prefers hypertrackers. He even expanded the hyper-structure by adding more layers to the song arrangement, and by adding more tables. But just like one of his big sources of inspiration (MusicLine) it also incorporates some of the UI-ideas from soundtrackers. For example, you can set absolute effects next to the notes, such as pitchbend or vibrato.

This mixture of sound- and hypertracker became very popular in the chipscene. LSDj inspired LittleGPTracker, and created a new momentum. One example is Pulsar, recently created by Neil Baldwin who made 8-bit game music already in the 80’s. Even more recently, I’ve seen previews of new demoscene software that is highly inspired by LGPT.

These programs are not made for keyboards. They are designed for handheld consoles and very few buttons. Another difference from other trackers is that they can be used for live performances. Most trackers are pretty useless for live improvisations, unfortunately. A third difference: they can maximize the hardware. Trackers are normally designed to leave resources for code and graphics of demos and games, but this new generation allows you to use nearly 100% of the available resources. That is a fundamental difference, which is why chipscene Gameboy music can be more powerful than game/demo music for Gameboy.

The chipscene made chipmusic stand on its own feet, independently from the visuals, and that has affected the software too. New conventions have been developed, and it seems like future chiptrackers will follow this new path inbetween sound- and hypertrackers. It might also be used for other platforms with few buttons or low memory. Arduino and Raspberry Pi come to mind, aswell as smartphones with their complete lack of buttons.

In those situations I’d guess that “tracker” is a precise enough term. Just like  Renoise is a tracker, in a world of piano rollerz. But if there should be a new term for it, I suggest acidtrackers.

* I agree with HVMEC that trackers and editors are not the same thing. Trackers are step sequencers, while editors require the user to set the duration of each note (more here). The term hypertracker excludes programs like Soundmonitor or Future Composer, because they are editors. On the other hand, I think those kinds of programs are rare today. And perhaps they share more with MCK/MML or piano roll sequencers, than with trackers?

After the Trackers: John Cage Bukkake

Trackers have remediated plenty of Western ideas of music. Typical time signatures (4/4) and tonality (12-TET) are the most obvious. Less apparent is the distinct separation between instruments and notation; sound and code. Most trackers force the users to make strictly defined instruments which sound basically the same every time it’s triggered. As such, trackers are essentially the opposite to modular synthesis, where anything can modulate anything (ideally).

Perhaps that’s why trackers never seem to go mainstream. They are too deterministic and controlling. Too manual. The contemporary way is to have fun with stuff you can’t understand: nothing is a mistake á la Cage. It’s okay too be lazy, 2 cool 4 skool. So trackers like Renoise are going that way too, and seems to be getting pretty bloated in the process.

In a similar way, some of the most talked-about chipmusic tools are not trackers. New physical interfaces like Gatari and C64 keytar are obvious examples, but sometimes software also gets some attention. Nanoloop, of course, can be seen as a precursor to the now popular grid interfaces. Viznut’s Ibniz is more of a mathematics tool, but it got a lot of attention earlier this year. It’s been designed to make text-based generative works with a tiny filesize (sometimes called ?bytebeat). Since Ibniz works with both visuals and sounds, it also blurs the boundaries between visual interface and content, like little-scale also demonstrated a few days ago.

For me it seems clear that visuals and music will melt together in new forms of interfaces in the future. Let’s look at two experiments that can give some pointers for the future of low-tekk composing: Gijs Gieskes’ TVCV-sequencer and Chantal Goret’s mouse-controlled Crazy Box!

Taketracker Mystery Solved?

In a previous post we saw that Taketracker was listed as the second most popular Amiga tracker, which was odd since it’s an obscure DOS-software. According to Redneckerz it was the most powerful 16-channel-tracker that used the MOD format.

I looked into it a bit more. According to Stone Oakvalley, the SOAMC collection used Deliplayer 2 to identify the music format. Other programs identify these “Taketracker-songs” as either Sound-, Noise- or Protracker. Below you can see Stone’s screenshots.

There are TT-songs in SOAMC that were made by Maktone, Dubmood, Zalza, Josss, Zabutom, Nagz and Cerror. Atleast some of their songs are made in Fasttracker II, saved as MOD. Also, if an FT2-songs are loaded into Protracker and saved again, they are still identified as Taketracker (according to Deetsay aka Tero).

On the other hand, Beathawk’s ambient-tune says in the sample-info “hopefully ucan play this right with pc”. Rez says that his TT-songs were made with Protracker 3.52. And if you search the TT-songs’ sample-infos for “amiga” you can find more songs that were definitely made on the Amiga. There are songs by Allister Brimble and Moby, but also those pioneerings works by 4-mat around 1990 (which were made in Protracker 1.3b).

The header of a MOD file only seems to have rough details about which tracker was used. Afaik, there’s no way to know for sure which tracker was used to make the songs. But there are obviously different ways to go, since the TT-songs are identified so differently by various software.

Stone decided to relabel all the TT-songs as Protracker. You can see a list of all the the conversions here. Hopefully someday there’ll be a proper MOD-identifier around. Until then, the Taketracker mystery is not solved!

The Most Popular Tracker Ever?

Stone Oakvalley had his Amiga 1200 playing for one year, 24 hours a day, to make the largest MP3-compilation of Amiga music. 255 000 songs and 200 days of listening. The info page shows 194 different formats, and I wanted a list of the most popular ones (amount of songs per format, rounded off).

• Protracker – 77 200
• Taketracker – 6300
• Noisetracker – 5800
• MED – 3500 + 700 in OctaMED
• Sierra AGI – 1800
• Soundtracker – 1400 + 300 in Ultimate Soundtracker

Just below 1000 we can find MusicLine, Startrekker, TFMX, Sonix, AHX and custom players. The synthetic trackers (SIDmon, Art of Noise, DigiBooster, Sonic Arranger, SoundMon, Future Composer) are all less popular. There are several noteworthy ones that I haven’t even heard about: The Player, SoundFX and MusicDiskTracker for example. Maybe worth checking out?

The high amount of Sierra AGI-songs, is probably caused by jingles and sound effects being interpreted as songs. Or is there a subculture of Sierra ravers that I don’t know about? Also – what is Taketracker? It seems to be an obscure MS DOS-tracker from 1994, but there are over 6000 songs made in it here. Is it because 4-channel Fasttracker songs were converted to Taketracker, which was Amiga-compatible?

[update: Stone used Deliplayer 2 to detect the format of the music. The songs that DP2 detects as TakeTracker are identified as as Soundtracker, Noisetracker or Protracker by other software (Eagleplayer, Delitracker, Aplayer, Hippoplayer, Noiseconverter & ExoticRipper), as Stone showed me. With artists like Maktone, Josss, Dubmood, Zabutom, Zalza, Cerror, it’s likely a PC-tracker]

As Stone suggested in an e-mail, it’s probably a specific version of a Pro/Noise/Soundtracker that is misinterpreted]

Although Soundtracker was developed (hacked) for 3 years, there are only 1700 songs. Perhaps it is because 1) people did not use the mod-format but rather the song-format (which did not include the samples), 2) the files have been re-interpreted as Noisetracker/Protracker, or 3) the music was poorly archived.

Or perhaps the home computer music revolution only got started with Noisetracker and MED in 1989? Both these programs were developed until 1991 (when MED became OctaMED). But in 1990 Protracker appeared (as a Noisetracker-hack) and became the new standard.

With almost 80 000 songs, Protracker is more popular than the PC’s Fasttracker, Screamtracker and Impulse-tracker combined (source). Did any other music software leave such a massive amount of non-recorded songs? On the C64, we can use SIDId to see that no program was used for more than 10 000 songs (scroll down here). I suppose those numbers don’t include sub-songs though.

Anyway. Protracker – the most popular tracker ever? Also, is it the most popular 8-bit tracker in terms of vinyl releases? In the timeline there are 100+ vinyls made on Amiga, most likely in Protracker or MED. Protracker was very popular in the-stuff-that-became-known-as-breakcore on labels like Bloody Fist, Digital Hardcore and Fischkopf.

Btw – check this great diagram on the history of trackers. Useful++