There are very few wedge-shaped synthesizers in existence. Some come close in the interesting console-style angles of the EMS VCS3, or the Korg MS20, or perhaps the atomic power plant panels of the Oberheim Four or Eight Voice systems. Convenience and compatibility, along with ease of manufacture, tend to favor the flat or the mildly rakish. This was not a consideration Triadex had entertained when they designed the Muse. For them, it was all about the triangular prism.

Perhaps the interest in mathematical shapes stemmed from the fact that the Muse came not from a typical musical instrument manufacturer, but instead from the scientific halls of MIT (Massachusetts Institute of Technology). It was designed in the late 1960s and early 1970s by professors Edward Fredkin and Marvin Minsky, who were far more interested in digital systems, artificial intelligence, and investigating the interaction of non-intelligent parts than specifically in innovating in electronic music. Minsky’s work had found its way into Arthur C. Clarke's novel 2001: A Space Odyssey, in which he was credited with making the breakthrough in AI that ultimately led to the creation of HAL 9000. Later, Minsky became an advisor to Stanley Kubrick for the film adaptation of the book. Fredkin trained as a jet pilot for the US Air Force in the 1950s and discovered an interest in computers when he was posted to the Lincoln Department of Defense R&D lab at MIT. He was later recruited by research firm Bolt Beranek & Newman (BBN) and worked on the PDP-1 computer prototype, writing an assembly language and its first operating system. This is where he first came across Minsky.

×

Fredkin’s output was ferocious. He invented a machine that let the Air Force scan and analyze radar information and photographs. The subsequent company went public, and he made enough money to buy an island in the Caribbean—quite literally. It had no drinking water, so he developed a reverse-osmosis technology to desalinate seawater. He later sold the island to British billionaire Richard Branson. He had a hand in developing intelligent chess machines and created the Fredkin Prize, offering a $100,000 reward to whoever developed the first computer program to win the world chess championship—a prize claimed, of course, by the IBM programmers of Deep Blue in 1997.

But we digress: in 1968, Minsky invited Fredkin to work with him at MIT. Fredkin went on to be the director of Project MAC, famous for groundbreaking work in AI, operating systems, and computations. Minsky won the Turing prize in 1969 and ran the AI elements of Project MAC.

Minsky, a lifelong musician, had a long-standing interest in electronic music—having previously worked on developing early computer music language EUTERPE, an early attempt at creating a "real-time" computer-based composition process. Between an interest in tape music, computer music, and algorithmic processes, Minsky collaborated with Fredkin to develop, build, and sell around 300 Muse synthesizers. Within the scope of each of their lives' work, Triadex is hardly a blip in terms of significance and influence—which, of course, makes it all the more fascinating.

Examining the Triadex Muse

×

According to the patent filed in 1971, the Triadex Muse was a "Digital Music Synthesizer". To quote the abstract:

In the apparatus disclosed herein, a note generator is controlled by a long-term, quasi-periodic function which is in turn generated by applying digital feedback in preselected combinations around a digital register. The register comprises means for holding a plurality of bits of digital information in a given order, e.g. a shift register or counter, the held information being changeable according to a predetermined pattern in response to input signals applied thereto. Digital feedback is provided by applying to the register at least one input signal, which is obtained according to a preselectable or adjustable code from bits of information obtained from various points in the register itself. The apparatus thus, in effect, composes music as distinguished from merely synthesizing sound.

As with any good patent abstract, that is a truly dense piece of prose!

To call the Muse a synthesizer is a bit of a stretch, considering it had only a simple square-wave oscillator and no other timbre or modulation controls. It used a "divide down" process of pitch generation, which uses a high-frequency clock to drive a counter that divides the frequency to the required pitch. It was more an experiment in algorithmic sequencing, where the oscillator served as a simple indicator of the process. They were attempting to show that a collection of digital systems could, if properly motivated, compose music.

×

[Above: Multi-position switches on the Triadex Muse, used to determine aspects of its algorithmic sound generation structure.]

Music was composed through the interaction between 8 slide switches and 40 possible positions. The human user's job was to set the conditions for the machine to bite down upon with its discrete digital circuits, logic gates, shift registers, and binary counters. There were 14 trillion possible note combinations, and apparently a Muse was left running for years in an exhibition at the Boston Museum of Science, playing a continuously changing piece of music.

The first four sliders are called "Interval" switches, and they select from a number of binary sources spread across the 40 positions. These can be simply on or off, clock pulses, a 4-bit binary counter, a 2-bit binary counter, or a 31-bit shift register. Interval A, B, and C are mapped to the major scale, and Interval D is all about octaves. The other four sliders are called "Theme" switches, and they use the binary sources to feed the next value into the shift register. The Theme sources are combined with an Exclusive NOR gate to give 1 on an even number of ones or 0 on an odd number.

×

[Above: The array of LEDs used to display generative melodic activity on the Triadex Muse.]

At a very basic level, the Intervals determine the notes, and the Themes determine the patterns, or order in which the notes appear. Without any Theme interference, you get a simple scale of repeating notes. As you add Theme switches, the sequence appears to randomize and evolve. This is down to the bits being fed back to what's called a Linear Feedback Shift Register (LFSR). With 31 bits available, it's possible to reach 2,147,483,647 steps before repeating and even more as you add further sources.

I would like to think I have some kind of handle on what is happening as I move the sliders and make certain decisions, but I don't. The Intervals are definitely graspable, but the Themes really start to get baffling, and all you can do is experiment and enjoy what happens. You can try an online emulation to get a sense of just how unusual this method of interaction can be.

×

[Above: Additional panel controls for the Triadex Muse's generative melodic behavior.]

There are other controls that add to the performance. On the left, you have sliders for volume, tempo, and pitch range. The sequencer can run automatically or be manually stepped through. You can switch on "Rest", which adds a rest instead of the lowest note. That is the entirety of the control you have over what could be a virtually infinite pattern of notes.

On the right of the Muse is a delightful column of cascading lamps that displays the state of the logic gates. It's very reminiscent of the shift register display on Eurorack modules such as the Turing Machine from Music Thing Modular. In fact, Tom Whitwell, the designer of the Turning Machine, cites the Muse as an inspiration for his module.

Triadex Muse Accessories

Minsky and Fredkin also designed some companions for the Muse to create a sort of complete system. One was the LS1 Light Show color organ, which housed four colored lamps behind frosted glass that would blink and follow the sequence. The other was the AS1 amplifier and external speaker to improve the built-in speaker, with the idea of giving better sound, although there really wasn't much sound to go on.

All three together, in the same wedge shape with wooden trim, looked rather splendid. The lights definitely add to the experience of using the machine. You could also potentially patch two Muse machines together to find harmonies. Unfortunately, the professors used a proprietary connection and control system, which means it's not compatible with any other synthesizers or music gear. It’s a shame, because that square wave can, of course, get pretty tiresome after a while.

The Muse in Hindsight

×

The lack of a real synthesizer voice inside the undoubtedly sophisticated and innovative sequencer was probably why it didn't sell very well. And, of course the two professors moved on to considerably more groundbreaking projects—the automation of composition was an intellectual curiosity, and the Muse was simply a byproduct of their experimentation in artistic uses for generative processes. Of course, it never sought to be an outright musical instrument, and instead was a device that simply could aid in the compositional process—or could, perhaps, provide an ever-evolving backdrop of musical ambience to your home, perhaps presaging many current-day electronic music/interaction methodologies.

It has recently acted as the basis for the Synthetic Sound Labs 2660 Matrix Sequencer, as has served as a more abstract reference for many other generative sequencing devices. It remains a lovely piece of work, and despite its rarity, it has made a lasting impression on the synthesizer community and modular makers who search for generative pathways in logic gates.

You Might Like

16hp Sale

Proteus Generative Humanoid Sequencer

Special Price $284.00 Regular Price $325.00

In stock

18hp Sale

Bloom V2 Fractal Sequencer

Special Price $437.00 Regular Price $499.00

In stock

12hp

2660 Matrix Sequencer

$350.00

In stock