Improvising with Machines case study: Mechanical Techno
The recent article for Organised Sound myself and Adam Parkinson wrote changed significantly from its first draft, which was weighted heavily towards case studies of our individual electronic music practices. Below is my case study, which describes one section of a typical live performance with the Mechanical Techno extended turntable setup.
In my Mechanical Techno project I use a standard DJ turntable as the basis for a spinning stack of modified vinyl records, each of which generates or triggers sound in a variety of ways. During live performances, each layer of the tower is added sequentially, separated by wooden cylinders skewered onto a knitting needle, as the whole structure continues to spin. As each new record is introduced a new musical component is added into the mix. The building of the stack dictates the structure of the music, but within that structure are various degrees of improvisation. The tower can be built in different ways using different combinations of components, modified records, external sound sources and effects. These can be combined in endless, unquantised ways.
The music made with this setup, whilst not always ‘techno’ by strict genre definitions, is repetitive, rhythmical, usually in 4/4 time and at between 122 and 144 bpm. In addition to a regular turntable tone-arm and cartridge, extended sound sources include analogue and digital synthesisers, small objects and acoustic percussion amplified with microphones, external sounds (eg radio static, oscillators or noise generators) amplitude- or filter- modulated by the system, and a range of unusual triggers and control mechanisms.
As soon as at least one of the sound sources is introduced, the platter will turn and sound will be produced: “it runs”. Intervention is always possible and might include changing the speed of the platter; repositioning tone-arms, triggers or controllers; introducing auxiliary effects or changing their settings; or nudging the angular position of records in relation to one another (thereby changing rhythmical relationships).
A typical configuration of part of a live performance might comprise the following:
Layer 0 (on the turntable platter itself): a 12″ white label of unidentified dance music, blanked off in alternate quarters with adhesive vinyl of the type used for graphics or signage on shop windows. An ordinary turntable cartridge is used, but the tone-arm is constrained in its lateral position by a piece of thread tied to a weight, positioned on the platform of the turntable. Each time a covered quarter of the record passes the needle, slight surface noise is heard but no other sound. On alternating sections, where the original vinyl is exposed, a snippet of audio is played back – a physically selected sample. Unless the original recording is at exactly 133 ⅓ bpm (or 90 or 180 bpm for a record intended to be played at 45rpm), the tempo will not match the rotation of the platter, and may or may not be in sync with the perceived rhythm. Additionally, the adhesive stickers themselves produce audible clicks which, due to their 90 degree separation, can emphasise the 4/4 beat. The grooves of the record are microscopically close together and, despite the thread, there is some potential for lateral movement of the stylus each time a blank part of the record passes. As such the needle sometimes jumps into a different groove of the record, playing a different sample in an unpredictable way, but forced into the same rhythm by the regular placement of the stickers: “it surprises” as physical wobbles of the turntable increase the likelihood of change.
Layer 1, 50mm above the first positioned on a wooded cylinder: another modified record, this time a smooth disk of vinyl, the unpressed B-side to a single-sided release. A line of scratches bisects the disk, around which sandpaper has been used to roughen the surface. A standard turntable cartridge is attached to a second tone arm, held in position to allow the needle to read the vertically raised disk. Twice each cycle, the roughly scratched section passes the needle and emits a burst of harsh, scraping noise. EQ settings on the desk and a short reverb shape the sound into something approaching an electronic handclap. The smooth parts of the record are not perfectly noiseless and sometimes sharp percussive clicks encroach on the rhythm. Over the course of a performance, the needle covers the same terrain hundreds of times, sometimes beginning to cut its own groove into the surface, and adding extra surface noise over time – “it evolves” to become noisier, and indeed no two rotations produce exactly the same pattern of noises.
Layer 2: A 300mm diameter disk made of 12mm MDF, laser cut with a regular pattern of 48 holes around its edge. Pegs are inserted into the holes to programme a rhythmical pattern. These holes are positioned to afford sixteenth note triplets. A contact microphone is attached to a wooden stick, held in position by a retort stand so that each time a peg passes, it flicks the piezo disk. The signal is fed to a Nord Drum electronic drum brain, triggering a digital (analogue modelling) kick drum sound. As the disk is now 10 cm above the platter of the turntable, and the faces of the wooden cylinders are neither perfectly flat nor completely parallel, some lateral and vertical wobble is apparent in the disk. Though at the trigger end this only translates to millimetres of variation, this can be enough to subtly change how the trigger responds: occasionally missing it, double-tapping, lightly brushing or playing earlier or later than in rigid metre – “it malfunctions” albeit in a minor way, the rhythm generated is imprecise and off-grid, microtiming variations affecting the groove in something akin to the difference between a drum machine and a live drummer playing the same pattern.
Layer 3: an optical reflection sensor, built for me by Tom Richards, reads white stickers placed on a standard vinyl record in programmed patterns. The sensor is a binary switch, sending a 5v control signal each time a sticker passes. The signal is sent to the gate input of an Arturia Micro Brute analogue synth with an internal sequencer. An independent LFO can vary the timbre of the resulting bass part over time. I have many different pattern disks to choose from, and each can have several ‘tracks’ of patterns, arranged around concentric circles on the disk, meaning they can be selected in real-time by changing the position of the tone arm. Part of the improvisation here is in choosing which disk to use, selecting which pattern from the disk, and also varying the relative rotational angle of the disk with relation to the existing rhythms. It’s something like using several step sequencers, all set to the same tempo but with non-quantised start-points. Additionally, the ‘locking’ of the synchronisation is dependent on the friction between the disks and the wooden blocks which separate them. It’s entirely possible that drag from one of the sensors or tone arms can shift the start point and therefore the rhythmical timing between each of the disks. Certain of the disks for the optical reflection sensor use divisions of eighth or sixteenth notes, but others use twelfths, divisions of ten or more random patterns (such as a series made from tracing irregular shapes from diagrams of bacteria). Whilst with practice it is possible to guess approximate rhythms by looking at the disks, in the moment of performance this is often a surprise. The sequencer on the Micro Brute can be programmed in real time. Each gate signal advances the sequence by one step, and the sequence can be anything between 1 and 64 steps long. As such, the number of triggers per cycle need not match the number of steps per sequence, and setting sequences which phase in and out of sync with the turntable’s loop point is simple and intuitive. These kinds of phasing patterns are not always predictable during programming, and can even change again if single steps are missed due to read errors – “it plays”, creating basslines, riffs and melodies which would be very difficult to accurately predict or plan for.
Layers 4 and 5 introduce more percussion sounds using the trigger methods above, playing another voice on the Nord Drum and using a second optical sensor to activate the clock of a Volca Sample sequencer loaded with percussion sounds. With each additional layer, the tower becomes slightly more precarious. Additional layers introduce more severe microtiming variations, asymmetric time-divisions due to the physical wobbling, friction and vibration which can affect other layers already in play.
Layer 6 is a wooden disk with evenly spaced pegs around its edge, each pair joined by an elastic band. Two piezo sensors dangle by wires over the disk, close to the edge but clearing the pegs. The sounds each make are tuned electronic percussion, somewhat like a thongophone or a donk (the electronic sound which gives that genre its name). Several ping-pong balls are introduced one at a time, rolling round on the platter and creating a random rhythm distinct from the more regular patterns the rest of the machine is producing. Physically flicking the ping-pong balls at the triggers changes the patterns. Whilst the pattern at first appears completely random, the diameter of the balls is such that the spacing between each when they are immediately adjacent is close to a sixteenth note at the edge of the disk. As such, bursts of donk rhythms can sound in time with the beat. The physicality of the setup is what gives the rhythm its character. The tension between randomness and regularity can feel like the machine is soloing on its two donk notes.
Working with the mechanical techno setup can feel like a collaboration with another entity, from my perspective as a performer, “it collaborates”. The available options – sample records, pattern records, trigger patterns, synth parameters – can be combined in endless configurations. But each time I make a decision it closes off certain other options. It feels as though the machine is limiting my choices at each stage. I think of the process of making a piece as one of optimisation, similar to a funk band ‘locking in’ to the groove of a track. Danielsen describes this kind of process giving an example of a James Brown funk track.
“For this reason it would be better to call it optimization than variation – optimisation of the different elements so that they become even more integrated and comfortable within the whole. This continuous optimization is often described as ‘locking’ or ‘nailing the rhythm.’ It is not a carefully considered process, and it never really ends; instead, it goes on automatically, continuously, manifesting in the form of better or worse periods of interaction.” (Danielsen 2018, p41)
Once a groove is up and running, my role can vary depending on how I view the pacing of the performance. Repetition is in the nature of electronic dance music, so allowing the machine to run is a valid performance decision: I can stand back and let the rhythm play for several bars and allow it to have its say. For me there is enough variation in microtiming, evolving noisebed, irregular and unexpected minor changes in timbre, timing and sample selection that the loop is never static.
Chris Cutler criticises the unchanging repetition of recorded loops: “Where biological systems are creative and unreliable – qualities which I believe are profoundly linked – mechanical or electronic systems are unerringly accurate, but mindless.” (Cutler 2018, p68) His defence of human-played repetitive music revolves around the players’ inconsistencies: “There may be endless repetitions in aural cultures, but there can never be loops because, as long as human agency is involved, the same thing is always going to be different.” (Cutler 2018, p68) My system, through its unreliability and fallibility, creates loops which are consistently variable, creating a groove with more character than an unchanging repetition.
Cutler, Chris (2018). Loops, Memories and Meanings, in Julien, O and Levaux, C (Eds.) Over and Over: Exploring Repetition in Popular Music. London: Bloomsbury Academic, pp 68-74.
Danielsen, Anne (2018). Time and Time Again: Repetition and Difference in Repetitive Music, in Julien, O and Levaux, C (Eds.) Over and Over: Exploring Repetition in Popular Music. London: Bloomsbury Academic, pp 37-50.
Graham Dunning 