Google Dubtechno Now

Tug of War: Digital Timestretching

29th of January, 2024

Sometimes, given problems are simple or discrete enough that we can think of a sort of ‘canonical’ approach to solving them. Since a game of chess is discrete—making a move involves choosing between a relatively small number of possibilities—we can come up with some strategies which are simple enough that we should expect anyone who studies the problem to come up with that same idea, like making a move at random or moving pieces according to a set, deterministic order. Making music isn’t discrete in this way so there are no canonical solutions to making music, but within limited domains there are canonical approaches to very specific sorts of processes within music. For instance, in speeding up and slowing down recordings, I venture that the canonical approach is to multiply the waveform by some value. Making a recording twice as fast using this approach means taking the whole waveform and multiplying its frequency by two, hence also doubling the pitch.

This is a technological emulation of speeding or slowing the flow of time, which of course there is some precedent for in a relativistic universe even if the phenomenon of sound that has undergone relativistic time-stretching effects has so far remained unheard by any human ear. Turntablism as an aesthetic depends on playing with and converging upon this phenomenon. The rotation speed of vinyl records is modified (often rapidly) by a human playing the turntable as an instrument, which means affecting the rhythm, pitch, and timbre of a recording all at once. Never before heard sounds like the scratched warble and cluck of the articulated human voice helped to characterise the early texture of hip hop and breakbeat: this was a pop musique concrète which sought after new transformations of sound for the sheer joy of hearing something fresh and impossible.

But what happens when the methodology of turntablism meets the eccentricities of digital audio manipulation? Speeding and slowing via the canonical solution means modifying pitch, but with the advent of digital samplers like the Akai MPC producers were free to stretch their recordings without modifying pitch, with the trade-off that instead they were modifying their texture. When hardcore producers sped up the breakbeats of the 80s and produced their 160 to 170 bpm darkside and jungle breaks there was only minimal loss of fidelity, but slowing a sample meant shredding it into tiny intervals and interpolating it with copies of itself. Vocal samples and short snippets of percussion were still legible as such, but their stretching meant articulating them with a screeching metallic resonance.

“I and I know that all of the youth will witness the day that Babylon shall fall.”

At some point in the 90s the specific timbre of these stretch-marks were somewhat standardised. Producers gravitated towards using similar pieces of equipment because they were accessible: either in that they were cheap or in that many people had access to a bit of kit via a shared recording studio. But with the advent of cheap computers and digital audio workstations, the techniques by which software shredded these samples, filled in the gaps, and put them back together, now depended heavily on the individual eccentricities and peculiarities of a set of bespoke algorithms. Producers looked for tricks and techniques for drawing out new textures from their software that could never be realistically repeated: the process of fluctuating speed was now firmly divorced from a public canon. Though artists were probably about as willing to share and collaborate as ever, their tools were distributed as software packages which were in direct competition with each other as digital commodities.

The multiplication of possible approaches had an individualising effect in the music as a result, as any given track would have slightly different (unique and unrepeatable) production artefacts. No one else could hope to produce the same sort of malfunctioning speech synthesis that babbled unnervingly at the beginning of Aphex Twin’s Afx237 v.7, nor the enormous scream which collapses into the dying gasps of a faulty generator at the beginning of The Chemical Brothers’ We Are the Night. Nor for that matter can the groaning bassline from Noisia’s Monster be understood without first hand experience of the many iterations of contortion and resampling that went into its creation. Dave Tipper is widely seen as a virtuoso of sound design, but that is not necessarily relevant to the fact that the sequence of reality bending folds, inflations, and strangulations that went into programming Tug of War can never be sufficiently approximated by anyone else.

Much of Tipper’s ‘00s output smells like IDM and downtempo, and most of his ‘10s work is characterised by the same kind of meticulous digital sound design shenanigans as the much younger cohort of bedroom producers making brostep and glitch hop at the time. Anyone who only knows him for this later work might find the fact that he was steeped in turntablism throughout his formative years in the late 80s and early 90s surprising. But there are certain standout moments that make this influence so obvious that every fibre of his idiosyncratic being feels like a love letter to those values and conventions. Tug of War is the filthiest, horniest love letter of the lot. Massive compressed kick drums, snares, and shakers feel clean as a whistle despite their distortion and crunchiness, but they scaffold a stumbling bassline which combines a long swelling motion and a rhythm of continuous deceleration to produce a groove of simultaneous flux and rigidity.

In the interest of compelling structure I’ve described this track out of order: the bassline is wonderful, but Tug of War introduces its fanciest trick in the first five seconds. A delayed crack of a hi-hat begins to ping around, faster and faster, until it produces a high pitched periodic note. The note is held for a tiny fraction of a second before tumbling back into its rhythmic mode just in time for a set of distorted, time-stretched belches to tumble out over the mixture. Tone freely folds into rhythm and folds back into tone as the tempo fluctuates wildly around a single implied groove. This is the texture of a digital turntablism: where tempo and pitch make a mockery of acoustic music and produce a time that accelerates into spirals—terminating themselves as they coil up into nothing or straighten out into lines—without ever compromising that heartbeat pulse that frames the whole thing; that makes it danceable.

When geniuses of the new artform of fucking about with records first developed this technique, the textures they created were informed by the records they chose to manipulate. Turntablism is first and foremost a logic of sound development: a set of processes of developing sound over time. But the premises of that logic had to be supplied from the outside, which meant taking other people’s funk and disco records and applying the revolutionary processes to their reproduction. They could never shake off the element of indeterminacy that came with folding someone else’s grooves into their revolution. Why would they want to? The digital development introduces an even further indeterminacy, since the syllogisms themselves now depended on unaccountable machinery, produced by tech nerds with brains but no hearts. When Tipper sends the repeating period of those hats past the ability of the ear to parse them and into quantum uncertainty, we must not be under any illusion about what the resulting whine depends on. It depends on an algorithm that some engineer wrote to determine how to handle the reality breaking phenomenon of a virtual drummer hitting a drum thousands of times a second.

Further Reading:

Forty-Four UK Dance Tunes From the Noughties: Part 1
(Slip & Slide) Suicide: Vocal Anonymity in Dance Music
Hyperoptics: Lessons From Cinematic Vision