Torso Electronics S-4

Torso’s ambitious ‘sculpting sampler’ may be a work in progress, but it promises to be something very special indeed.

Torso Electronics are a young Danish company with a preference for user interfaces that minimise the gap between ideas and realisation. Their first product, the T‑1 sequencer, has the hands‑on directness of a performance instrument and for their second, the S‑4, a similar philosophy is applied to audio capture and processing.

The S‑4 is described, rather mysteriously, as a ‘sculpting sampler’. Our mission today is to discover what that means — and then try to decide if it’s a creation worthy of Michelangelo, or the product of just another bloke with a chisel...

So What Is It?

Listing the current functionality feels like a safe enough way to kick off. S‑4 workflow is based upon four stereo tracks, each of which can act as a virtual tape recorder, a polyphonic sample player or a gateway for incoming audio. The tracks are then subjected to a series of ‘sculpting’ options — currently a granular effect, a morphing resonator, several flavours of distortion and a combined reverb and delay. Additionally, the S‑4 can act as a USB audio interface — one that’s capable of expanding its I/O automatically whenever a class‑compliant device with extra channels is attached. Lastly, for instant results in any environment, it features a built‑in microphone.

Onboard storage (4GB) is provided for samples and projects, plus there’s a USB‑C connector to enable file transfer from a PC, Mac or peripherals such as external drives. Arguably, the only omission is an SD card slot, or a similar means of boosting the storage unobtrusively, because, if my experiments are anything to go by, that 4GB will fill up pretty darn quickly.

Internally the S‑4 sports a — now ubiquitous — Raspberry Pi with a quad‑core 1.5GHz processor. Audio is a clean 24‑bit, 48kHz throughout and, while the stereo audio inputs and outputs are on quarter‑inch jacks, all the other connections are 3.5mm, including the headphone socket, MIDI in and out, plus the analogue clock I/O. A single adaptor (Type A) is supplied to turn mini‑jack MIDI into the regular five‑pin type and the power adaptor ships with a selection of country options; you can pack them away in readiness for those sunlit days when touring is viable once more.

On the rear panel we find 3.5mm sockets for headphones and MIDI and sync I/O, quarter‑inch sockets for audio I/O, a USB‑C port and power supply input.

As hardware devices go, it’s a petite package measuring a mere 156 x 242mm and standing just 39mm from its rubbery feet to the tips of its encoders. The RGB buttons are chunky‑feeling and those encoders map to objects on the 3.5‑inch pin‑sharp LCD in ways Elektron users will instantly recognise. Admittedly, the encoders don’t handle quite as nicely as those of an Elektron but my main misgiving is their black on black aesthetic — this design choice feels needlessly obtuse, at least in my darkened studio. That said, my fingers eventually learned their way around without troubling my brain too much.

Sculpting Time

On power‑up you’re greeted by an empty project and so immediately have choices to make — such as what role(s) the S‑4 should perform today. Hopefully in some future update there will be an option to start from your own defaults, or from where you left off previously, but for now you must either load a project manually or begin with Torso’s preferred settings. Incidentally, stored projects appear in a simple list, sorted alphabetically — and it’s not hard to predict this will soon become long and unwieldy.

Beneath the encoders are buttons representing the chain of audio devices available. From left to right, these are: Material, Granular, Filter, Color and Space. All can be bypassed if necessary but, at present, Material is the only one to offer multiple modes — Tape or Poly. When Material is bypassed, that track is ready to receive incoming audio.

The S‑4 boots as a virtual four‑track tape machine with an interface that’s so friendly it invites you to begin recording and overdubbing right away. However, you quickly realise there’s no indication of how long you’ve been going — or how much time you’ve got left. Further investigation revealed that a maximum of six minutes is allocated per track, but some kind of time or bar counter visible during recording would be very much appreciated.

Recordings are based on a global tempo value — either sourced internally or taken from your selected input (MIDI or analogue). Synchronisation arrived quite late (version 1.1) with further improvements and hotfixes following in 1.2.x. It’s perfectly straightforward to adjust the start and end of each reel on the fly, shift the audio relative to other tracks or chop to precise bar intervals when a perfectly sync’ed loop is required. That said, I’d love to see a zoom function implemented so you could more easily fine‑tune those starts...