Sound design

Everything a track does — how its trigs play, which sample engine it uses, how it is filtered, shaped, distorted and modulated — sits on six parameter pages, one per PARAMETER key. Once you can jump between them, check values at a glance and undo experiments, every other topic in this chapter gets quicker.

TRIG page settings are saved with the pattern, not with the preset. The SRC, FLTR, AMP, FX and MOD settings, together with the sample, make up the track’s preset — see Editing a preset.

Any track parameter can also change from step to step: hold a TRIG key and turn a knob to set a parameter lock.

The TRIG page decides how a track’s trigs play rather than how they sound: pitch, velocity, length and the chance that a trig fires at all. Page 2 adds retrigs — fast repeats for rolls and stutters — and portamento. The values act as defaults for every note trig on the track; hold a TRIG key and change one to give that step its own setting.

TRIG page 2 is the retrig and portamento page. A retrig repeats a note trig at the chosen RATE, while VFAD fades the velocity of the repeats in or out over LEN. Retrigs only work on audio tracks, and per-step retrig settings are stored in the active pattern.

The SRC (source) page is where you shape how a track plays its sample. What you find here depends on the track’s SRC machine: each machine is a sound engine that uses its own slice of the sample playback engine’s abilities, with a parameter set tailored to that job. Change the machine and the page changes with it.

Any SRC machine can go on any track, and a machine can only be swapped for another of the same kind — one sample engine for another. A track with any machine other than MIDI counts as an audio track — see Audio tracks, MIDI tracks and machines.

SRC page 1 holds the machine’s parameters. SRC page 2 shows the same parameters as graphics, next to a view of the sample’s waveform. Edit on either page with the DATA ENTRY knobs.

Oneshot is the machine a track uses unless you choose another, and the most direct: it plays the sample as recorded, from a start point for a set length, forwards or backwards, once or looping. It is the only SRC machine with its own loop point, so use it when a sound should sustain by repeating part of itself. If a loop has to follow the tempo, pick Werp, Stretch or Repitch instead; to chop a sample into pieces, pick Slice or Grid.

For a walkthrough of tuning a sample and trimming it with STRT, LEN and LOOP, see Sample playback: tune, play mode, start, length and loop.

Werp makes a sample or loop follow the tempo of your project or pattern. It cuts the audio into small time segments and plays them one after another, lined up with the tempo, so the loop keeps time as you change the BPM. What sets it apart from Stretch and Repitch is control over those segments: you choose their size and whether each one plays forwards, backwards or on a loop.

All three use BARS to know how long the sample should last, measured in bars at the current tempo.

Stretch also keeps a sample in time with your project or pattern, but goes about it differently from Werp: it chops the audio into tiny grains and fades from each grain to the next. There are no segment controls; instead you get STRT and LEN, as on Oneshot, to choose which part of the sample plays, while TUNE sets the pitch. For how it compares with Werp and Repitch, see the table in Werp machine.

Pitch snapping on TUNE works as on the other machines that have it: press and turn for semitones, hold FUNC and turn for octaves.

Repitch is the third way to make a loop follow the tempo. Rather than cutting the audio into pieces, it repitches the whole sample to match the target tempo, so the pitch moves along with the speed. Because the tempo sets the pitch, this machine has no TUNE parameter. Choose Repitch when you’re happy for the pitch to follow the tempo; choose Werp or Stretch when you want to set the pitch yourself.

The Slice machine, added in OS 1.15, chops a sample into pieces you can play one by one — a drum break, say. Unlike Grid, which simply divides a sample into equal parts, Slice lets you shape every slice: start from an even grid, with automatic transient detection if you like, then move each slice’s start, end and loop points in the slice editor, split slices and remove them. It can also spread the finished slices over trigs you’ve already placed.

SLICE picks the slice that plays and LEN how many consecutive slices play from there: with SLICE at 2 and LEN at 3 you hear from the start of slice 2 to the end of slice 4. SRC page 2 shows the waveform with the selected slices shaded.

The slice editor is where each slice gets its exact shape. It shows the waveform with the current slice highlighted and puts its controls on the DATA ENTRY knobs. Open it with EDIT SLICE POINTS, or let CREATE SLICE GRID take you there.

Grid is the quick way to chop a sample: tell it how many slices you want and it spreads them evenly over the whole sample, with no slice editor and no points to place. That suits material that already divides evenly. When slices need to land on hits that don’t fall on a regular grid, use Slice instead, where you set every slice point yourself.

GRID sets the number of slices, SLICE picks which one plays and LEN how many consecutive slices play from there. On SRC page 2, a chequered overlay on the waveform marks the slice selection.

The FLTR pages carve a track’s tone. Page 1 belongs to the track’s FLTR machine — one of a collection of filters and EQ — and its parameters change with the machine. Page 2 holds a simple base-width filter, plus settings for the filter envelope and keytracking.

There are six FLTR machines, each covered in Filter machines; Comb+ arrived in OS 1.10. The MACHINE menu (FUNC + SRC) sorts machines by category, and a machine can only be swapped for another of the same kind — one filter for another.

The base-width filter is a high-pass and a low-pass filter connected in series. BASE sets the base frequency and WIDTH how far the range reaches above it, so together they frame the band of frequencies that gets through.

BW.RT chooses whether the base-width filter sits before or after the FLTR machine’s filter. The OS 1.10 release notes list configurable routing of the base-width filter and of distortion as new; the overdrive’s position relative to the filter machine is set with OD.RT on the FX page.

FLTR page 1 belongs to the track’s filter machine. All six share the same core — an ADSR envelope (ATK, DEC, SUS, REL), a frequency control (FREQ) and an envelope depth (ENV) — and differ in the one or two controls that give each its character. Swap machines in the MACHINE menu (FUNC + SRC); The FLTR pages has the steps, plus FLTR page 2 with the base-width filter, envelope delay and keytracking.

Multi-mode changes filter type with a single knob: turn TYPE to morph from low-pass through band-pass to high-pass. RESO adds a resonant peak at the cutoff.

Lowpass 4 is a four-pole low-pass filter with a 24 dB/octave slope — twice as steep as the Legacy filter’s 12 dB/octave — so it takes out high frequencies more firmly. RESO adds a peak at the cutoff.

EQ is a parametric equaliser rather than a filter: use it to lift frequencies you want more of or cut ones you don’t. FREQ sets the centre frequency, GAIN the boost or cut around it and Q the width of the band — the higher the Q, the narrower the band.

Comb- and Comb+ are comb filters: they add metallic, pitch-tuned, resonant overtones. FREQ tunes the resonances, FDBK sets the gain of the feedback signal and LPF sets a low-pass filter inside the feedback path. The difference is the feedback: Comb- feeds back negatively for a hollow, tube-like sound, while Comb+ feeds back positively for a more string-like one. Comb+ arrived in OS 1.10.

Legacy is modelled on the filter in the original Digitakt I. TYPE switches it between low-pass and high-pass; both are 2-pole with a 12 dB/octave slope.

The AMP page controls how each hit’s loudness moves over time — the amp envelope — along with where the track sits in the stereo field and how loud it plays. The envelope runs in one of two modes, chosen with MODE: AHD (attack, hold, decay) or ADSR (attack, decay, sustain, release).

Think of HOLD = NOTE as a gate: the sound holds for exactly as long as the note is on — the trig length in the sequencer, the time you hold a TRIG key, or the time you hold a key on an external keyboard — and then decays. A fixed HOLD time gives the same shape every time, however long the note.

The FX page gives each audio track its own grit — bit reduction, overdrive and sample rate reduction — and sets how much of the track feeds the shared delay, reverb and chorus.

DEL, REV and CHR are sends: they set how much of this track reaches each effect. The effects themselves are shared by the whole pattern and edited on the SEND FX pages — see Editing send FX and mixer pages, Delay, Reverb and Chorus. For where each stage sits in the voice, see Sound architecture.

Each audio track has three LFOs (low-frequency oscillators) for adding movement: they modulate parameters on the track’s SRC, FLTR and AMP pages. MOD page 1 controls LFO 1, page 2 LFO 2 and page 3 LFO 3. All three pages have the same eight parameters, so once you know one you know them all. MIDI tracks have two LFOs.

LFO speed in steps. With MULT set to a BPM value, the table below shows how many sequencer steps one full LFO cycle takes for each SPD and MULT combination. A value of 8 means one cycle every eight steps; 1/2 means two cycles per step.

Each LFO’s DEST parameter on the MOD pages decides what it moves. The list is grouped into categories — META, LFO1, LFO2, SRC, FILTER, AMP and FX on audio tracks — and holding FUNC while you turn DEST snaps from one category to the next. MIDI tracks have a shorter list of their own. The tables give each destination as the list names it, next to the parameter it moves.

MIDI track parameters are explained in The MIDI machine: sequencing external gear; the LFO settings themselves are in The MOD pages (LFOs).