Skip to main content

The paradigm

If you understand five ideas, you can predict how almost everything in Lockstep behaves.

2.1 Scope + verb​

Every editing and performance action is built from two halves:

  • A scope key declares what you're about to operate on. Scopes are held modifiers: the eight-key left cluster (Func, Track, Phrase, Scene, Morph, Song, Mute, Fill — 7.9 naming), plus a held step (Trig) and a section key. Two modifiers — one from each column — can be held together to combine scopes (the compound chord). (Cue is the audition scope, entered as the compound Func+3: hold it to pre-listen the focused track, Cue+step to audition a step — writes nothing.)
  • A verb key declares what to do. The verb set is small and fixed (Y U I O P): Snapshot (push checkpoint), Record (= copy), Play (= paste), Clear, Yes — the Func layer gives Restore / Panic / Delete / No — plus turning an encoder for live value tweaks.

You hold the scope, then press the verb. For example:

  • Hold a step + Record → copy that step.
  • Hold a track + Record → copy that whole track.
  • Hold a section + Record → copy that section's parameters.

The verb meant "copy" in all three; only the scope changed. Once you internalise this, new features stop being new shortcuts to memorise — they are old verbs applied to new scopes.

2.2 Override-ELSE-Base​

Every value that can vary per step follows exactly one rule:

Effective value = step override (if present) ELSE track base

There is no "reset" value to manage and no precedence flags. A value is either overridden on this step, or it falls through to the track's base setting. This single rule governs:

  • Machine parameters (P-Locks): the per-step override of a sound engine's controls.
  • Trig fields: note, velocity, gate length, and trig condition can each be overridden per step or left to the track default.

This is why "holding a step and tweaking" always means the same thing — it writes a step override. Tweaking with no step held writes the track base.

2.3 The edit context (held step = step scope)​

Whether your edit lands on a step or on the track depends on one piece of state: is a step currently held?

  • No step held → parameter writes go to the track base.
  • A step held → the same writes go to that step's override (a P-Lock, or a per-step trig override, depending on which section is active).

This is true regardless of the source of the edit — a QWERTY key, a MIDI CC from a hardware knob, or the on-screen encoder all resolve through the same gate. Holding a step and turning a knob P-Locks it; holding a step and playing a note records that note onto the step.

2.4 The container hierarchy​

Set
 └── Song ×16
      ├── per Track: Kit (the sound) + Phrase ×16 (note content)
      └── Scene ×16 (a launchable row: each track's phrase + active-mask + core time)
  • Set — one plugin instance. Owns all Songs, the sample pool, MIDI mappings, and global settings. (Matches an Ableton "Set".)
  • Song — a self-contained song: per-track Kits and Phrase pools plus a set of Scenes. The bank-sized unit; switching Songs is a full performer reset (clears live deviations).
  • Phrase — pure per-track musical content: the trig grid, per-step overrides, P-Locks, length/divider, trig defaults, conditions. No sound. Each track has a pool of 16.
  • Kit — the per-(track, Song) sound: which engine the track hosts, its base parameters, post-machine FILTER/AMP, and sample references. Chosen via Track + hold(SRC); the sound itself is the operand of the Machine scope (Func+Track). (The dissolved Octatrack "Part".)
  • Scene — a launchable cross-track row: the global phrase index (which row all tracks default to), the active-mask (who sounds), core time, and the Morph snapshot. Per-track phrase deviations are live/RAM-only — never saved. Launched live (Scene+step), not chained into a written arrangement.

Tracks are polymetric: each has its own length (1–64 steps) and musical subdivision (1/64 to 4/1, straight / dotted / triplet), so a 7-step track and a 16-step track phase against each other naturally with no master-bar concept.

2.5 The sound path: what the names mean and where they sit​

The container hierarchy above says where musical content lives. This section is the other axis — what a track is, and the order sound flows through it. Every name here recurs throughout the manual:

Phrase ──trigs──▶ Machine ──▶ Foundation (FILTER → AMP) ──▶ Inserts ×2 ─┬─ (×sendA) → Send bus A
                                                                           ├─ (×sendB) → Send bus B
                                                                           └────────────────────▶ track sum
                                    Send bus A ──▶ Send Return A ─┐
                                    Send bus B ──▶ Send Return B ──┤
                                    track sum ────────────────────▶ Σ ──▶ Master Insert 1 ──▶ 2 ──▶ output gain
  • Machine — the sound engine a track hosts (sampler, FM, Analog, drum synth, MIDI-out…). Each machine declares its own parameters; the section keys (5–0) page through them. The machine is part of the Kit, so swapping Kits swaps machines.
  • Foundation — the post-machine FILTER and AMP blocks that the sequencer owns, identical on every track regardless of machine. This is what "track-foundation row" (Track + section) refers to: a guaranteed filter and amplitude envelope that exist even when the machine has none of its own. A machine's own filter, if it has one, still lives behind the canonical FILTER key — the foundation filter is in addition, downstream.
  • Scope overlay / colour — section params resolve through an ordered layer stack (machine → track → …). Unqualified, each key shows the machine's page, falling through to the track page where the machine owns nothing; hold Track to peel the machine layer and see the track pages only (machine-only sections go dim). Section keys and the MZ are coloured by the scope a page comes from — machine pages read neutral, track pages read cyan — so you always know which layer you are editing.
  • Inserts — two per-track IEffect slots after the foundation. Loaded via the hold FX picker (or Song + hold FX for master), edited on the FX section, momentarily bypassed with the Animate gesture (FX + step). MIDI-out tracks have none (no audio).
  • Send buses — post-insert taps from each track. Set Send A / B levels on AMP page 2. Send returns are processed before master inserts.
  • Master FX — four Song-scope units (2 inserts + 2 send returns) over the summed output. Loaded via Func+Song+FX; cycle with repeated press (FX1→FX2→Snd A→Snd B). Master-only effects are compacted out of the track picker (they simply don't appear — no greyed gaps), and the External send only appears on the two send units. When a catalogue spans more than one page of 16 cells, the picker pages with Nav ←/→ (a P1/2 indicator shows in the picker's status line).
  • Output routing & buses — each track's finished signal has one destination, the Out slot in the CHANNEL block: Master (default), Track N, Aux N, or Off. Out is single-destination by construction (out-degree ≤ 1) — choosing a destination replaces the previous one; a track is never on Master and an aux at once. Routing a track to Track N removes it from the master sum and feeds it into track N; if N hosts a Route machine it becomes an aux/sub-bus that reads the sum of everything routed into it, processes it through its own FILTER/AMP/FX, and sends that onward. Aux N sends the track to one of the plugin's host Aux output buses instead of the master sum — the DAW mixes it (the exclusive tap is deliberate: the host, not Lockstep, decides how the aux returns). A track on an Aux whose host bus is disabled (e.g. standalone, or the DAW left the bus unconnected) folds back to Master so no audio is lost. Ordering is solved automatically (per-block topological sort); a routing that would form a feedback loop is refused. This is how you build drum buses, parallel chains, and resampling. The Out rotary only steps through valid destinations (Off / Master / current buses) and shows the target live as you turn — you never jog through unusable tracks. On the track/VU row, a bus and its feeders share a colour so groups read at a glance. An audio track's button fills with its output VU; a MIDI-out track instead fills with a magenta velocity meter (note-ons add a velocity-proportional loudness that decays like audio), and its top-right dot flashes on CC activity (the left cyan dot is the trig pulse). Solo is routing-aware: soloing a bus keeps its feeders audible (you hear what flows in), and soloing a feeder keeps its downstream bus chain audible (so it still reaches master).
  • Audio input — a machine can consume audio instead of synthesising it, via an input_source tap: None, Ext1–Ext4 (four stereo plugin/device input buses), or Master (the prior block's master sum, for whole-mix resampling). Only the first input bus (In) is enabled by default; enable In 2–In 4 in the host to feed Ext2–Ext4. The Route machine is the pure router — it adds nothing, just passes its input (an external bus, or the bus sum routed to it) through the canonical FILTER/AMP/FX. A fresh Route defaults to None, so it is a silent sub-bus until you route audio in or pick a source. The source rotary omits any selection that would feed back (e.g. Master on a track whose own output reaches master, or a cyclic Track tap), exactly as the Out routing rotary omits cyclic destinations — and a stale selection loaded from disk is muted at run time as a second safety net. The Out rotary also offers only destinations that accept inbound audio. A track qualifies as a destination when its machine exposes an input source (and is not MIDI-out) — Route, Loop, Tape, Record and the other capture machines do; an ordinary synth track does not, and so never appears in another track's Out. That is why grouping a kit onto a sub-bus begins by giving the bus a Route machine: you are not choosing a mixer channel, you are pointing at something built to be fed. Off and Master are always offered; the Aux buses appear only when the host has enabled them, so in standalone there are none.

And the state that feeds this path is layered, finest layer winning (this is "more specific scope wins" applied to values):

  • Base — the Kit's stored value for a parameter. What you edit with no step held.
  • Morph — a per-Scene A/B pair of sparse parameter maps blended by the crossfader, sitting over the base. Continuous, performed.
  • P-Lock / trig override — a per-step override, sitting over both. Discrete, authored. A step that has one ignores Morph and base for that parameter.

So: a Phrase decides when notes happen, the Kit (machine + foundation + inserts) decides what they sound like, Morph bends the sound continuously across a Scene, and P-Locks pin exact values to exact steps. The Scene chooses which Phrase each track plays and who is audible; the Song holds it all; the Set is the plugin.