Open-source sequencer and sampler — by warft_ctrl
Say it like turn. As in: turn the knobs. Turn a doodle into a beat. Turn “I can’t play piano” into “wait, that slap though.”
TŒRN is a sequenced musical device for beginners, makers, and anyone who likes blinking lights more than plugin menus. Think Etch-A-Sketch™ energy, but instead of drawing a wobbly staircase you draw drums, synths, and happy accidents — on a vivid 32×16 RGB LED panel (standard; 16×16 optional) that is both the screen and the keyboard.
It sits on a Teensy 4.1, runs from a Micro SD card full of your WAVs, and is happy on a desk, in a backpack, or on battery at the park where strangers will ask “what is that?”
Status: under very active development. Things move fast. Stay tuned, and don’t be surprised if yesterday’s weird bug is today’s charming feature (or the other way around).
You get:
- 8 sample voices for drums, one-shots, loops, and “that weird recording from the metro”
- 1 three-voice polyphonic synth with presets (bass, keys, chiptune, pads… the usual suspects)
- 2 monophonic synths with LFO and arpeggiator for leads that refuse to sit still
- A 32-step × 8-page sequencer (256 steps of room to overthink; 16×16 mode uses 16 steps × 16 pages), page chaining into songs, MIDI in/out, filters, bitcrush, reverb, and enough real-time knobs to keep your hands busy and your laptop jealous
No laptop required once it’s built. Samples live on the SD card. Patterns autosave. You twist, tap, and suddenly it’s 2 a.m.
For the friendly guided tour (first beat in about a minute), open the handbook in this repo.
TŒRN builds on the legacy of the NI404, with more features, fewer cables dangling off the side, and a custom PCB where everything mounts cleanly. The older “extra boards and hope” era is over.
Meet the glow-up:
- RGB-I2C rotary encoders — smoother, prettier, and they light up like they mean it
- Custom PCB — codec, charging, and jacks on one board
- USB-C — because micro-USB was a character-building experience we no longer need
- LiPo charging port — optional battery for true “make beats on a train” energy (battery not included; we trust you with electricity)
- Proper audio I/O — 6.35 mm headphone out, line in/out, mic in, plus a built-in mic for “sample the room right now” moments
- TRS MIDI in/out — play nice with the rest of your setup
- On/off switch — underrated luxury
- Touch buttons — for the things knobs shouldn’t monopolize
- Laser-cut acrylic case — files included; looks finished, not “breadboard cosplay”
- JST connectors — expansion leads plug in
Open source as always: schematics, code, and design files are here for you to poke, fork, and improve.
- 11 instrument voices on a 16-row grid: 8 sample voices (1–8), one 3-voice poly synth (11), two mono synths (13–14)
- Standard 32 steps per page × 8 pages → 256 steps per pattern; optional 16×16 mode uses 16 steps × 16 pages
- Voice mode: each voice keeps its own pattern length. It loops the pages that hold its notes, pitches included, while the other voices keep looping on their own
- FIRE (ETC → ADV menu): optional playhead sparks when a chosen voice’s notes light white
- Store up to 999 patterns, 999 samples, 999 samplepacks
- Autosave / autoload so your 3 a.m. genius doesn’t vanish at boot
- Switch patterns live without stopping playback
- Song mode: chain up to 64 page assignments from the current pattern
- Pattern ops: copy/paste, transpose page/channel (±1 octave), smart random notes (major/minor + rhythm ideas), clear page/channel
- SWAP exchanges notes between voices, with octave offsets; optional SUBP file subpatterns extend arrangements
- Up to ~12 seconds per voice at 44.1 kHz
- Browse and load samples while playing (yes, really)
- Per-voice start/end trim, forward/reverse
- TIME stretches or shortens samples independently of pitch
- Samplepacks (including pack 0 = custom per-voice assignment)
- Record from built-in mic / line-in / mic-in straight in the sample browser
- Save or merge used samples back to the SD card as a pack
Channel 11 — poly synth
- 3 oscillators per voice × 3 voices
- Presets: BASS, KEYS, CHPT, PAD, WOW, ORG, FLT, LEAD, ARP, BRSS
- Per-voice ADSR, filter envelope, pan, volume, pitch (cents & semitones), waveform
Channels 13–14 — mono synths
- 2 oscillators each
- ADSR, LFO on cutoff/resonance, MIDI arpeggiator (step/span; does not pitch-shift grid notes)
- Waveforms: sine, saw, square, triangle
- ADSR — attack / decay / sustain / release (0–32 range)
- Bitcrusher — bit depth 1–16 (0 = bypass), sample-rate reduction down to glorious crunch
- Filter — LP / HP / BP, 0–10 kHz, resonance 0.7–5.0, smooth transitions
- Reverb — room size, damping, wet/dry; denser six-comb/four-diffuser network on reverb-equipped voices
- Fast filter access on encoder #3 during runtime (assignable default per channel)
- TRS MIDI in/out, clock send/receive, notes, transport
- BPM 40–300, internal clock, external MIDI clock, tap tempo via touch
- Live recording / looping (touch-hold), up to ~12 s per voice
- Note probability, velocity, condition triggers (those “only every other bar” tricks)
- Global transpose, per-channel detune & octave, fine cents on the poly synth
- Mute / solo, pattern modes (OFF, ON, SONG, NEXT)
- Stereo routing options
- USB Audio + Serial: send the mix to a computer while using the SD tool; musical MIDI uses the TRS ports
- Analog clock out (J25), clock in (J26) and voice-triggered sidechain out (J27)
- TOUCHY performance pads, live fills, and per-voice audition
- Browser tools for firmware flashing, SD transfers, sample conversion and the simulator
- Demo animation: open ETC → INFO, press touch3 / BUTTON_C; press and release Touch 1 / BUTTON_A to return to INFO
| Brain | Teensy 4.1, PSRAM fitted. Audio codec is the SGTL5000 on the carrier |
| Memory | 16 MB PSRAM target; Micro SD (handbook suggests up to 64 GB, SanDisk Ultra) |
| Audio | 44.1 kHz, 16-bit mono WAV |
| Controls | 4× RGB illuminated I2C rotary-push encoders, 5 capacitive touch buttons |
| Display | Standard 32×16 RGB LED panel, 512 LEDs (FastLED); 16×16 optional |
| I/O | 6.35 mm headphones, line in/out, mic in; built-in mic; TRS MIDI in/out |
| Power | USB-C or optional LiPo + charging |
| Path | Contents |
|---|---|
platformio/ |
Arduino sketch tabs, headers, custom audio in src/, build scripts and platformio.ini |
hardware/PCB/ |
KiCad boards, schematics, manufacturing files and purchasing BOMs |
hardware/housing/ |
Fusion archives, laser DXFs and panel previews |
docs/devdocs/ |
Developer documentation → docs.toern.live |
docs/handbook/ |
Operator handbook → handbook.toern.live |
docs/BUILD.md, docs/info.md |
Hardware assembly and firmware build instructions |
tools/firmware/ |
Browser firmware flasher (GPL-3.0) |
tools/sdtool/ |
Browser SD tool |
tools/sim/ |
Browser simulator |
tools/audioconvert/ |
Standalone sample converter |
tools/case/ |
Standalone case generator; case.toern.live planned, not deployed |
tools/colorsheme/, tools/_shared/ |
Color editor and shared tool assets |
_debug/ |
Host tests, analysis scripts and local diagnostic captures |
_private/ |
Local business data/helpers; annotation tool in _private/helpers/annotate/ |
website/ |
Public project website only; local/deploy-only |
See the complete directory map for local-only items marked ***.
Build from the repository root with pio run -d platformio, or cd platformio first.
For a personal hardware build, follow the assembly and laser-cutting guide. Start with the BOM directory: check assembly-part stock before ordering the PCB. The pre-order list lists parts to secure for JLCPCB assembly. Then buy the Teensy, PSRAM, encoders, matrix, cables and case hardware separately. The website’s Build chapter covers the same assembly order.
This project leans on several libraries. Some are stock; some got lovingly (and necessary-ly) poked.
-
ResamplingReader.h (
teensy-variable-playback) Variable rates, interpolation, multi-channel, loop types (repeat / ping-pong), crossfading — inplatformio/src/resamplerReader.h -
Treeverb (
effect_treeverb) Full-rate reverb with balanced 6/4 diffusion in RAM1 —platformio/src/effect_treeverb.*; the oldereffect_freeverb_dmabufremains a host-test reference. -
FastLED
FASTLED_ALLOW_INTERRUPTS 0so LEDs and audio don’t argue mid-beat; customFastLEDshow()/FastLEDclear()wrappers -
MIDI Bigger Serial8 buffers for dense clock traffic (
RX 2048,TX 128) and custom TRS MIDI settings
- TeensyPolyphony by Nic Newdigate — polyphonic heart of the thing
- Audio Library (PJRC)
- WS2812Serial, i2cEncoderLibV2, FastTouch
Order a TŒRN at toern.live. Questions go to info@toern.live.
When it arrives, open the handbook and start at First beat. Power is USB-C. A battery and a microSD card are not in the parcel; the buy page and the handbook hardware chapter name what to get.
Firmware changes: info.md to flash, docs/devdocs/ for how the code is shaped, firmware-edits for a new menu page. Fabricating a board from the design files is BUILD.md — that is not how a TŒRN is ordered.
- Software / code: MIT — free for personal and commercial use.
- Firmware loader (
tools/firmware/): GPL-3.0, because that tool is Teensy Loader Javascript. GPL-3.0 applies only to that folder. - Hardware design files (schematics, PCB layouts, Gerbers, housings, etc.): CC BY-NC 4.0 — personal / non-commercial use and modification. Jan-Peter Kuntoff is the only person authorized to sell TŒRN hardware (bare PCB, assembled PCB, kit, or unit), including boards sold without a Teensy. Teensy 4.1 is PJRC’s; it is not licensed in this repo. Anyone else needs written consent to sell.
Commercial hardware licensing: Jan aka warft_ctrl — jpkuntoff@gmail.com (Hamburg, Germany)
Huge thanks to Paul Stoffregen and the PJRC crew for Teensy, to the wider open-source audio community, and an especially loud shoutout to Nic Newdigate for teensy-polyphony — basically the soul of this project.
Made with too many late nights and exactly the right number of RGB LEDs.
Jan aka warft_ctrl (formerly soundpauli) Hamburg, 2026