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 16×16 RGB LED matrix 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 16-step × 16-page sequencer (256 steps of room to overthink), pattern 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.
- 16 channels in spirit: 8 sample voices (1–8), one 3-voice poly synth (11), two mono synths (13–14)
- 16 steps per page × 16 pages → 256 steps per song arrangement space
- 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 menu): optional playhead sparks when a chosen voice’s notes light white
- Store up to 999 patterns, 999 samples, 100 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 patterns into something resembling a composition
- Pattern ops: copy/paste, transpose page/channel (±1 octave), smart random notes (major/minor + rhythm ideas), clear page/channel
- 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
- 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 (rate/depth), arpeggiator (step/span)
- 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
- 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, optional external WS2812 strip for ripple eye-candy synced to the music
| 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, 3 touch switches |
| Display | 16×16 RGB LED matrix (FastLED); up to 2 modules → 32×16 |
| LED strip | External 5 V WS2812 connector (configurable length) |
| 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 |
TŒRN does not mount the Micro SD as USB mass storage. That would be too easy, and also fight the audio engine. Instead:
While Menu → ETC → SD is open (screen: WAIT → OK), you can browse and transfer files over USB Serial.
- On the device: open Menu → ETC → SD
- On a computer: open https://sdtool.tyng.app in desktop Chrome or Edge
- Click Connect, pick the Teensy serial port
- Stay on ETC → SD for the whole session
WAV uploads are converted in the browser to 44.1 kHz mono 16-bit. Drag folders, multi-file drops, and zip archives are supported.
Any USB Type that includes Serial (e.g. Serial + MIDI):
cd standalone-tools/sd-tool-standalone
pip install pyserial
python3 toern_sd.py -p /dev/cu.usbmodemXXXX list /
# also: put / rm / mkdir / get
python3 toern_sd.py web # optional local UI at http://127.0.0.1:8787Insert or eject the SD card only with the unit off — the firmware re-indexes on boot.
| Path | What’s there |
|---|---|
toern.ino + toern_*.ino |
Firmware (the fun / terrifying part) |
src/ |
Modified audio bits (resampler, freeverb DMA) |
BUILD.md |
Order the rev H board, what to solder, first power-up |
info.md |
Compile and flash (PlatformIO) |
PCB/toern_revH/ |
Current KiCad board, gerbers, JLCPCB BOM |
handbook/ |
Human-friendly operator’s guide (how to use the device) |
devdocs/ |
Docusaurus code and hardware (rev H) docs → published at /docs/ |
tools/ / standalone-tools/ |
Helpers; SD tool, sample converter, browser sim → tyng.app |
standalone-tools/web-standalone/ |
Browser simulator → sim.tyng.app |
standalone-tools/teensyloader-standalone/ |
Firmware flasher → firmware.tyng.app (GPL-3.0) |
website/ |
Project site (local/deploy-only; handbook + docs copied in manually) |
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 — insrc/resamplerReader.h -
Freeverb (
effect_freeverb_dmabuf)
DMAMEM-friendly reverb for Teensy 4.1 —src/effect_freeverb_dmabuf.* -
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, 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 (
standalone-tools/teensyloader-standalone/andtools/teensyloader/): GPL-3.0, because that tool is Teensy Loader Javascript. GPL-3.0 applies only to those folders. - Hardware design files (schematics, PCB layouts, Gerbers, etc.): CC BY-NC 4.0 — personal / non-commercial use and modification. Commercial hardware use needs written consent.
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