TŒRN — Samples

Two layers: pack & wave

A pack (covered in Sample packs) decides which numbered folder of samples the eight sample voices read from. The sample browser covered here — opened via WAVE in the menu or with a both-touch (left-first) from the grid — picks one .wav file under samples/ for the current sample voice (channels 1–8), with trim, reverse, and load steps before the audio actually replaces what the voice plays.

Up to 999 patterns, 999 samples, and 100 sample packs are addressable. Samples are mono 16-bit WAV at 44.1 kHz, up to 12 seconds per voice.

Sample browser

Open from the menu (WAVE) for sample voices 1–8, or with a both-touch left-first gesture from draw / single / menu when the current channel is a sample voice. The screen shows four large hint glyphs on row 1 — P (magenta) preview, Y trim end, G load, W browser — plus a waveform display.

Browsing

Encoder 4 turns through a single combined list: parent link [../], subfolders, then .wav files in the current directory under samples/. The matrix changes display per row type:

On a .wav: peak / waveform view between the trim start and trim end.
On a directory or [../]: breadcrumb path on row 6 (green dot for root, yellow chevrons, then one letter per folder segment).

File row selected: peaks + trim bar on y=3.

Folder row selected: breadcrumb path replaces the peaks.

Folder navigation

Encoder 4 short behaves contextually:

On [../] or a folder → enter that folder.
On a file → go up one directory (deliberate no-op at the samples root, so you can't accidentally exit).

Trim window

Trim is expressed as 0–100% of the file's decoded length.

Encoder 1 turn — moves the trim start. Stops just before the trim end.
Encoder 2 turn — moves the trim end. Stops just after the trim start.
Row 3 of the matrix shows a dim blue baseline; columns left of the trim start tint green (pre-roll), columns right of the trim end tint red (post-roll). The lit window in between is your active slice.

Preview & reverse

Encoder 1 short or long → forces a manual preview from RAM regardless of auto-preview state.
Encoder 2 short — reverses the highlighted sample. The waveform on screen flips end-for-end and the preview re-triggers from the start so you can hear the reversed take. Hit it again to flip back.
Auto-preview policy is set under PLAY → PREV: ON auditions every selection automatically, PRSS waits for an explicit gesture.

Loading

A short press on encoder 3 loads the highlighted file into the voice and immediately autosaves the pattern: the current trim window plus reverse state become the voice's permanent sample until you load another one. Encoder 3 turn is intentionally unused in the sample browser.

Sample browser — encoder map

1 Encoder 1

Turn	Trim start (%)	Restarts auto preview when on.
Short / long	Manual preview	Audition the highlighted file without loading it.

2 Encoder 2

Turn	Trim end (%)	Stops just after the trim start.
Short	Reverse preview	Mirrors RAM buffer + trim sliders.

3 Encoder 3

Short	Load	Commit the trimmed sample to the voice and autosave.
Turn	—	Unused.

4 Encoder 4

Turn	Browse list	Combined parent / folder / file list.
Short	Folder logic	Enter folder OR go up; no-op at samples root.

Recording new samples

From the sample browser, hold touch 3 to enter record mode for the current sample voice. Recordings are written as numbered files such as rec_01.wav inside the current browse folder.

Idle: threshold column, live level column, RDY text, Y/R/G/X hints.

1 Encoder 1

Turn	Threshold	0–100; column on x=1 visualises the bar.
Short	Arm threshold	Enables auto-start on level cross.

2 Encoder 2

Short	Start recording	Manual fire, pairs with threshold arming.

3 Encoder 3

Short	Play / stop review	SD preview of last take; also acts as global stop while recording.

4 Encoder 4

Short	Exit to sample browser	Stops recording if needed, drops onto the new file for trim/load.

Threshold-armed recordings auto-stop after ~1 s if the input falls under half the threshold. Maximum take length is 12 seconds per voice. The REC submenu in the menu chooses input source (MIC / LINE), mic and line gains, fast-rec trigger style, and what happens to existing notes on the channel after a successful recording.

How the SD card is organised

The card has a simple, opinionated layout. You can edit it from a computer when the unit is off — re-scan with ETC → RSET → SD after power-up.

samples/ — root of the sample tree. Subfolders become packs and browse folders.
samples/<pack-N>/map.txt — optional manifest assigning files to voices.
patterns/NNN.dat — pattern slot files saved from FILE.
patterns/NNN.txt — optional sidecar text loaded with encoder 3 toggle in FILE.
palette/ — optional custom colour files for ETC → COLR.

Naming: keep folder names short and ASCII — the marquee on the device shows the first letter of each segment, so long names get truncated visually.

Live recording on a voice

Beyond the formal record mode dialog above, the firmware supports an 8-channel live-loop style flow: touch-hold during playback can capture audio onto a voice for up to 12 seconds, and "one-shot" recordings can drop a take onto a single channel. Behaviour depends on REC → CLR (clear/merge policy) and REC → TRIG (trigger style). Treat this as the "performance" front-end for the same recording engine that powers record mode.