Autosample - GUI + CLI

A lightweight Rust autosampler for macOS, Windows, and Linux, focused on a desktop GUI workflow with a CLI also available. It triggers notes over MIDI, records audio input, and exports organized multisamples with metadata.

---

Overview

Autosample helps you sample instruments quickly and consistently. Pick your notes and velocity layers, record each one automatically, and get clean files ready to use in your sampler.

It gives you:

• Automatic MIDI note playback and recording
• Desktop GUI for setup, device selection, and session control
• Optional cleanup tools (trim and normalize)
• WAV or MP3 export
• Resume mode so long sessions are easy to continue

---

Features

• Note and Velocity Parsing: Supports single values, lists, and ranges
• Timing Control: Independent preroll, hold, and tail durations
• Round Robin: Multiple takes per note/velocity
• Output Organization: Flat or by-note directory layout
• Session Metadata: Writes session.json for every run
• Resume Mode: Skips already existing target files
• Cross-Platform: macOS, Linux, and Windows

---

Requirements

Runtime

• A MIDI output target (like hardware synth)
• An audio input device
• ffmpeg (only required when exporting MP3 or both)

Development / Build

• Rust toolchain (stable): https://rustup.rs/

---

Installation

1. Clone

git clone https://github.com/ariel10aguero/autosample
cd autosample

2. Build

cargo build --release

Binary output:

target/release/autosample

---

Quick Start

1. List devices

./target/release/autosample list-midi
./target/release/autosample list-audio

2. Run a small test

./target/release/autosample run \
  --midi-out 0 \
  --audio-in 0 \
  --notes "C4" \
  --vel "127" \
  --output "./output" \
  --prefix "test"

---

Usage Notes

Notes syntax (--notes)

• Range: C2..C6 (ascending only)
• List: C4,E4,G4
• Single: A4
• Accidentals in input: C#4, Db4

Velocity syntax (--vel)

• Range with step: 127..15:16
• List: 127,100,64
• Single: 100

Timing model

Total recording time per sample:

preroll_ms + hold_ms + tail_ms

---

Output Layout

Autosample writes to:

<output>/<prefix>/

With --output-organization flat:

output/Piano/
  Piano_C4_vel127.wav
  Piano_Cs4_vel127.wav
  session.json

With --output-organization by-note:

output/Piano/
  C4_060/
    Piano_C4_vel127.wav
  Cs4_061/
    Piano_Cs4_vel127.wav
  session.json

Accepted values:

• flat
• by-note

---

Common Run Example

./target/release/autosample run \
  --midi-out "Your MIDI Device" \
  --audio-in "Your Audio Device" \
  --notes "C3..C6" \
  --vel "127,100,64,32" \
  --hold-ms 1000 \
  --tail-ms 2000 \
  --preroll-ms 100 \
  --round-robin 2 \
  --format wav \
  --output "./output" \
  --prefix "Piano" \
  --output-organization by-note \
  --resume

---

Troubleshooting

• MIDI device not found: Verify with list-midi and use exact name/index
• Audio device not found: Verify with list-audio and use exact name/index
• Attack is cut: Increase --preroll-ms
• Release is cut: Increase --tail-ms
• Low level output: Use --normalize peak
• Interrupted session: Re-run with --resume

---

GUI Screenshot