Setup for Experienced Python Users

Basic Installation

SCAMP requires Python 3.12 or greater. If you’re already familiar with Python and you know how to install packages using pip, installing SCAMP is a one-liner:

pip install scamp

The prebuilt wheels on PyPI bundle the FluidSynth library for Linux, macOS (both Intel and Apple Silicon), and Windows, so soundfont playback works out of the box on all three platforms — no separate FluidSynth install is required.

To pull in the optional extras (LilyPond export, MIDI I/O, keyboard/mouse input) all at once:

pip install "scamp[all]"

This installs abjad==3.31 (pinned for compatibility), python-rtmidi, and pynput.

Optional (but recommended) Components

FluidSynth: bundled vs. system

You usually do not need to think about this section — the bundled FluidSynth that ships with the SCAMP wheels is enough to play notes on all supported platforms. It’s worth knowing the mechanism, though, in case you want to swap in a system copy (especially on Linux, where a package-managed FluidSynth is often better integrated with the host audio stack — JACK, PipeWire, PulseAudio routing, etc.).

Two playback_settings flags govern what gets loaded:

try_system_fluidsynth_first — if True, SCAMP tries to dlopen the system libfluidsynth first and falls back to the bundled binary if that fails. Defaults to True on Linux and False on Mac/Windows. Set it to True on any platform if you’ve installed FluidSynth yourself (e.g. brew install fluid-synth, sudo apt install fluidsynth) and want SCAMP to use it.
use_bundled_pyfluidsynth — if True (the default everywhere), SCAMP uses its own copy of the pyfluidsynth Python wrapper, which can dlopen either the system or bundled libfluidsynth. Set to False to use a separately pip installed pyfluidsynth instead.

You can change either of these via PlaybackSettings and persist with save_to_json(). To inspect what actually got loaded in the current process, see the testing section below.

Installing python-rtmidi

For midi input, and also to generate an outgoing midi stream (which could, for instance, be routed to a DAW), you will need the python-rtmidi library. You can get this by running from the command line:

pip install python-rtmidi

On Linux, if you’re running into an error you may need to first install the python3-dev package, for instance with the command:

sudo apt install python3-dev

For any other python-rtmidi installation woes, take a look at the installation instructions here.

Installing abjad and LilyPond

For LilyPond output, you will need the abjad library. To do so, run the following:

pip install abjad==3.31

Note the ‘==’ in the command, which specifies the exact version of abjad to install. This is the version that SCAMP has been built to be compatible with. abjad has historically made breaking API changes between releases, so newer versions may not work; if you’re not sure which version you have, the simplest path is pip install "scamp[all]", which pins this version for you.

After installing abjad, you will also need to download and install LilyPond, since abjad relies upon it. You may run into security issues, which are further explained in Easy Setup for New Python Users.

Installing scamp_extensions

The scamp_extensions package is the place for models of music-theoretical concepts (e.g. scales, pitch-class sets), additional conveniences for interacting with various types of input and output, and in general anything that builds upon SCAMP but is outside of the scope of the main framework.

The easiest way to install scamp_extensions is by running the command:

pip install scamp_extensions

To install the most up-to-date version (assuming you have git installed), you can instead run:

pip install git+https://github.com/MarcTheSpark/scamp_extensions

This will install the latest version from this repo.

Testing the install

A quick sanity check that everything is wired up:

from scamp import test_run, print_dependency_status

print_dependency_status()
test_run.play(show_lilypond=True)

print_dependency_status() prints a one-line-per-dependency report covering FluidSynth (with the wrapper and the resolved libfluidsynth binary path, so you can confirm whether the system or bundled copy got loaded), sf2utils, python-osc, python-rtmidi, pynput, and abjad (with installed-vs-tested-version comparison). States are ok / warn / missing. If you’re excessively nerdy and would rather consume the data programmatically, dependency_status() returns the same information as a list of (name, state, detail) tuples.

test_run.play() plays a short piano gesture sweeping inward toward middle C; with show_lilypond=True it also exercises the abjad → LilyPond → PDF path to display notation.