fluidsynth

A real-time software MIDI synthesizer using SoundFont 2

TLDR

Play a MIDI file with a SoundFont

$ fluidsynth [/usr/share/soundfonts/soundfont.sf2] [path/to/file.mid]

Specify the audio driver

$ fluidsynth -a [pipewire|pulseaudio|alsa|jack] [soundfont.sf2] [file.mid]

Render a MIDI file to a WAV audio file (no real-time playback)

$ fluidsynth -T wav -F [output.wav] [soundfont.sf2] [file.mid]

Run as a MIDI server accepting external MIDI input

$ fluidsynth -s [soundfont.sf2]

Batch process multiple MIDI files without interactive shell

$ fluidsynth -ni [soundfont.sf2] [file1.mid] [file2.mid]

Set sample rate and master gain

$ fluidsynth -r [44100] -g [1.0] [soundfont.sf2] [file.mid]

Disable reverb and chorus effects

$ fluidsynth -R 0 -C 0 [soundfont.sf2] [file.mid]

SYNOPSIS

fluidsynth [options] [soundfont] [midifile...]

FluidSynth is a real-time software MIDI synthesizer implementing the SoundFont 2 specifications. It synthesizes MIDI input or files into audio output using SoundFont instrument definitions.The synthesizer can operate in multiple modes: playing MIDI files directly, running as a server for external MIDI input (keyboards, sequencers), or fast-rendering to audio files. An interactive shell allows real-time control of synthesis parameters.FluidSynth supports various audio backends (PulseAudio, PipeWire, ALSA, JACK) and MIDI backends (ALSA sequencer, JACK MIDI), making it suitable for both standalone use and integration into larger audio setups.

PARAMETERS

-a, --audio-driver driver

Audio output driver (pulseaudio, pipewire, alsa, jack, oss, etc.).

-m, --midi-driver driver

MIDI input driver (alsaseq, alsaraw, jack, oss, etc.).

-F, --fast-render file

Render to audio file as fast as possible (no real-time playback).

-T, --audio-file-type type

File type for fast-render output (wav, raw, aiff, au, flac, oga).

-n, --no-midi-in

Disable MIDI input.

-i, --no-shell

Disable interactive shell.

-s, --server

Run as background server process.

-r, --sample-rate rate

Set audio sample rate (default: 44100).

-g, --gain value

Set master gain (0.0-10.0, default: 0.2).

-L, --audio-channels num

Number of audio channel pairs.

-R, --reverb 0|1

Enable or disable reverb effect.

-C, --chorus 0|1

Enable or disable chorus effect.

-o setting=value

Set arbitrary synth setting (e.g., `synth.polyphony`).

-c, --audio-bufcount count

Number of audio buffers.

-z, --audio-bufsize size

Size of each audio buffer in frames.

-V, --version

Display version information.

-h, --help

Display help message.

CAVEATS

Requires at least one SoundFont file (.sf2 or .sf3) to produce sound. Many distributions provide default SoundFonts in /usr/share/soundfonts/. Real-time synthesis requires adequate CPU resources; complex SoundFonts may cause audio dropouts on slower systems. The -F fast-render option disables real-time audio output.

HISTORY

FluidSynth was created by Peter Hanappe in 2000 as an open-source software synthesizer. Originally named "IIWU Synth," it was renamed to FluidSynth and has grown into a widely-used MIDI synthesizer for Linux, macOS, and Windows. The project continues active development, with version 2.x introducing significant improvements to the API and audio quality.