NAME
sonivoxrender — Render MIDI song files into raw PCM audio
SYNOPSIS
sonivoxrender [-h|--help] [-v|--version] [-d|--dls soundfont] [-r|--reverb 0..4] [-w|--wet 0..32767] [-n|--dry 0..32767] [-c|--chorus 0..4] [-l|--level 0..32767] [-g|--gain 0..196] [-V|--Verbosity 0..5] [-R|--reverb-post-mix] [-C|--chorus-post-mix] [-s|--sndlib 1..3] song_file
DESCRIPTION
This program is a MIDI file renderer based on the sonivox synthesizer library. It reads .MID (Standard MIDI Files), .RMI (RIFF-RMID), and .XMF (Extensible Music Files) file formats, and writes an audio stream to the standard output as raw 16 bit stereo PCM samples.
Options
- -h, --help
-
Prints brief usage information.
- -v, --version
-
Prints the version numbers.
- -d, --dls soundfont
-
Optional DLS or SF2 soundfont file name. If not provided, it uses an internal embedded soundfont.
- -r, --reverb reverb_preset
-
Reverb preset between 0 and 4: 0=no, 1=large hall, 2=hall, 3=chamber, 4=room.
- -w, --wet reverb_wet
-
Reverb wet level between 0 and 32767.
- -n, --dry reverb_dry
-
Reverb dry level between 0 and 32767.
- -c, --chorus chorus_preset
-
Chorus preset between 0 and 4: 0=no, 1..4=presets.
- -l, --level chorus_level
-
Chorus level between 0 and 32767.
- -g, --gain master_gain
-
Master gain between 0 and 196, default is 100 (+0dB). The number is relative to 100, in 1dB increments, e.g. 120 = +20dB, 80 = -20dB.
- -V, --Verbosity verbosity
-
Verbosity level between 0 and 5, where 0=no, 1..5=severity levels.
- -R, --reverb-post-mix
-
Ignore CC91 reverb send level. The reverb effect will apply to mixed output audio, which is the old behavior.
- -C, --chorus-post-mix
-
Ignore CC93 chorus send level. See also --reverb-post-mix.
- -s, --sndlib index
-
EAS sound library to use:
- •
-
1: wt_200k_G (default) - WT-only bank, used in Android devices. Also named “Common”. Support 22050 Hz and 44100 Hz sample rates, 8-bit and 16-bit samples.
- •
-
2: GMdblib-3 - FM-only bank. Support all sample rates. Sample bit depth does not matter here.
- •
-
3: hybrid_22khz_mcu - Hybrid bank. Use WT synth for drums and FM for melodic instruments. This bank is a combination of
GMdblib-3andSonic_20Khz_Drums. Support 22050 Hz sample rate and 8-bit samples only.
Note: This option does not affect DLS/SF2. They will always use the DLS synth engine. If the selected sound library is not compatible with the build configuration, the program will fail with an error message.
Arguments
- song_file
-
Input [MID|RMI|XMF] file name.
EXAMPLES
The following examples assume the default option USE_44KHZ=ON, which means an output sample rate = 44100 Hz.
Example 1: Render a MIDI file and save the rendered audio as a raw audio file:
-
$ sonivoxrender ants.mid > ants.pcm
Example 2: pipe the rendered audio thru the Linux ALSA aplay utility:
-
$ sonivoxrender ants.mid | aplay -c 2 -f S16_LE -r 44100
is equivalent to:
-
$ sonivoxrender ants.mid | aplay -f cd
Example 3: pipe the rendered audio thru the lame utility creating a MP3 file:
-
$ sonivoxrender ants.mid | lame -r -s 44100 - ants.mp3
Example 4: pipe the rendered audio thru the sox utility creating a WAV file:
-
$ sonivoxrender ants.mid | sox -t s16 -c 2 -r 44100 - ants.wav
Example 5: pipe the rendered audio thru the PulseAudio’s pacat utility:
-
$ sonivoxrender ants.mid | pacat
Example 6: pipe the rendered audio thru the PipeWire’s pw-play utility:
-
$ sonivoxrender ants.mid | pw-play --rate 44100 -
Example 7: pipe the rendered audio thru the FFmpeg’s ffplay utility:
-
$ sonivoxrender ants.mid | ffplay -i - -f s16le -ar 44.1k -ac 2 -nodisp -autoexit -loglevel quiet
Example 8: pipe the rendered audio thru the mpv media player:
-
$ sonivoxrender ants.mid | mpv --demuxer=rawaudio -demuxer-rawaudio-format=s16le --demuxer-rawaudio-rate=44100 --demuxer-rawaudio-channels=2 --no-video -
BUGS
See Tickets at GitHub <https://github.com/EmbeddedSynth/sonivox/issues/>
LICENSE AND COPYRIGHT
Licensed under the Apache License, Version 2.0
Copyright (c) 2022-2026 Pedro López-Cabanillas and contributors
AUTHORS
Pedro López-Cabanillas <[email protected]>.