1.1.9

Bug Fixes

• fix build if no MIDI drivers are available (thanks to @carlo-bramini)
• fix return value of fluid_file_set_encoding_quality() (thanks to @ReinholdH)
• fix use-after-free in fluid_timer
• fix memory leak in pulseaudio driver
• fix memory leak in rvoice_mixer (thanks to @mawe42)
• fix dumptuning shell command displaying uninitialized values (thanks to @rmattes)
• fix a resource leak in source shell command (thanks to @carlo-bramini)
• avoid a buffer overrun when mixing effects channels in fluid_synth_nwrite_float() (#287)

New Features

• implement handling of FLUID_SEQ_ALLSOUNDSOFF events in fluid_seq_fluidsynth_callback()
• support for registering audio drivers based on actual needs (#218)

1.1.8

• fix a bug around synth.audio-groups setting, which caused improper multi-channel rendering (#225)

1.1.7

Bug fixes

• consistently relicense libfluidsynth under LGPL-2.1+ (also addressing fluid_chorus.c, #165)
• fluid_synth_set_channel_type() was not exported properly
• fix calculations for modulators (#194)
• fix SysEx parsing issues (#127, #148, thanks to Erik Ronström and Stas Sergeev)
• fix mangling with illegal sample loops, causing audible glitches (#171, #149, thanks to @mawe42)
• fix inverse logic of audio.jack.multi option (#135)
• fix channel fine tune RPN to use correct range (#187, thanks to @mawe42)
• fix timing problems when changing the sequencers scale from a callback event (#195, thanks to @imhcyx)
• workaround incorrectly rendered audio when requesting more than 64 frames from fluid_synth_write_*() (#192)
• adjust ALSA MIDI port type (#139)
• avoid voice_count becoming negative (#151, thanks to Jean-Jacques Ceresa)
• avoid notes with a fixed key generator playing forever (#159)
• avoid TCP/IP connections from closing in an undefined manner (b75c8fd)
• a lot of memory leaks, NULL dereferences and SegFaults (thanks to Surabhi Mishra)

New Features

• add support of vorbis-compressed sf3 sound fonts (#140, thanks to @fabiangreffrath)
• add sostenuto pedal to the synth (#47, #134, thanks to Jean-Jacques Ceresa)
• add vbr quality when encoding with libsndfile
• re-implement routing reverb and chorus to distinct buffers in fluid_synth_nwrite_float() (#135)
• add IPv6 support to socket API (#124)

New API calls

• channel, key, velocity and state getter for fluid_voice_t
• tempo, bpm, total length and currentBeat getter for fluid_player (#190, thanks to @quiasmo)