These changes are meant to address missing percussion in GM and GS
MIDI files. The GM standard says that note-offs should be ignored on
MIDI channel 10 for all but Long Guiro and Long Whistle. In addition,
the MIDI standard requires that for each note-on there is a
corresponding note-off. Many GM/GS MIDI files have very short
duration percussion notes, since its assumed that note-offs are
ignored. Ignoring the percussion note-offs is the right thing to do
and is what composers for GM/GS files expect.
All this is enabled only if GM/GS SYSEX messages are received which
enable these modes or the user explicitly assigns GM or GS mode via
the FluidSynth settings. Otherwise FluidSynth behaves as normally and
all note-offs are honored. This is just the beginning of adding GM/GS
support, there are other things that need to be addressed as well.
Hope that answers your question.
Josh
Quoting "S. Christian Collins" <[email protected]>:
Josh,
How does this affect the GS sound effects bank?
-~Chris
[email protected] wrote:
So I went ahead and added GM On/Off and GS Reset SYSEX handling.
There are now 2 parameters synth.midi-mode=normal/gm/gs and
synth.midi-mode-lock=no/yes. If midi-mode-lock is set to yes then
SYSEX MIDI mode messages have no effect (useful for forcing
FluidSynth to a particular mode). If the current MIDI mode is GM
or GS, then percussion note-offs are ignored for all except Long
Guiro and Long Whistle, as per the GM spec. I tested this with a
GS file and it seems to work as expected (no more missing drums).
TODO:
1 Reset MIDI mode as part of fluid_synth_system_reset()
2 Reset controllers when GM is turned on (and other values
mentioned in gmguide2.pdf)
3 Figure out how to synchronize the midi-mode setting with the
midi_mode variable, when it is being assigned from the synthesis
context.
For item 3. I'm thinking of adding value get callback
functionality to settings, so that a callback can be registered for
obtaining the current value (so for example the midi-mode string
setting could be generated from the current value of the enum
synth->midi_mode). Perhaps some hack can just be implemented for
now though, but calling settings functions from within synthesis
context isn't really a good idea (mutex locks, mallocs, etc).
Feedback and testing would be appreciated :) The "normal" value
for midi-mode doesn't sound great, but I couldn't think of a better
word.
I'm going to do my best to get FluidSynth 1.1.0 released by the end
of this month (in a nice working/stable state with good API
documentation). Testing FluidSynth in various environments and
reporting success/problems would be really helpful. There are some
fairly significant changes with the whole synthesis path and the
more testing FluidSynth in real word scenarios the better.
Cheers!
Josh
_______________________________________________
fluid-dev mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/fluid-dev
_______________________________________________
fluid-dev mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/fluid-dev
