Josh,
I've looked at the source code, but I'm not really a programmer. How
hard would it be to make velocity, volume (cc7) & _expression_ (cc11)
calculated using fluid_cb2amp (which is at -200) and let the
SoundFont's internal attenuation setting continue using the
fluid_atten2amp setting (at -531.509)? I think this would be the best
solution, but then again... I'm not a programmer :)
-~Chris
Josh Green wrote:
Hello Chris,
The whole attenuation issue with Creative Labs sound cards has been
discussed before in regards to FluidSynth, but I don't think it was ever
resolved correctly. As you noted, they do not follow the SoundFont spec
in this regard. I've contemplated whether FluidSynth should just resort
to AWE 32/64/Live!/Audigy compatibility mode when the SoundFont
indicates that it was made for one of those devices (EMU8000, EMU10K or
others ??? in the Sound Engine field). If you have any ideas in this
regard, that would be great. I imagine Swami and other SoundFont
editors should probably follow this anomaly as well.
Cheers.
Josh Green
On Thu, 2008-07-17 at 02:28 -0500, S. Christian Collins wrote:
Well, I have found a problem with the change I made to the velocity
scale. While it improves the velocity responsiveness, as well as
related controllers (volume & _expression_), it messes up the
attenuation level as set within the SoundFont.
On the Live!/Audigy, setting the attenuation to 8 dB actually
attenuates the signal by about 3.1 dB. The unmodified FluidSynth will
attenuate the same patch by 2.9 dB, so it's pretty close. However,
with my change, the attenuation is now exact to what is set in the
SoundFont (so setting it to 8 dB actually attenuates it by 8 dB).
Honestly, I don't know why the Live!/Audigy synth is set up the way it
is, because you would think that setting it to attenuate by 8 dB
*should* actually attenuate by 8 dB. However, since most SoundFonts
expect a Live!/Audigy type response to the attenuation setting to
render correct sound balance, this should IMO be fixed in FluidSynth
to avoid breaking compatibility.
I will peruse the source code to see if I can find what to change so
that the velocity curve can remain modified, but since I'm not a
programmer, maybe someone else here might be able to find it more
quickly?
Anyway, time for bed for me--2:26 am here...
-~Chris
Josh Green wrote:
On Sat, 2008-07-12 at 11:21 -0500, S. Christian Collins wrote:
Dear FluidSynth developers,
I have created a compilation of changes that I have made to the
FluidSynth code. You can view the changes in PDF format by
downloading from this link:
http://www.schristiancollins.com/temp/20080711_changes.pdf
I would like to propose that these changes be implemented in the next
version of FluidSynth (1.0.9?). I would like to publish SoundFonts
designed for FluidSynth, but my SoundFonts will require some of these
changes in order to function 100% as designed. I hope you will agree
that these changes are for the better of FluidSynth and consider their
inclusion in the project. Please let me know if anybody has problems
viewing the PDF.
Thank You,
S. Christian Collins
Hello Christian,
Nice to hear of your changes and work to make FluidSynth synthesize in a
more expected manner. I'm currently traveling and will be back home in
a few weeks. Its long overdue for me to put in some time into
integrating the various changes that users have sent. Thanks again for
the contributions.
Josh Green
|