Re: [fluid-dev] FluidSynth velocity to attenuation curve

[Ebrahim Mayat]

> Anyway, here is what is tested in the file I sent:
> Test #1: velocity->attenuation
> Test #2: attenuation as specified by preset (patch changes must be 
> working to hear the dynamic changes in this test)
> Test #3: volume (controller 7)

Chris

I did the tests (results below) on Fedora 8 using the Planet-CCRMA
realtime kernel 2.6.22.10-1.rt9.1.fc8. The master volume was maintained
at 50% throughout the experiments.

Firstly, it is not possible to render the 'softsynth velo curve
test .mid' file using `fluidsynth -a jack 'softsynth...sf2'
'softsynth...mid' ` within the terminal. 

If I only substitute the MIDI file with one of my own then I can render
square waves. 

If however, I keep your MIDI file and substitute the soundfont with one
of my own,there is no sound. 

So, there probably is a problem with the MIDI file.

Then, I tried rendering your MIDI file from within rosegarden using the
fluidsynth-DSSI plugin. Now, I can hear sounds.

There is definitely a much broader dynamic range in tone series 3
compared to the other two and tone series 2 has the most restricted
dynamic range. In tone set 3, the tones start at a much lower volume
than those in set 1. 

So, the  velocity to attenuation curve is evidently restricting the
dynamic range. 

What is a bit puzzling though is that in the second tone set, the first
and second tones are loudest and there seems to be a little audible
difference in the subsequent tones. As a control, when I substituted
your soundfont with a GM one, preset changes could clearly be heard for
tone series 2.

Regards.
E