[fluid-dev] SoundFont handling

[Lutz Morelater]

Hi,

SoundFont files are an essential part for fluidsynth. I have two questions 
related to them.

1.) I would like to know which is the cleverest way to work on them?

There is swami, I have used that, but swami is probably best for finetuning 
sounds and creating a SF2 file and not for reorganizing large soundbanks 
spread over several files. And swami has crashed on my laptop quite 
frequently when I was not doing much apart from browsing (listening) a few 
(two, three, four) SF2 files. There was no error message, it was just gone.

What I need is some way to handle SF2 files.

For example: I have downloaded several rather large GM compatible collections 
of sounds and now I want to keep only those instruments (presets) of them 
that really sound great to me. I am going to discard any drumset in a SF2 
file and create a seperate file for drumsets, and my soundbanks should have 
no more than 100 sounds each (that condition is caused by my MIDI hardware 
controller panel that is unable to select soundnumbers greater than 99).

I have found that the file size often is misleading, i.e. large files contain 
nice sounds and crappy sounds as well, as do small sized collections not only 
contain crappy sounds but also real gems.

So the question is how to effeciently reorder/resort/reorganize instruments so 
that I get my very own collection of sounds.

2.) The next thing I need to know is if there is a tool that helps to identify 
doublettes of sounds. Many SF2 files contain similar or even identical 
instruments that differ from each other in hardly any detail, if at all. It 
would be great if a tool, fed with SF2 file names, could browse all presets 
in the files and remove all sounds that can already be found in one of the 
SF2 files checked.

Thanks for any help,
Lutz.