[linux-audio-user] Musical linuxsampler?
markknecht at gmail.com
Wed Nov 8 12:50:49 EST 2006
On 11/8/06, Markus Herhoffer <markus at herhoffer.net> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
> I am not a big friend of sampling as I am used to the sounds of real
> instruments. But I decided to give qsampler/linuxsampler (CVS version
> from Gentoo pro-audio overlay) a try. And as expected my virtual
> symphonic orchestra sounds disgusting.
> It seems to me that this is not a problem of the sampling quality. I use
> some mid-price samples in Gigasampler format. The real problem is the
> "statics" of the sound. Every tone sounds the same. This is extremely
> annoying on woodwinds and brasses.
> Now my question: Are there any dynamic MIDI parameters
> linuxsampler/libgig is able to react to? I've noticed that the engine
> interprets the pitch data. Are there any other parameters (like
> aftertouch, sustain, breath ctrl etc.) the engine responds to? Or are
> there other concepts for making the music more sound like real music?
Reduce your MIDI velocities and you'll start to get more
expressiveness. This is true with both LS (which I don't use anymore)
and GSt. (which I use all the time) I find keeping average velocity
way down in the 40, 50, 60 ranges is what more library vendows expect
for average notes. The libraries I use are all 4, 8 & 16 level
libraries. If you are using an inexpensive 1 or 2 level library then
your results will likely not be as good.
IIRC LS had some problems when you used high velocities with some
sound cards. LS created distortions that GSt did not.
Set up orchestral based panning to place the sounds in more real
locations and things will likely also improve.
More information about the Linux-audio-user