2010-08-26 09:04, Bernd Casper skrev:
> Hi David, Gents,
>
> many thanks for the latest changes. It seems to be a great improvement in
> sound quality, whatever the reason may be. Much more details, amazing
> transparency, less gain level distortion.
>
> Overflow settings.
> On my (relatively powerful) system I cannot realize big overall changes, if I
> do alter the weighting. Though, one detail seems very important to me:
> To me it seems, under certain circumstances, FS overflow handling is allowed
> to cut off voices inside a single multi-sample soundfont instrument
> ("Instrument" level).
> To me it was always one of the biggest advantages of FS, to allow as many
> samples in an instrument I wish, without any clipping. All my soundfonts
> building bases on that fact.
> So from my point of view, it should be absolutely avoided, to cut off samples
> inside of an instrument which is already sounding.
>From another point of view - there are often presets made of samples
where on sample is the "main" one and the other is a small, almost
unhearable "effect". In that case, cutting off the "effect" sample would
be the right thing to do.
Anyway, as a future improvement, we could consider a switch changing the
behavior: so that when FS runs out of polyphony, it adds the score of
all voices belonging to the same note together, then kills all voices
belonging to the note-on with the lowest score, instead of killing
individual voices.
> I tried to ensure this by setting age to 9999 and volume to -9999, but I
> noticed no change in handling.
I would probably set the synth.overflow.volume to 0 in that case -
setting it to -9999 would cause *louder* notes to be clipped first and
more silent notes to remain active.
> Release tickles.
> Please listen to my sound example. You'll realize some notes having a little
> tickle in the release (1st & 5th note in the 1st group, none in the second
> group, 3rd & 5th note in the third group, and the very final note). This
> effect is always present currently, while playing - by chance, it seems. It's
> always a little delayed, about 0.25 seconds - like releases cutted off.
> http://bca.free-artists.net/data/tempdl/bca_FS112alpha_clipping_demo.ogg
I'm sorry, but I can't hear anything. Could be a difference in our
listening environments, or that the ogg compression removes the
"tickles". But without hearing them and without a sound font, it is very
difficult for me to do anything about it.
// David
_______________________________________________
fluid-dev mailing list
fluid-dev@nongnu.org
http://lists.nongnu.org/mailman/listinfo/fluid-dev
