On 11.4.2019 13.09, Kristian Amlie wrote:
On 11/04/2019 07:29, redu wrote:Kristian, I guess -30dB zero value would be just good. Have got no way to check in practice, however.
On 10.4.2019 12.14, Will Godfrey wrote:I recently ran into the part leakage problem as well, as part of a new
On Wed, 10 Apr 2019 09:39:16 +0300Thank you for this information Will. This is good news! Of course you
redu <reino.urala@xxxxxxxxx> wrote:
“Parts leakage” in Yoshimi (and in ZynAddSubFX)This is a known issue and is due to the fact that the volume
Ref:
There is a low level signal leak from PARTS signals. Worst case leaking
is audible when using sinus signal synthesis and multiple activated
parts (Sinus voices are great to reveal all kinds of distortions and
intermodulations too):
calculation never
gets to zero but covers a range +13dB to -40dB.
Changing the range to get the bottom end lower (it could never
actually be zero)
would mess up all previously defined work, as all the intermediate
values would
be shifted down.
There is one possible solution I can think of, but I'd want opinions from
others first.
This would be to completely bypass the part if the volume control was
set below
(say) -39.9dB
As a point of interest there was a related problem we solved several
years ago.
This was when the cutoff for fading note tails was -40dB as well. You
could
distinctly hear the end of a slow decay with no effects or other notes
sounding.
In that case we just changed it to -60dB
are to "bypass the part if the volume control was set below say -39.9dB".
However, the midi signal volume resolution is quite low (0...127)? Maybe
I am wrong, but to play safe, should an optimum threshold level be
checked in practice before freezing it to the master?
feature I'm working on (will soon post to the list!). Me and Will are
still discussing it, but it's possible we will make an exception for the
zero value. Right now that corresponds to -30dB [1], but we could just
turn it all the way to -inf dB (zero amplitude). Would that help your case?
[1] I think it's not -40dB, because it goes to +10dB, not to 0dB