Roberto Gordo Saez wrote:
As far as I know, soundfonts do not allow crossfading
:-(
That is a pity. It would be great to have velocity layers that overlap,
and automatic attenuation of each in opposite directions such that the
overall volume remains constant for each velocity value in the region of
the overlap.
If that were the case in soundfonts, I think one could make a very good
piano soundfont with just 2-3 samples at various dynamic levels
(loudness/softness) of each sampled note/frequency.
Without crossfading, velocity layering becomes more complicated. I made
a copy of your disklavier soundfont and deleted all "instruments" from
the piano patch except the piano samples at ~100, ~117 and ~122 velocity
value. Using just those three instruments as layers, I made the 100
layer extend the full velocity range, the 117 extend to cover about the
top (higher value) 2/3 of the velocity range, and the 122 cover the top
1/3. Then I attenuated the 117 layer by about 8-10 dB, and the 122 by
about 15-18 dB.
I made some other adjustments (some samples, in the bass register for
example, were noticeably louder than the samples for adjacent note
ranges), and the result sounded OK to me, but the jump in volume when
two or three velocity layers became active and sounded at the same time
was still very noticeable, of course. That's when I began to search for
the crossfading controls in the soundfont editor-- :-(
Another feature that I miss is the capability to place
multiple
redundant samples per note and velocity layer, so they are randomly
choosen. This way the soundfont will be even bigger (sorry), but it
solves a particular problem: when playing the same note in a quick
sequence, our human ears notice very soon that exactly the same sample
is playing each time.
I'm learning to play drums on a MIDI electronic drum set (a Yamaha
DTXpress IV "Special"), and have learned that this is one of the common
complaints of drummers who use electronic drums. When playing a snare
drum, for instance, very rapidly, it can sound "like a machine gun"
because exactly the same sample is repeatedly triggered by the rapid
strokes of the drumsticks. In the drumset I'm using, the sound does vary
to some extent if the sticks strike the drum head in slightly different
locations, as they normally do. I'm guessing that the same sample is
modified by digital filters and effects processing to cause the
variation in sound.
Best wishes,
Steve