MIDI Velocity and Dynamics

Hi,

I'm looking into dynamics. Arbitrarily, as a starting point, I am saying
that ppp is note velocity 10, pp is 19, p is 31, mp is 47, mf is 63,
f is 81, ff is 107, and fff is 127.

Here are the results of my initial tests.

Each audio instrument has its volumn set to -20 dB.
And, each instrumental section feeds into an
Apple reverb.

For the Westgate Studio Woodwinds extended edition

http://www.westgatestudios.com/wc_expanded.htm

I used one solo instrument of each type: piccolo, flute, oboe,
English horn, Eb clarinet, Bb clarinet, bass clarinet, bassoon,
and contra bassoon, and assigned them each a two measure
phrase in the middle to higher portions of their range. No
two instruments doubled one another (so they were not
absolute octave doublings, some doublins in thirds, some
in fifths).

Each time the group played the two measure phrase, I increased
the note velocity. So, the first time the two measure phrase
was played, representing ppp, all the MIDI velocities were 10.
The second time the two measure phrase was played, representing
pp, all the MIDI note velicities werw 19, and so forth, all the
way up to fff which I made velocity 127.

I did not worry too much about the flute not really being able to
play ppp in its upper register. The general idea was to accept
that it could (and when I write real music, I would simply never
notate a flute at ppp in its upper register).

The complete group as a whole was very well integrated in its
dynamics and timbral changes as the "notated" dynamics
changed. The resultant decibal levels for the complete
woodwind group at each dynamic was

ppp -53
pp -49
p -45
mp -40
mf -36
f -30
ff -24
fff -21

I did the same experiment using the standard string sound for
Garritan Orchestral Strings (i.e., not warm and not tender). In
this case, strings can play softly and loudly in any register.
I assigned Violins I, Violins II, Violas, Cellos, and Double Basses
each to its own unison line, the same two bar melody used
above with these results in dB:

ppp -50
pp -47
p -42
mp -37
mf -33
f -28
ff -27
fff -24

When I played the woodwinds and the strings together, the results
in dB were:

ppp -49
pp -45
p -41
mp -35
mf -31
f -27
ff -23
fff -19

and the two sections balanced reasonably well (and if it had happened
that they did not balance, then I could always lower the volume of
any given section, complete woodwinds or complete strings, since
each has their own bus).

In general, then, it appears that if I use the numerical scheme above.
where ppp meants note velocity 10, and so forth, that the woodwinds
as a group remain coherent and reasonably in balance, as to the
strings, and as do the groups when combined. This is good, in that
it means that when writing I do not constantly need to tweak the
dynamic of each note (because I am not looking for an absolutely
professional sounding final sound, just a sound which doesn't sound
to awfully out of balance).

I ran into problems using this technique with the SampleTekk
brass

http://www.sampletekk.com/proddetail.php?prod=STDELIVER-018-FORMAT

This is not to say that the brass are at fault, after all, can a trumpet
play ppp in its upper range? Nevertheless, their trumpet 1, for
instance, didn't match well with the timbral changes and dynamic
changes as the musical dynamics changed from ppp to fff, whereas
the Apple trumpet and the Westgate Studios free trumpet did a
better job (though not necessarily a perfect one.

So, I may back-off from the SampleTekk for a while and use Apple
and Westgate Studio stuff until I have the brass integrated with the
above dynamic plan. Then, consider whether I want to go into
EXS-24 and start formulating my own musical instrument (for instance,
I would need to change which patches (piano, mf, or forte) are sounded
for the trumpet, and perhaps adjust the volume as a function of
key velocity), skills I do not yet have.

Thanks,
Sys6