                     VL Programming Guide and Tutorial
                       Yamaha Corporation of America
                             Copyright  1996

Version 1.0

ABOUT THIS GUIDE

This document is provided free courtesy of Yamaha Corporation of America
and Manny Fernandez. It originated as a reference for YCA programmers
creating voices for the VL70m. As such, it only covers the VL Version 1
parameter set that is also common to the VL70m. The new parameters added in
the VL Version 2 software are not discussed. All parameters discussed are
identical in function for the VL1, VL1m, VL7 (Versions 1 &;2) as well as
the VL70m.

This is a text version of the guide.It is also available in HTML 
format with GIF figures.

There is also an example voice file, vlvoices.hqx. 
This is in VL v2 Expert Editor for Macintosh format. The
construction of these voices is discussed in detail in the tutorial
section, and seeing the Instrument Page parameters while reading the
Tutorial will help you to learn how different types of voices are
programmed.

Again, this version of the Editor's Guide is being provided free of charge,
with no support planned by either Yamaha or Manny Fernandez for Version
1.0.

An expanded and enhanced version of this Guide is being developed to cover
the new Version 2.0 parameters, as well as detailed discussions of
Fricative, Throat, Mixing, Crossfade and Interpolate -- including a new
voice example library. It will be available directly from and supported by
Manny Fernandez. For further information on the Version 2.0 Expert Editor's
Guide, Manny's EMail address is 74071.351@compuserve.com

If you would like to make comments on this Version 1.0 of the Expert
Editor's Guide, send them to Manny -- but please do not ask specific
questions about the Version 1.0 Guide or VL programming.

Thanks for the interest, and have a modelingly good time with your VL!


                             Table of Contents

      About this Guide        VL Programming Guide and Tutorial

      Table of Contents

      Chapter 1               The Modeling Paradigm
                              Editor Overview
      Chapter 2
      Basic Construction

      Chapter 3               Algorithm Overview
      Chapter 4               Mode Behaviors

      The VA Algorithm - Parameter Overview and Definitions

      Chapter 5               Pipe/String Overview
      Chapter 6               Driver Overview
      Chapter 7               Parent Reed Overview
      Chapter 8               Child Reed Overview
      Chapter 9               Embouchure Control Overview
      Chapter 10              Pressure Control Overview
      Chapter 11              Beat Overview
      Chapter 12              Excitation Overview
      Chapter 13              Compensation Overview
      Chapter 14              Pipe Length Overview
      Chapter 15              Throat Formant Overview
      Chapter 16              Let's Make a Model !!

      VL Modeling Tips & Tutorial

      Chapter 17              General Single Reed Tips
      Chapter 18              General Jet Reed Tips

      Chapter 19              Miscellaneous Tips That Work, But With
                              No Scientific Basis Whatsoever!

      Final Words
      Chapter 20              The Circle of Life
      Chapter 21              Embouchure Controlling Tables

The Modeling Paradigm

Virtual Acoustic synthesis is unlike traditional types of synthesis because
you are programming the physics of instrument behavior and construction
(not the sound!). This reproduces in the holistic interaction of pitch,
timbre and behavior which occurs in acoustic instruments.

                                  [Image]
Indeed, it is the skillful control of pitch, timbre and behavior which
allows acoustic instrumentalists to create nuances and expression in their
playing that traditional synthesis cannot reproduce.

This is because traditional synthesis is based on the reduction of sounds
into discrete pitch, timbre and behavior components i.e. Assign a tuning
(pitch), choose a waveform or sample (timbre), then alter with filter and
amplifier envelopes (behavior). There is the ability to modify behavior
with realtime controllers like Mod Wheel for vibrato, Foot Controller for
filter sweeps and the Pitch Wheel for bend articulations, but each of the
effects is discrete and does not recreate the interdependent interactions
that happen in a acoustic instrument.

Programming the VA system will put you in the role of instrument designer,
and you will need a new paradigm in your approach to creating and editing
sounds. Namely, you are no longer directly dealing with the 'sound' itself,
but instead are dealing directly with the instrument . Thus you change the
construction of the instrument to indirectly affect the desired change in
the 'sound.' For example, you are no longer making the 'sound' brighter by
opening a filter. You are now making the 'sound' brighter by increasing the
airflow speed through the instrument, widening the flare of the bell and
changing the dampening characteristics of the wood, metal or plastic body.
The essence of VA synthesis dictates that now the 'sound' is brighter, but
it now also has a slightly different harmonic structure, the pitch and
scale temperament is altered, and the attack/decay and legato/staccato
articulation behavior is different as well. The fact that VA synthesis
always responds with this interaction between pitch, timbre and behavior
gives it the ability to recreate the nuances and expression of acoustic
instruments.

The goal of this guide is to give you insight into the basics of the VA
algorithm structure and it's interactive nature so you can create and edit
sounds to your desired result, as well as provide tips and templates for
generalized instrument construction. By it's design, VA synthesis emulates
the real physics, so you can easily create instruments that don't, won't or
can barely play. Traditional acoustic instrument designs were created and
refined over generations, and reflect refinements in design to maximize
specific physical behaviors. The first time Adolphus Sax tried to create a
metal woodwind, it was a far cry from a Selmer Mark VI !!


Editor Overview -

The VED editor is divided into 7 pages of parameters, two Common pages for
voice level parameters that are shared by the two elements, and five
Element pages of parameters that are independently programmed for each
element. These pages are:

Common Miscellaneous Page
Common Effects Page
Element Controllers Page
Element Miscellaneous Page
Element Instrument Page
Element Effect Page
Element Envelopes Page

All parameter fields followed by the [>] are breakpoint scalable by note
number. The number of breakpoints available depends on the parameter. This
is for VL Version 1.0 software. Some Version 2.0 Parameters are not
discussed (Straight Horn Insertion, etc.)

Basic Construction

Algorithm Overview -

The VA algorithm can be divided in to two types of parameter categories --
parameters that are integral to defining the physics of the instrument, and
parameters that reshape the timbre produced by the given physical
definition. You can think of this first group as parameters that are
'inside the model' and interact holistically in regards to pitch, timbre
and behavior. Changing any one parameter to alter timbre requires a
corresponding change in other parameters to keep the pitch and behavior
consistent. These parameters must be edited via Software and consist of the
all Element Instrument Page parameters, as well as the Throat Formant
parameters on the Element Miscellaneous page (which can be edited from the
VL from panel). The second group of parameters are 'outside the model' and
act independent of one another as in traditional synthesis. All of this
second parameter group are editable from the VL front panel. This guide
will focus on the Instrument Page parameters that define the physics of the
virtual instrument.

The VA model as implemented in the VL-1/-1M/-7 /-70m is based on the
saxophone, where a single reed mouthpiece drives a conical tube resonating
body. It contains two main portions. First is the Pipe (or String) which
describes the resonating body or tube, and the second is the Driver, which
describes the reed behavior and airflow dynamics therein. There are
additional portions that allow for the generalized sax model to be
extrapolated to approximate other instrument types. The VA algorithm
defines an instrument's physical behavior and is capable of creating the
following behavior categories (with some examples of corresponding
instrument timbres):

   * Single Reeds: Sax, Clarinet, etc.
   * Jet Reeds: Flute, Recorder, Shakuhachi, Pipe Organ pipe, etc.
   * Lip Reeds: Trumpet, French Horn, etc.
   * Bowed Strings: Violin, Kokyu, etc.
   * Plucked Strings: Guitar, Bass etc.

Double Reed instrument timbres such as Oboe and Bassoon can be produced,
although the VA algorithm cannot reproduce true double reed behavior. In
addition, a wide range of 'synthesizer' timbres can be created. Remember
that behaviors are being created, and any specific timbre is a side effect
of a given behavior.


Mode Behaviors

It is important to understand the tuning mode behavior of instruments when
programming VA synthesis. Three recommended references are Orchestration by
Walter Piston; Horns, Strings and Harmony by Arthur Benade; and The
Acoustical Foundations of Music by John Backus. Acoustic brass, wind and
reed instruments have limited note fingering but can play over many octaves
because they change modes i.e. the player blows the first, second, third,
or fourth harmonic overtone etc. while repeating the basic fingering. For
example, the basic fingering on a sax selects the 12 half steps in an
octave. When you play the lowest note on the sax, you're playing the
instrument in first mode. Think of it as it's 'fundamental' mode. After you
have fingered the first octave in the scale, the register key in pressed
and the pitch jumps up to another octave -- you are now in second mode.
With the same fingerings the next octave of pitches are available. A
trumpet has only 7 valve positions, yet it is capable of playing over three
octaves because it can play in 9 modes. Six notes can be played in each
position.

Since the VL notes' are selected via a keyboard or MIDI, the algorithm must
be programmed to automatically change modes as you play up and down the
keyboard. Modes can also be selected in realtime via MIDI controllers if
the virtual instrument has been properly constructed and the Embouchure
Control parameters are properly set. These Embouchure Control parameters
mimic the way that a wind, reed or brass player changes modes by varying
their embouchure/lip tension, blow angle or breath velocity. Thus flute
'overblow' effects and trumpet 'shake' effects are possible in realtime and
contribute to the dynamics and expression of a VA voice.

                             The VA Algorithm -

                     Parameter Overview and Definitions

Pipe/String Overview -

The VA algorithm resonating body consists of a mathematical approximation
of a conical tube of variable length and taper as well as variable damping
characteristics. It also incorporates a loss function that weakens all
even-numbered harmonics to produce saxophone-type register (octave) key
behavior.

To approximate the conical tube instrument driven from the small end of the
cone, the total length of a straight tube is divided into two parts, the
Short Length and the Long Length. The Short Length is how far from the end
of the tube that the driver (mouthpiece) is connected to the resonating
body.

                                  [Image]
                           FIGURE 1: SHORT LENGTH

The VA algorithm actually generates different pitches by expanding or
shrinking the size of the virtual instrument. The pitch depends on the
total length of the tube. The VA algorithm allows for two ways for the
total tube length to be determined -- by either varying both the Short
Length and Long Length together, or by just varying the Long Length alone.

Parameters:

Short Length Mode -- Selects Short Length tracking mode. Relative mode
means that the Short Length changes relative to pitch of the instrument, so
that the Short Length/Long Length are changing together and their ratio
remains constant for any pitch, similar to a trombone of a slide whistle.
Absolute mode means the Short Length is fixed at its programmed value
regardless of pitch, with only the long length changing. Thus the Short
Length/long length ratio is variable with each pitch. This is similar to
opening/closing the tone holes on a sax or clarinet, etc.

Short Length/Ratio [>] <16 Breakpoints> -- (See Figure 1) This is the
distance from the closed end of the conical tube at which the driver
connects, and is the primary origin of the harmonic structure of the sound.
The Short Length sets the template of possible harmonics in the sound. The
distance of the driver from the closed end of the tube creates reflections
that will phase cancel or weaken certain harmonics. For example, if the
Short Length is set to 25% (of the total length), all 4th order harmonics
will be significantly reduced or missing, similar to a pulse wave with a
width of 25%. Also the harmonic mode capabilities are dependent on the
Short Length. Longer Short Length settings have less harmonic mode
possibilities compared to shorter Short Lengths, as more harmonics are
missing or weakened at longer Short Lengths. The 25% example would create
an instrument that cannot play in 4th, 8th, 12th mode etc. as the 4th, 8th,
12th etc. harmonics will be significantly damped. A multimode instrument
such as French Horn (up to 15 modes !) will generally have a much shorter
Short Length than a Clarinet that typically only plays in first and third
modes.

Absorption [>] <8 Breakpoints> -- This filter sets the cutoff frequency for
the harmonics that reflect back into the tube at the open (bell) end of the
instrument. The open (bell end) of an instrument causes an acoustical
reflection of the sound wave because the rapidly expanding diameter of the
tube at the bell creates a virtual nodal point similar to that at the
closed end of the tube. In physical terms, this parameter controls the
ratio of conical tube taper to bell flare:

                                  [Image]
          FIGURE 2: Absorption High. Cone tapered, minimum flare.
        -----------------------------------------------------------
                                  [Image]
      FIGURE 3: Absorption Low. Cone nearly straight, greater flaring.

This parameter has a great effect on the pitch of given notes of an
instrument, and the effect is more pronounced with higher pitches. Always
re-initialize the tuning after making significant changes in the
Absorption.

Damping/Decay [>] <6 Breakpoints> -- This parameter has two settings; one
key-on value and one key off value. It controls the damping of the sound
wave as a function of the friction between the moving air and the tube
walls. As the friction increases (the Damping/Decay decreases ), the sound
loses energy faster, thus it 'decays' more quickly (if you think of 'decay'
as in an ADSR envelope). The Decay setting gives the value for the system
while a key is being held, and the Damping setting gives the value for the
system when the key is released.

Register Key Open -- This selects the note at which the register (octave)
key opens. Again, since the VL notes' are selected via a keyboard or MIDI,
the algorithm must be programmed to know what pitches require the resister
key to be open as you play up and down the keyboard. All notes at and above
the value set here play with the 'register key open'.

Low/High Mode balance (1st Harmonic Damp) [>] <4 Breakpoints> -- This
parameter controls the loss of harmonic energy that occurs for notes at and
above the value set for Register key open. Because the VL algorithm is
based on a sax, this controls the loss of energy in all the odd harmonics .
This recreates the octave pitch shift between a given note and the same
fingered note with the register key added. In contrast, a clarinet is a
cylindrical instrument instead of a conical instrument, and it has a
different mode behavior. It's register key causes a shift of an octave and
a fifth with a given fingering because it dampens all even harmonics.


Driver Overview -

The Driver portion of the VA algorithm defines the physics of airflow into
the instrument. The algorithm takes the value of the MIDI controller
assigned to Breath Pressure (see Element Controller Page) and converts it
into 'virtual airflow' in real time as specified by these parameters. The
'virtual airflow' now exhibits the physical behaviors created by these
parameters. The 'virtual airflow' then interacts with the reed of the
instrument.

The Child Reed is the equivalent to the single reed of a sax, the lip edges
on a trumpet, the air jet over a flute, etc. It vibrates at a given
fundamental frequency in response to the ' virtual airflow'.

The Parent Reed is an extension to add the influence that a players' lip
mass has over the vibrating lip edges in brass instruments.

Embouchure Control is the realtime response of the reed and airflow
behaviors to an assignable MIDI controller -- it allows for the algorithm
to create timbre and behavior responses to changing embouchure, which is
the player's muscle tension and position on the mouthpiece. An acoustic
instrument player's realtime control of their embouchure is a major
determinant of their individual 'sound'.

Pressure Control allow for the Breath Pressure to control some of the
Embouchure Control parameters.

Beat is an extension that adds the physical effect of the oscillating reed
hitting against the mouthpiece edge.

Excitation is an extension that enhances plucked string simulations, and
defines the shape of the 'pluck'.

Compensation allows for Breath Pressure to control pitch. It 'compensates'
for the natural pressure/pitch interactions created by the algorithm.
Acoustic players are always adjusting their embouchure in realtime to
maintain stable pitch at all pressure and dynamic levels.

Pipe Length consists of the delay settings that create the tuning of each
individual note (see Tuning Page).

Driver Parameters:

Pressure Input Gain [>] <8 Breakpoints> -- This controls the gain of the
incoming 'virtual airflow' derived from the MIDI controller set to control
Breath Pressure. Simply, it allows for control over how hard or soft the
system is being blown. High values give brighter and louder sounds, but
often the harmonic details can get lost -- especially those from subtle
variations in Breath Pressure. At low values, the response is more dynamic,
but the instrument may be more difficult to play. Also, note that multimode
instruments will tend to have low values for Pressure Input Gain as it is
important not to overpower each individual mode frequency of the reed.

Pressure Level to Reed [>] <5 Breakpoints> -- This controls how much of the
reed surface area is influenced by breath pressure. Think of a player
resting their lips around the mouthpiece and reed. As the mouthpiece is
moved in and out of the player's mouth, more or less of the reed is covered
by their lip. Higher values mean that more of the reed area is exposed to
oscillate freely. In general, low values will give faster generation of the
fundamental and accurate harmonic intervals. High values will usually give
faster attacks, but also the tendency for the reed to jump to different
harmonics, and harmonic overtones have a tendency to be slightly out of
tune.

Slit Sign -- This parameter determines which direction the reed oscillates
from it's rest state. Positive (+) means that the reed oscillates by
closing towards the mouthpiece with Breath Pressure. Negative (-) means
that the reed oscillates by opening even further from the mouthpiece with
Breath Pressure -- or think of it as playing the instrument by sucking
rather than blowing !

Additional Reed Aperture -- This adjusts the at gap between the reed and
the edge of the mouthpiece when the reed is at maximum closure. Can effect
the loudness and brightness of a sound. If set too high, the system may
begin to self-oscillate even after Breath Pressure drops to zero. Not to be
confused with Slit Initial Aperture.

Slit Saturation Feedback Balance -- This controls the balance between inner
(simple) and outer (complex) versions of slit saturation feedback, which is
the mathematical portion of the model governed by the Mouthpiece Narrowness
parameter (see Embouchure Control, below). The timbre and behavior
differences between inner and outer depends on the actual values set for
Mouthpiece Narrowness -- the higher the values for Mouthpiece Narrowness,
the more distinct the inner and outer versions become. Outer is more
dynamic and tends to be better for sax-style voices.

Graham Function Argument [>] <5 Breakpoints> -- The Graham Function is
actually the mathematical description of the fluid dynamic effects of a
medium, in this case air, flowing through a gap, where there is a given
pressure differential on either side of the gap. In the VA algorithm, this
value controls how quickly the airflow reaches saturation (maximum flow
velocity) at any given Breath Pressure input. The effect of this parameter
is highly interactive with the Slit Initial Aperture, Mouthpiece
Narrowness, and Reed Flexibility in determining the airflow speed through
the system. Simply, the airflow speed through the system is a
multiplication of the air pressure and the Graham / Embouchure Control
parameters. The overall timbre and behavior of the instrument is highly
dependent on airflow speed -- the potential harmonic structure created by
the Short Length setting is highly modified by the Graham/Embouchure
Control parameters. The Short Length setting selects the possible
harmonics, and the Graham/ Embouchure Control parameters are the major
determinant of their amplitude and behavior. General suggestions will be
presented in "Modeling Tips", and "The Circle of Life."

Driver Output shut off -- When set to 'shut off' the output from the Driver
is forcibly set to zero at zero Breath Pressure input. When set to
'through', any energy from the Driver that remains after Breath Pressure
reaches zero will continue to drive the Pipe, as dictated by the max/min
values of the Damping/Decay. This can make for subtle attack variations,
especially in repeatedly tongued notes.

Driver Output Level [>] <5 Breakpoints> -- This is the overall output gain
of the Non-Linear driver. Higher values give louder, brighter sounds.
However, this parameter interacts with the Graham Function Argument
parameter in producing chaotic (non-harmonic) noise components in the
sound. Too keep the instrument 'chaos free', the sum of the Graham value
and the Driver Output value should be close to zero. If the sum is too
greatly positive, the sound will have pronounced chaotic noise components.

Parent Reed Overview -

The Parent Reed parameters are an extension in the VA algorithm to improve
the synthesis of lip reed (brass) sounds. This is because the oscillating
edges of the lips that function as the 'reed' of brass instruments are
influenced by the mass of the lips as a whole. Since the mass of the lips
as a whole is much higher than the edges, their oscillation frequency is
much lower, and typically will not oscillate to the extent to cause
resonance. This effect usually manifests as a brighter, buzzier sound, and
occasionally as subtones. In extreme settings the Parent Reed can oscillate
well into the audio range and can produce a bright, 'biting' FM like
metallic character to the tone.

Parent Reed Parameters:

Force to Parent Reed -- Controls the gain of pressure input to the Parent
Reed. Higher values tend to be brighter.

Output Level -- Controls the overall output from the Parent Reed. Higher
values tend to be brighter and attack more quickly.

Lip Collision [>] <5 Breakpoints> -- This parameter simulates the
repulsing, or rebound effect of the lips colliding against one another
while they oscillate. Low values are 'soft' collisions, like bouncing a
rubber ball on carpet, while high values are 'hard' collisions, like
bouncing a marble on a glass table. High values tend to be brighter, but
the net overall effect is dependent on the Parent Reed cutoff frequency as
well. Interestingly, with negative settings, the lips can be made to
collide through one another, yielding interesting and strange results,
especially if the Addition to Slit opening is high.

Cutoff [>] <5 Breakpoints> -- Sets the LPF cutoff frequency of the Parent
Reed. The sloped line in the key scale window is the baseline for the
fundamental frequency. The combined values for this parameter and the Lip
Collision parameter can create noise-like chaos in the system if both are
set too high.

Child Reed Overview -

The Child Reed is the equivalent of the vibrating reed of the mouthpiece.
Its thickness, flexibility, aperture, response to airflow, etc. are very
important to creating the behavior of a sound. It has both static and
dynamic characteristics, with the dynamic characteristics vitally important
to the controllability that the player has over shaping the sound, both
timbrally and behaviorally. The Dynamic parameters are grouped under
"Embouchure Control" below. The following parameters comprise the static
parameters

Child Reed Parameters:

Force to Child Reed -- Controls the gain of pressure input to the Child
Reed. Higher values tend to be brighter.

Reed Displacement Output Level -- This is the amount of energy output from
the Displacement portion of the Child Reed. The VA algorithm allows for two
different mathematical methods to determine the output energy of the Child
Reed. One type of output energy is determined by its displacement at any
given moment in time, like the stored energy in a stretched spring. The
other is the output energy determined by its momentary velocity, or
momentum. Depending on the specific type of reed being modeled, either Reed
Displacement Output or Reed Velocity Output (or some blend) would be more
accurate. Reed Displacement Output is typically used for Single Reeds and
Bowed Strings. Higher values tend to be brighter and attack more quickly.

Reed Velocity Output -- This parameter controls the output energy
determined by the velocity of the Child Reed as mentioned above. Reed
Velocity Output is typically used for Jet Reeds.

Resonance Control -- Selects keyboard tracking the mode for the Child Reed
Resonance frequency. Absolute means that the reed pitch is set to the true
values of the Child Reed Resonance frequency parameters. Relative means
that the frequency of the Child Reed tracks the keyboard, with the
fundamental pitch (Child Reed cutoff) being set relative to the overall
tube length as determined by the Pipe Length settings created in the Tuning
subpage. A by-product of this setting is the appearance and function of the
Tuning subpage Initialization window. See Child Reed Cutoff and Tuning.

Resonance Lag Time -- The controls the speed at which the Child Reed
changes frequency when different notes are selected via the keyboard. Most
useful for brass instruments where the pitch played is highly dependent on
the constantly changing tuning modes at which the lips oscillate. When set
too high, it resembles portamento.


Embouchure Control Overview -

The dynamic aspects of Child Reed behavior are defined here. The following
five parameters allow for setting low and high limits for their values as
determined by the real time position of the MIDI controller assigned to
Embouchure Control on the Controllers Page. The purpose of allowing
realtime control is to recreate the dynamic embouchure of the acoustic wind
player. The tightness of the lips and their position on the mouthpiece
greatly influence the timbre and behavior of the sound. The low limit
values are output when the controller is at minimum, with the high limit
values being output with the controller at maximum.

Each instrument behavior type -- single reed, jet reed, lip reed etc., have
very specific ways that these parameters change whether the player uses a
'loose' embouchure (controller at minimum) or 'tight' embouchure
(controller at maximum). For example, in a sax, if the embouchure tightens,
then:

Slit Initial Aperture and Mouthpiece Narrowness both decrease, as the lip
pressure closes down the reed.

Reed Flexibility also decreases because the reed has effectively become
thicker and less flexible.

Child Reed Resonance Frequency goes up because the effective length is
shorter, and thus it vibrates faster.

Child Reed Resonance Amount goes down as the reed is being somewhat damped
by the increase in lip contact area.

This is the general rule for single reeds --however this is a strict
interpretation, and you can set the algorithm to respond any way that gives
a satisfactory result. See "Embouchure Control Tables" for strict examples
of various the VA instrument types.

Embouchure Control Parameters:

Slit Initial Aperture [>] <8 Breakpoints> -- This controls the position of
the reed at neutral pressure, where breath pressure input is equal to the
back pressure in the system. If you think of a simple harmonic oscillating
system such as a pendulum, the normal neutral state is at the
perpendicular, on-axis position. Slit Initial Aperture is analagous to the
initial horizontal displacement of dragging the pendulum off-axis prior to
starting oscillation. Increasing the Slit Initial Aperture shifts the reed
more toward its maximum closure for a given pressure input. Of course, this
assumes the slit sign is (+). The inverse condition is true for negative
Slit sign.

Mouthpiece Narrowness [>] <8 Breakpoints> -- This controls the hysteresis
effect -- the tendency of the reed to 'stall' at it's physical limit of
movement. A wide mouthpiece (high values) creates less back pressure to
help open the reed, so the reed 'stalls' more giving a brighter sound. A
narrow mouthpiece (low values) creates greater back pressure, so the reed
does not have as much 'stalling' tendency because the back pressure helps
to open it after maximum closure. This gives a darker sound. Mouthpiece
Narrowness directly interacts with the Slit Saturation Feedback, which is
separated into an inner (simple) and outer (complex) feedback loop for
greater versatility, selected by the Slit Saturation Feedback Balance
parameter. Inner feedback is generally brighter, with more simple
harmonics. Outer feedback is more correct to the strict model and gives a
more complex harmonic spectra, though somewhat darker. Increasing this
parameter makes the differences between inner and outer Slit Saturation
Feedback more distinct.

Reed Flexibility [>] <8 Breakpoints> -- A more flexible reed is able to
undergo changes in its vibratory state more rapidly. Thus, as the value of
Reed Flexability increases the reed is capable of making these faster
transitions, giving a brighter sound that generally will have a faster
attack transient. Also, can be thought of as the reed thickness, keeping in
mind that more flexible reeds are thinner .

Child Reed Resonance Frequency [>] <12 Breakpoints> -- This sets the
vibrating frequency of the Child Reed. This interacts with the Short Length
setting to create the basic harmonic structure. If the Child Reed Resonance
Control mode is absolute, then this sets the true resonating frequency in
Hertz. If the Resonance Control mode is set to relative, then the values
are given in octaves relative to the 'fundamental' frequency as determined
by the Pipe Length settings. At a setting of zero (in relative mode), the
Child Reed frequency is exactly the same as determined by the Pipe Length
setting for the tube length/pitch.

Child Reed Resonance Amount [>] <12 Breakpoints> -- This sets the amount of
Child Reed resonance, or how much energy it has in it's resonating state.
This setting is important when creating mode selection via Controllers that
change the Child Reed Resonance Frequency. Even if you cause the Child Reed
Frequency to jump two octaves with the correct settings for the Embouchure
Controlling, Child Reed Resonance Frequency parameters, the mode may be out
of tune, weak or not heard if this resonance amount setting is too low. The
reed needs more energy to resonate in the higher modes. Another way to look
at it is at higher resonance settings, the Child Reed has a greater
tendency to lock to one of the allowed harmonics determined by the Short
Length. At lower settings, the reed can remain 'out of tune' relative to
the Short Length, and create out of tune harmonic structures. Also, higher
settings are brighter.


Pressure Control Overview -

Similar to the embouchure Control parameters, acoustic players naturally
change their embouchure in concert with how hard they blow. These
parameters allow for certain related Child Reed parameters to be controlled
by Breath Pressure. The parameters are variable from zero to 16. When set
to 16, then the dynamic range is the full differential between the low and
high limits. Values from 1 to 15 are preset fractions of the total dynamic
range defined by the low and high limits. These two parameter's net result
interacts with the corresponding net result of the Embouchure Control
parameters to give the realtime value for Reed Thickness and Child Reed
Resonance Amount.

Pressure Control Parameters:

Pressure to Reed Thickness -- This allows for Breath Pressure control of
Reed Flexibility

Pressure to Reed Resonance -- This allows for Breath Pressure control of
Child Reed Resonance Amount.


Beat Overview -

Beat is for simulating the sound of the reed collides with the mouthpiece
edge. It helps in making sound brighter and more buzzy. When set improperly
(or 'properly' if that's what you want !), Beat has the ability to drive
the system to such an extent to override the effect of the Child
Reed/Parent Reed drivers. Much subtlety and detail will be lost in the
sound, and it's dynamic behaviors will be severely altered. Similarly noise
and other chaotic elements that are essential to 'acoustic' sound will be
masked as well. Beat is best used in moderate amounts.

Beat Parameters:

Amount [>] <8 Breakpoints> -- This parameter controls the overall Beat
amplitude. It has both positive and negative value settings. Similar to
slit sign, negative beat values simulate vibration in the opposite
direction. In general, sounds have the same sign, + or -, for Slit Sign,
Parent Reed Lip Collision and Beat Amount. Interesting things will result
if the signs are mixed, and the sound can be very unstable and
unpredictable !! Also, Beat, like Throat, can interfere with the accuracy
of the autotuning, so final tuning must be done manually. See Pipe Length.

High Frequency Emphasis [>] <6 Breakpoints> -- This sets the tuning of the
Beat effect. Since the intent of Beat is to add high-end brightness, the
usual approach to setting the beat emphasis is to match the approximate
frequency of the Parent/Child Reed. This allows the for the addition of the
Beat brightness without having the Beat override the Child/Parent Reed
driver's behavior. If the Beat emphasis is set too low for a given Beat
level, then the Beat pitch can overwhelm the Parent/Child Reed drivers and
introduce inharmonicities into the sound.


Excitation Overview -

Excitation creates an energy 'impulse' that can be used to start the VA
algorithm oscillating without pressure input. The intent is to provide a
burst of energy to help start oscillation when breath is applied. This is
because some types of sounds require a burst of pressure to start sounding,
and may not sound or may attack too slowly with less drastic pressure
increases. Excitation mimics the 'plosive' effect (as in 'explosive') of
letting your breath exhale after momentarily letting the pressure first
build in your cheeks. Say the word "pump". The sound of the 'p' is created
because of the 'explosive' pressure that builds up as your lips do not open
until a split second after your diaphragm as exhaled. Compare the sound of
the word "pump" to the word "hump". The word "hump" sounds only with the
release of breath from your diaphragm because your lips are already open --
it's attack transient is purely dependent on your diaphragm strength.
Excitation is triggered only by key on, and consists of a variable width
pulse waveform. It's output can be applied either to the reed or to the
pipe.

Excitation Parameters:

Level to Pipe [>] <3 Breakpoints> -- This controls the amount of excitation
applied to the linear portion. This can allow for the creation of 'plucked'
sounds without any output needed from the Parent/Child Reed drivers.

Level to Reed [>] <3 Breakpoints> -- This controls the amount of excitation
applied to the reed. This helps start the Parent/Child Reed to oscillation
with an added pulse of energy.

LPF Cutoff Frequency [>] <3 Breakpoints> -- This is a standard lowpass
filter to control the brightness of the pulse waveform used for excitation.

Velocity to -Level Sens. -- This controls the key-on velocity response of
excitation output. A setting of zero means the level will not respond to
key velocity, a setting of 16 gives maximum dynamic response.

Velocity to -Filter Sens. -- The excitation lowpass filter cutoff can also
be dynamic to key-on velocity. Zero means no dynamic response, 16 gives
maximum dynamic response.

Velocity to -Width Sens. -- The excitation waveform pulse width can also be
dynamic to key-on velocity. Zero means no dynamic response, 16 gives
maximum dynamic response.

Pulse Width [>] <3 Breakpoints> --This sets the basic pulse width of the
excitation wave.


Compensation Overview -

All natural instruments exhibit pitch variations with dynamic level. One of
the first skills learned by any player is control of pitch at various
dynamic levels, accomplished by micro adjustments in their embouchure. Once
learned, this becomes essentially automatic for the player. The VA
algorithm accomplishes this automatic correction of pitch with dynamic
level with three parameters -- Compensation, Pressure to Slit sens. and
Pressure to Resonance sens. Pressure to Slit sens. and Pressure to
Resonance sens. were discussed above previously and are concerned with the
reed dynamics to pressure. Compensation controls the pitch/dynamic range
stability, and it directly changes the overall system tuning in response to
pressure input.

Compensation Parameters:

Pitch Compensation [>] <12 Breakpoints> -- It is natural for most
instruments to play a bit flat at low dynamic levels and to get more sharp
as they get louder. Setting compensation levels positive will make the
pitches at low breath pressure become more sharp relative to the maximum
breath pressure pitch. Conversely, setting negative compensation values
will make the low breath pressure pitch more flat relative to the maximum
breath pressure pitch. Depending on the settings of Pressure to Reed
Thickness and Pressure to Reed Resonance, the Beat and Throat settings and
Child Reed Velocity Output, there can be odd fluctuations both flat and
sharp in the middle pressure ranges. High amounts of Beat and Throat also
cause unpredictable low pressure tuning variations.


Pipe Length Overview -

The Pipe Length settings are the stored Total Delay tuning values for each
note. They are accessed via the "TUNE" button on the Instrument page. The
tuning window has two displays depending on the Child Reed Resonance
Control mode. In both the "Relative" and "Absolute" Child Reed mode
versions the group of black horizontal lines that slope down to the right
are the mode lines. Modes 1 through 5 are displayed. Modes 6 and above are
too close to resolve, so are not displayed though they are present. The
version displayed when the Child Reed Resonance Control mode is 'relative'
shows three red outlined boxes connected by red lines that overlap the mode
lines:

                                  [Image]

The red boxes are used to set the desired mode when initializing the tuning
table for the instrument. The additional values in the lower left corner of
the tuning window show the key number and delay value for each of the three
boxes. There are three subpages of the tuning page, selected by chossing
the desired "bullet"-- "Start Tuning" (shown above), "Edit 1 Key" and "Set
Conditions."

Start Tuning : This is where you initialize the total delay tuning table
and run the autotune function. The slope of the mode lines and the amount
of curvature is a byproduct of the Absorption cutoff frequency value, and
its only importance is that it shows that at lower Absorption values, some
pitches will sound only in higher modes because of the increased energy
loss. Set the desired ranges and modes, set the multipler (see above, Pipe
Length Overview) and click the "Init." button to initialize the tuning
table to the set parameters. Similarly, click on the "Start" button to run
the autotune function. To save the results, click on the "Go" button, and
the tuning table will be stored and you'll
exit out of the tuning page. NOTE: If you ever try to exit the tuning page
via "Cancel" and you have tuning table values that have been changed, you
will be shown a dialog to either Save or Discard the changes before
exiting.

Instruments that use 'absolute' Child Reed mode are not usually multimode
instruments, and their basic modal structure template will be first mode
and second mode. Note you cannot set the desired mode with the red boxes as
when the Child Reed Resonance Control mode is set to Relative:

                                  [Image]

The note where the mode change occurs is determined by the Register Key
Open parameter. The actual modes after initialization depends on the
Multiplier value. If the multiplier is set to 1.00 (0), then the tuning
table will initialize to first mode for all notes below the Register Key
Open note, and to second mode for all notes at and above the Register Key
Open note. If the multiplier is set to 1.98 (63), the table will initialize
to second mode for all notes below the Register Key Open note, and to
fourth mode for all notes at and above the Register Key Open note. By using
the high and low key range values and the proper multiplier values, the
initialization when in Child Reed absolute mode can create modes 1/2 , 1,
2, 3 and 4. (1/2 mode is the sub-octave below the fundamental Child Reed
frequency).

In multimodal instruments that use 'relative' Child Reed mode, the
aforementioned boxes on this subpage choose the note ranges and desired
modes, and the Register Key open parameter has no effect on the tuning
table initialization. Again, by setting the boxes on the appropriate mode
lines and choosing various multiplier values, modes 1/2 through 10 can be
initialized. Note that the line connecting adjacent boxes must be on or
slightly above the desired mode line to select that mode.

Since the Child Reed cutoff mode is used as a reference for initialization,
if you wanted to make a 'absolute' Child Reed instrument that plays in
modes 5 and above, simply change the Child Reed cutoff mode temporarily to
relative for tuning initialization, then change it back to absolute for the
remainder of voicing and final tuning.

Tune 1 Key: This page allows for tuning of any one key manually by directly
accessing the delay settings. The best way to go when final tweaking a
voice in the 'Circle of Life.' High amounts of Beat and Throat tend to
create auto tune errors and are best tuned by hand. Change the tuning of
the selected key sharp or flat via clicking on the appropriate arrow(s).
You can also autotune a selected key by clicking the "Start" button. You
may notice in Multimode voices that as you try to tune a note that is flat,
it gets sharper and sharper until just below the desired pitch and then it
jumps down suddenly in pitch -- what has happened is that it has manually
jumped into a new mode. This tells you that given key cannot be in tune to
that pitch in the given mode for the given Short Length, Child/Parent Reed
and Absorption parameters. Remember to select the "go" button to save your
changes.

Set Conditions: Sets the parameters for the autotune software. These are
too complex to go into here, suffice it to say that it is best to leave
them at their default settings. It's much easier and faster to run the
'Start Tuning' procedure and then tweak via Edit 1 Key than to try to find
the exact values that will auto tune any specific instrument. However, if
you like to experiment, you have the ability to save and load any different
settings that yield good results for you. The first time you open the
Expert Editor, there will be all zeros enetered for the Input Gain,
Resonance, PLL Gain and Input Gain of Intgrator, and the Auto-tune will not
work properly. The correct default values are:

BPF: Multiplier:
Input Gain = 91 PLL Gain = 1.00 (0)
Resonance = 8.01 (96) Input Gain of Intgrator= 1.00(0)

Enter and save these values so you have the proper default conditions.


Throat Formant Overview -

The Throat Formant parameters recreate the influence that the player's
Throat and trachea has on creating formants in the sound. Think of it as an
extra piece of pipe that can placed before the reed, and as such it
consists of a basic resonating system. Throat is very useful in creating
harmonic characteristics that make the final voice sound 'more acoustic'.
Also, it is very good at creating some of the unusual overtones present in
the upper registers of wind instruments as well as other effects such as
the 'honk' or 'growl' effect of a sax.

Parameters:

Pitch Tracking -- Selects the tracking mode for the Throat delay. Similar
to Child Reed Resonance Control mode. Fixed mode gives frequencies as
specified by parameter settings. Key Track mode gives octave values
relative to the fundamental pitch of the sound.

Pitch [>] <8 Breakpoints> -- Sets the delay parameter values that create
the Throat pitch. When in Fixed mode, the sloped line indicates the
fundamental pitch, and the true frequencies are set in Hertz. When in
relative mode, the values automatically track keyboard pitch, and the range
is +/- 2 octaves relative to the fundamental.

Intensity [>] <4 Breakpoints> -- This is similar to the Damping/Decay of
the Instrument linear system. It controls the friction of the air in the
Throat. Higher values mean less friction, and therefore a brighter waveform
will result from the Throat output.

Amount [>] <4 Breakpoints> -- Controls the breath pressure input gain into
the Throat system. High values give brighter and louder Throat waveforms
that will be output into the instrument as a whole. Remember that the
actual Throat output is controlled via the Controller Page, with Throat
level determined by either a fixed set value or dynamically modulated by a
controller.

HPF [>] <3 Breakpoints> -- Unlike the Pipe portion of the Instrument page,
the Throat system has an adjustable highpass filter to alter the
transmission of low frequencies in the Throat waveform. It controls the
cutoff frequency of the Throat waveform reflected back into the Throat
system via a highpass filter. This parameter can be also used as a
secondary tuning in conjunction with the delay settings, especially at high
input gain and Throat LPF
settings.

LPF [>] <3 Breakpoints> -- This parameter is similar to Absorption on the
Instrument page in that is sets the cutoff frequency for the portion of the
Throat waveform to be reflected back into the Throat system via this
lowpass filter. The reasoning for including both HPF and LPF control for
Throat is to allow for more specific selection of frequencies within the
Throat waveform. Generally, the HPF and LPF value will be set up to create
a narrow bandpass type filtering of the Throat waveform.


Let's Make a Model !! -

Programming a new voice will be done either from tweaking an existing voice
or creating one from scratch. From programming other synthesis systems from
scratch, you are familiar with the 'Init Voice' concept where there is a
preset condition of a very basic voice. The concept of an 'Init Voice' is
different in VA synthesis. For one, the amount of programming from default
values that must be done is significant, as some very specific programming
must be done to create a model that plays predictably in the first place!
Second, since general instrument behavioral structure can be quite
different (flute versus trumpet for instance), each behavior category
requires a different initial state to minimize the amount of programming.
For convenience, consider five different 'Init Voices' :

Init Reed (a single reed)
Init Jet (a jet reed)
Init Lip (a lip reed)
Init Bowed (a bowed string)
Init Pluck (a plucked string)

Each of these five 'Init Voices' have significantly different relationships
in the values for their Instrument Page parameters, including their total
delay tuning and modes. The five aforementioned 'Init Voices' are provided
in the VL Examples file, along with voices to use as study examples.

Regardless, if you are starting from an existing voice or one of these
'Init Voices', there are many suggestions and tips for achieving some
desired changes in the sound. They can be grouped into two groups:
Timbre-only modifiers and Behavior/Pitch/Timbre modifiers.

The timbre-only modifiers will affect the timbre only without causing any
pitch or behavior changes in the sound, and include:

Tap
Harmonic Enhancer
EQ
Impulse Expander
Resonator
These areas can cause drastic changes in timbre, yet the given behavior and
pitch will not be affected.

The behavior/pitch/timbre modifiers will affect the behavior components,
with side effects relative to the pitch and timbre, and include:

Throat
Instrument Page
Pitch EG
Embouchure EG
These are the areas that will affect the voice as a whole, in any part,
with of course the interactive 'side effects' inherent in modeling.


VL Modeling Tips & Tutorial-

The following are tips and turorial voice examples related to the five
behavior types of which VA algorithm is capable, and represent the
'scientific' approach to these types of voices.

General Single Reed Tips -- Reference voice Init Reed.

Definitive characteristics of the single reed family include moderate to
high Absorption cutoff, use Displacement Child Reed output, Child Reed
Resonance Control absolute mode, Short Length absolute mode (with a wide
range of possible lengths). The overall timbre will be created by the Short
Length and Tap settings. Since the VA algorithm is tailored for the single
reed, all of the Driver, Child Reed and Embouchure Control parameter values
will typically have a wide range of possible settings. The Feedback Balance
will usually be towards outer, but a mix of inner can make the sound
brighter. Slit Initial Aperture, Mouthpiece Narrowness and Reed Flexibility
will be in the middle parameter value ranges, as well as Child Reed
Resonance Amount. One important characteristic of single reeds is that the
reed frequency is alway slightly above the resonating frequency of the
Pipe. Therefore, the Child Reed Resonance freq. will be at or above 'the
line' (remember that the sloped line in the Child Reed Resonance Freq. key
scale window is the fundamental pitch line). The example voice Init Reed
has moderate settings for the Driver, Child Reed and Embouchure Control
parameters. Beat can be used to enhance the brightness of the sound, and to
impart some small amounts of inharmonicities in the sound. Reference the
voice Init ReedB. This is essentially the same as Init Reed except for the
addition of Beat. Notice how Beat improves the sax-like character of the
sound.

Creation of double reed voices is a timbral edit rather than a behavior
edit on the VA algorithm (remember that double reed physics cannot be
approximated by this model). Generally shorter Short Lengths, more use of
variable Tap, and use of the Harmonic Enhancer will transform any single
reed voice to a double reed timbre. Also, since the airflow is much more
restricted through most double reeds, and the instruments tend to be
straighter (rather than cones), lower values for the Damping/Decay and
Absorption cutoff are appropriate. This will usually result in a voice that
may be more difficult to play relative to the breath pressure and tonguing
dexterity needed, which is arguably the most significant behavior
difference in double reeds versus single reeds -- witness the single reed
mouthpiece oboe for the beginning player. It's timbre is virtually
identical to a standard oboe, but much easier for the beginner to play and
learn Embouchure Control.


General Jet Reed Tips -

Reference Voice Init Jet

Definitive characteristics of the jet reed family are high Absorption
cutoff, small Short Length values (either absolute or relative mode),
Velocity Child Reed output, Child Reed Resonance Control relative mode, no
Mouthpiece Narrowness, no or low reed Flexability, and minimal Child Reed
Resonance Amount. The Embouchure Controlling min/max settings for Child
Reed cutoff will be centered around or near zero. Attack is controlled by
the Slit Initial Aperture, as well as the Graham Function Argument and
Driver output settings. Slit Initial Aperture is equivalent to the blow
angle across the mouthpiece opening. Increasing the Slit Initial Aperture
increases the blow angle down into the mouthpiece. Low Slit Initial
Aperture values give more fundamental but usually a slower attack, while
high Slit Initial Aperture values give a faster attack but more noisy
non-linear chaotic components. To modify the attack speed, increase in
small amounts the Slit Initial Aperture , Reed Flexibility and Driver
output and decrease the Graham Function Argument.

The jet reeds are more tolerant to the generation of the chaos elements
when the sum of the Graham and Driver output becomes more positive, and the
Addition to Reed Aperture can control the noise components -- set negative
to remove chaos/noise, set positive to increase. There is no Beat in jet
reeds as there is no physical reed to collide with the mouthpiece, and
Throat is also not typically used, as the player is not physically coupled
with the instrument.

For a different example of jet reed mode behavior, reference the voice Jet
Mode. This voice is initialized to second mode over its entire range, so
that MW2 selects the typical mode frequencies of the Child Reed: first mode
with MW2 at minimum, second mode with MW2 centered, and fourth mode with
MW2 at maximum. Thus you can play the keys in the C5 to C5 range and have a
full three octave range with the appropriate MW2 setting.

---------------------------------------------------------------------------

General Lip Reed Tips -- Reference voice Init Lip.

The typical structure for lips reeds includes low Absorption values, low
pressure input gain and high pressure level to reed, use of the Parent
Reed, and multi-mode total delay settings. The Short Length will tend
towards lower values, and either absolute or relative mode for Short Length
are appropriate. Reed Flexability will usually be high, and the Slit
Initial Aperture can have a wide range of moderate to high settings. The
Child Reed will be in relative mode, with high Child Reed Resonance Amount
values. The Child Reed Resonance Lag Time will be in the range of 1-10 for
creation of the pitch sliding effect characteristic of quickly played notes
for the lip reeds due to the muscle response time of the player's
embouchure to select modes. The high/low limits of the Child Reed Resonance
Frequency will be very wide, centered around zero for creating the brass
mode 'shake' behavior with Embouchure Control.

There can be wide variation of settings for the Graham Function and Driver
output level to affect the overall brightness and the attack
characteristics, and these need to be adjusted inversely as noted
previously -- decrease the Graham and increase the Driver output to create
brighter timbre and faster attack. Because the lip reeds are multimodal
instruments, care must be taken with the Short Length settings as certain
modes will be very unstable or not sound with certain Short Length
settings, and may also be difficult to tune. If you have notes that play
different pitches depending on breath pressure articulations (soft gradual
attacks vs. hard tongued attacks) or fast legato runs of adjacent notes (a
scale plays different pitches playing up than playing down is a common
example), then check the Short Length settings, as well as the Parent Reed
cutoff frequency.

To accentuate the brightness of brass sounds, use the Harmonic Enhancer --
set the carrier high pass filter to about 75 or above using the 'normal'
signal source, and set the overdrive amount to about 30 or above. Make sure
the HPF settings approximately track the keyboard. The Impulse Expander can
also be used to make the sound more 'metallic.'

Beat can also be used to accentuate the brightness, but often lip reed
voices can 'misbehave' when tuning (jump modes right as you approach the
desired pitch), so add Beat with careful attention to the Beat High
Frequency Emphasis settings to keep the overall sound as in tune as
possible.

---------------------------------------------------------------------------

General Bowed String Tips -- Reference voice Init Bowed.

The characteristic of approximating bowed string behaviors is setting the
slit function parameters to mimic the bow sliding across the string. The
hysteresis must be very high , with high values for the Reed Flexability as
well (set feedback balance = inner). This creates a force interaction
between breath pressure and the reed that approaches the sawtooth shape of
the bow sliding across the string. The Pressure level to reed parameter can
be thought of as the bow angle, or how much of the bow surface is in
contact with the string.

The sum of the Graham function and the Driver output will again be
important, with the sum of these tending to be equal or less than zero if
the Slit Initial Aperture is high, or positive if the Slit Initial Aperture
is small. Again, the balance of the Graham Function with the Driver output
influences the attack -- negative Graham with positive Driver output will
attack faster than positive Graham and negative Driver output. Since
pressure level to reed will usually be in the upper half of the parameter
range, the pressure input gain should be negative or close to zero.

Settings for Damping/Decay and Absorption can be widely varied, but values
of 90-110 for the Damping/Decay and values of 60-90 for Absorption work
well. Short Length mode is best if absolute, and the values can be widely
set anywhere from 130 to 190 depending on the Child Reed / Embouchure
parameters. Child Reed Resonance Control mode should be relative, with the
Child Reed Resonance Frequency set anywhere from zero to and octave above
the fundamental. Child Reed Resonance Amount will typically be very low.

In addition, using Parent Reed output helps to define the noise-like nature
of the 'bow scrape' transient, with the Parent Reed cutoff and Lip
Collision governing the timbre of the 'scrape'. Higher Parent Reed cutoff
with higher Lip Collision will give more 'bite' to the sound. Beat can also
accentuate the 'bite' and make the sound brighter, but is best used in
small amounts so it doesn't mask the behavior.

Because such specific Instrument Page settings are needed for the bowing
behavior approximation, timbre tweaks are best left to the tap, Harmonic
Enhancer, EQ and Impulse Expander/Resonator. Tap mode should be key track.
Use roughly equal amount of linear output and tap output. Notice that the
Init Bowed example uses only EQ. Try setting the Harmonic Enhancer balance
to full wet, and turn on the Impulse Expander and Resonator and listen to
the timbre change.

For the adventurous, use Throat as a timbre tweaker, but be aware that the
low pressure behavior will may change (use BC to control Throat amount, or
you may set a fixed value).

---------------------------------------------------------------------------

General Plucked String Tips -- Reference voice Init Pluck.

This is basically a purely Pipe system sound, because once the string has
been plucked, there is no longer any interaction between the string and the
driver. The driver can either be a short transient generated by the Child
Reed, Breath Noise or Excitation. Note that the sound will always decay to
zero, regardless of the Absorption and Damping/Decay values.

There are three ways to generate the pluck. One is by Breath Noise, where
an ADSR pressure EG that decays at maximum rate to zero creates a short
'burst' of noise that excites the linear system. Another is using the Child
Reed, again controlled by a fast decaying ADSR pressure envelope. The last
is using the Excitation wave, with maximum output of excitation to pipe.
Interestingly, there is a fourth way that entails no driving source other
than the algorithm in 'transition' state itself -- If you set the pressure
input gain, slit initial aperture and Driver output level to maximum, use
an AR envelope and 127 for all rates, set the Damping/Decay key off value
to zero, and theDriver output shut off to through. As long as the system
interpolate speed is not too slow, every note on will generate enough
energy to drive the Pipe system!

The example Init Pluck has Element 1 "Excite" as an example of the
Excitation method, and Element 2 "Fricative" as an example of the Breath
Noise method. Adjust the volumes on the Common Miscellaneous page and notes
the differences -- these are the same linear system settings with the only
changes being the method used to excite the Pipe into resonation.

The Excitation method is the most versatile method to use. First, you have
greater timbral control because the overall timbre is controlled by three
areas of the model -- the Short Length, the excitation pulse width and the
Tap output settings. The Breath Noise method only has two of these
available -- the Short Length and the Tap. For the Tap parameters, use
variable mode for the Tap position fix, mix roughly equal level of the
linear and Tap output, and use the ratio of Tap position as the timbre
modifier. Second, the pulse width excitation wave has it's own dynamic
amplitude and filter control independent of the global Dynamic Filter and
Amplifier/Filter EG.

Lastly, and most importantly, the pressure EG can have sustain level with
the excitation method and the voice will still behave as plucked -- the
Breath Noise and Child Reed methods will not work with non-zero sustain.
This will allow for the creation of plucked sound that can have infinite
sustain by proper settings of Parent Reed, Child Reed and Beat. Reference
ExciteBeat. In this example, the excitation supplies the initial energy for
the attack, and Beat supplies additional energy to the system to maintain
it's sustain indefinitely. The timbre is controlled by Short Length, with
subtle attack variation by changing the excitation pulse width.

Be careful when adding Beat that the High Frequency Emphasis is set to
match the tuning and pitch tracking of the total delay settings. You'll
notice that small changes in the Beat High Frequency Emphasis will change
the tuning of notes.

   Miscellaneous Tips That Work, But With No Scientific Basis Whatsoever!

Tip #1
Faster Attacks -- First check the Pressure EG and see if Attack Rate 1 and
2 are set to maximum (127). Note that even if both are already set to 127,
the envelope can in fact be faster if you increase the attack rate offset
parameter. If this is not sufficient, then go to the Instrument Page. The
first adjustment should be to the Graham Function Argument and the Driver
output level. Decrease the Graham and increase the Driver output by the
same amount. You can also increase the Pressure level to reed. WARNING: If
the pressure input gain is already high, increasing the pressure level to
reed can actually slow the attack.

Second, you can also raise the Mouthpiece Narrowness and Slit Initial
Aperture a bit.

Third, keep in mind that shorter Short Length typically settings attack
slower than longer Short Lengths -- so you could also increase the Short
Length a bit.

Tip #2
Keep in tune -- often. It's very easy to get so involved with creating a
behavior and timbre, you gradually create a voice that is very out of tune.
Remember that the tuning software alters the Pipe system portion of the VA
algorithm. Since a given behavior/timbre/pitch relies on the relative
values of both the Pipe and Driver system parameters, tuning the voice can
destroy the behavior and timbre for which you worked so hard when it tries
to correct the pitch. Make sure you often re-initialize and re-tune the
voice BEFORE the pitch drifts too far from the desired tuning while you are
creating of desired behavior and timbre. This will avoid spending too much
frustrating time in "The Circle of Life."

Tip #3
If you ignore Tip #2, you may find that you can't tune the voice to concert
pitch via "The Circle of Life" without completely destroying it's
behavior/timbre character. Consider creating a 'key' voice, i.e. a B-flat
clarinet where the element must be transposed by Element note shift on the
Common Miscellaneous page to play in concert pitch. After all, consider how
many orchestral instruments play in keys other than C. The reason is that
physically changing the instrument to play in concert pitch ruins it's
timbre and behavior !

Tip #4
Investigate the upper modes. You can have a very uninteresting sax-like
single reed sound that magically becomes a gorgeous ethnic double reed if
you simply initialize it to play in second and fourth modes instead of
first and second modes. Also try creating a sound that plays only in third,
fourth or fifth etc. mode. Adjust the Element note shift on the Common
Miscellaneous page accordingly for proper transpose correction, if it is
needed. Reference the voices Mode 1, Mode 2, Mode 3, Mode 4. Mode 5 and
Mode 6. These have the exact same voice parameters with the exception of
their total delay / mode settings. Notice the differences in timbre and
behavior.

Tip #5
Multi-mode sounds (brass) favor low Absorption values to enhance the
selection of the various modes via the Child Reed. Remember that low
Absorption values correspond to a straighter tube with a greater bell
flare, as in brass instruments. While you may think that such low
Absorption values would result in a dark sound, the addition of Parent Reed
output, Lip Collision, and the proper values for Graham Function Argument,
Driver output, Slit Initial Aperture and Reed Flexability will restore much
brightness. Correspondingly, the Short Lengths settings will tend to the
shorter values as well.

Tip #6
Use Throat as a timbre modifier. Set a fixed value for Throat between
90-110 on the controller page. Then use the Throat Delay, Intensity,
Amount, HPF and LPF settings to alter the timbre by creating beats,
subtones and other small inharmonicities in the sound. Especially useful
for string timbres -- reference Init Bowed, and assign Breath Pressure to
control Throat, and listen to the subtle timbre change below G3. It is best
to add Throat early in voice creation after you have created a basic stable
sound and before you do your final timbre and behavior tweaking . It can
alter the behavior in unpredictable ways, so you want the Throat Formant's
contribution to the total sound while doing your holistic final tweaking.

Tip #7
Often when tweaking the Compensation to achieve proper low breath pressure
pitch tuning, you may notice that the sound may varies sharp and/or flat
while progressing from the low pressure pitch to the maximum pressure
pitch. This usually occurs when the Compensation values start getting high.
To help correct this phenomenon, note the interaction between the
Compensation settings and the Pressure to Reed Thickness and Pressure to
Reed Resonance values. Raise the Pressure to Reed Thickness and Pressure to
Reed Resonance and decrease the Compensation to help smooth out the pitch
instability. Note that sometimes the pitch/pressure behavior contribution
from Pressure to Reed Thickness and Pressure to Reed Resonance is
backwards. If so, then decrease them as well as decreasing the Compensation
values. If it is still stubborn, try a 1-2 decrement in the Short Length
and try again.

Tip #8
Since acoustic instrument resonating bodies are complex formant filters, it
is very useful to use the EQ portion of the Element Effects to set up bands
of boost/cut with very high Q settings and high boost/cut amounts. However,
be aware of overdriving the EQ -- It can impart a diffuse, fuzzy character
to the sound. To correct, lower the input into the EQ portion, and use the
post EQ boost.

Tip #9
Many acoustic instrument exhibit weak fundamental frequency energy. Also,
usage of sounds in ensemble with other sounds and instruments usually
requires 'thinning' of the sound so all work well together in a mix. Use
the EQ's high pass filter to attentuate the fundamental if necessary.

Tip #10
Tune using the BC while blowing in your breath pressure 'comfort zone'
using Edit 1 Key. This allows for you to play in-tune pitches if your
comfort zone is below maximum pressure (remember that the autotuning
software for maximum pressure). Also, this gives you the advantage to play
slightly sharp with maximum pressure, which is a natural acoustic effect.
When used in conjunction with BC controlled Throat or Scream amount, it
creates much more realistic 'overblow' effects when actually overblowing.

Tip #11
Use Breath Attack to control Scream to accentuate tonguing effects. The
Attack Time constant should be relatively fast, less than 200 msec., with
the gain at 90 or above. Set the curve relatively high so that only
hard-tongued notes bring in the effect. Set the depth relatively small as
Scream is pretty powerful -- values less than 15, but realize than this can
be higher if the Breath Attack Gain is low. Useful for adding 'splat' in
tounged brass attacks.

The last Tip
Ignore everything presented here any try random experimentation. The are
hundreds of happy accidents still waiting to happen inside of VL. The
'perfect' saxophone may be hiding within a variance of 10 values each
across a dozen different parameters from a sickly, difficult to play party
kazoo -- And no amount of logical thinking will get you there.

                           -- HAPPY MODELING ! --


The Circle of Life -

The final polishing and fine tweaking of a voice is based on trying to
maintain the desired timbre and behavior while also playing the proper
tuning. Because of the integrated pitch --> timbre --> behavior
interactions, tweaking of a voice to maintain a specific pitch with a
specific timbre with a specific behavior requires a patient, systematic
approach. The core parameters dealt with here will be Short Length,
Absorption, and Child Reed Resonance frequency along with the Pipe Length
setting. A collateral parameter set will be the Tap output parameters.

The following example is very common with VL -- you created a sound you
like for a given range of notes where the timbre and behavior is how you
desire. Unfortunately, the pitch on the given note range is flat (sometimes
quite a bit). You run the auto-tune software and one of two things happens:

A) The note is tuned, but the timbre and the behavior has changed

B ) The note cannot be tuned to the proper pitch, and still the timbre and
behavior is altered.

The approach to fix A) with a pitch that is flat is as follows. First, exit
the tuning mode WITHOUT saving the auto tuning results, then:

1) Increase the Child Reed Resonance frequency for the affected range a few
(2-3) values.

2) Decrease the Short Length for the affected range a few (2-3) values.

3) Temporarily save the voice to the MI buffer

4) Tune the notes sharper using the Edit 1 Key mode and tune manually --
play the note repeatedly as you tune so you can hear the resulting changes
in pitch, timbre and behavior. If the behavior or timbre starts to change
noticeably, back down a few tuning values and proceed to the next note(s).
When the range is finished, save all conditions and exit tuning.

5) Increase the Absorption 1-2 values.

6) Return to the tuning mode and use Edit 1 Key to again try to tune,
playing repeatedly to verify behavior consistency. If proper tuning is
achieved and the behavior is acceptable, then go to 7). If behavior is
again lost when the note(s) are tuned, then go to step 8). Save all
conditions and then exit tuning mode.

7) Adjust the ratio of Tap position and Tap output values to make any
necessary timbral correction in the given key range, and adjust final EQ
and Impulse Expander/Resonator settings if desired.

8) If behavior was again lost when pitch(es) were tuned correctly, the
difficulty lies in an overall imbalance of many other parameters. Make
small (1-3) value increases in the pressure input gain, pressure level to
reed, Driver output level and Child Reed Resonance Frequency. If Parent
Reed is being used, then decrease lip collsion and increase the Parent Reed
cutoff a few (2-3) values.

9) Return to the tuning mode, re-initialize the range, and start tuning via
Edit 1 Key, and return to step 1).

10) Repeat 1-10 as necessary.

If the original situation was that the note(s) were sharp prior to tuning,
then the same protocol applies with the opposite direction for the
adjustments of the relevant parameters.

With situation B) above, the problem will usually lie in the settings for
Throat, Beat and Parent Reed. Sounds that derive a significant portion of
their harmonic structure or behavior from these areas are often very
difficult, if not impossible, for the tuning algorithm to auto tune,
because their pitch is determined by parameters that are not tuned by the
autotuning algorithm.

You must tune via Edit 1 Key, but the important adjustments are the
settings in the Beat, Throat Formant and Parent Reed, even before you
attempt the Edit 1 Key tuning. Your goal is to get the note as close to in
tune as possible before changing the Pipe Length settings, with keeping the
behavior as near to the desired behavior as possible. The first adjustments
should be made to the Beat High Frequency Emphasis. Raising the HFE value
will make a flat note more sharp.

The same applies to Throat -- small adjustments in the Throat pitch and
LPF/HPF can help tune the note as well. The Parent Reed cutoff can be
adjusted, but it's overall effect will be limited by it's output value, and
of course, the Lip Collision amount. If you need to raise the Parent Reed
cutoff, it will be likely that you may have to lower the Lip Collision.

When these adjustments get you close, then go back and run around the
"Circle of Life" as described above.


Embouchure Controlling Tables -

Below are the general relationships for the Max/Min settings of the
Embouchure Controlling parameters. When assigning to a MIDI controller,
Tight Embouchure is at controller maximum, and the Loose Embouchure is at
controller minimum. Note there is no table for Plucked Strings, as the only
variation is the force of the initial pluck, as there is no interaction
with the system after excitiation. The plucking force is the Excitation
amount or Fricative amount. Excitation is controlled by velocity directly.
In the case of Fricative, it has it's own MIDI controller (if fricative
control balance is set to 'controller') or by velocity scaled Pressure EG
(if fricative control balance is set to Slit).

                                  [Image]

Keep in mind that the purpose of Embouchure Control is to create your
desired timbral/behavior change with controller position. Use these tables
as a guide only -- feel free to deviate to obtain your desired real-time
timbre/behavior control.

                       Yamaha Corporation of America
                             Copyright  1996
 