Orchestral Programming

A Brief Guide

Introduc)on
Hello everybody! My name is Guy Michelmore. I
am a film and television composer and I’ve been
working with orchestral samples for most of my
working life. Ge@ng even the best sample
libraries to sound realisAc takes a great deal of
skill and knowledge, and that is what I’m going
to aCempt to impart to you in this micro course.
We will look at each of the five most important
aspects of orchestral programming in turn:

• Day 1: ArAculaAons
• Day 2: Dynamics & Controller
• Day 3: Legato
• Day 4: Size of the SecAon
• Day 5: The Room, Ambience and Mic PosiAons

Day 1: Ar)cula)ons
If you take a teacup and start hi@ng it with a pencil you will very quickly hear
how many different sounds can be created just by taking something as simple
as a piece of crockery and an item of staAonery and banging them together.
Now imagine taking something as sophisAcated and complex as a violin and,
rather than hi@ng it with a pencil, playing it with a bow and seeing how many
different sounds you can create from that instrument. As a music lover you will
be very familiar with the range of sounds that an instrument can produce, and
will be expecAng the subtlety, the emoAon, the different intonaAon, the
dynamic contrast that a professional player brings to every line of music they
play.

Enter the world of Digital Audio WorkstaAons (DAWs) and sample libraries.
How on earth are we going to recreate that kind of subtlety using digital
samples?
What exactly is a sample? Put simply it is a digital audio recording of a
musician playing one note. If you try pu@ng that into a sample player within
your DAW you will quickly find the limitaAons of working with a simple
sample. Firstly, as soon as the sample is played more than a tone away from its
original pitch it sounds extremely arAficial and unrealisAc. Secondly when you
start playing the sample louder or quieter, although the volume goes up and
down, it doesn’t sound realisAc because a real instrument has a different
Ambre depending how loud it is played. Add to this the problem that when you
play a note repeatedly on a real acousAc instrument it will introduce subtle
variaAon in the note whereas the digital recording remains absolutely idenAcal,
and you start to see why sample instruments are incredibly complex creatures
that require hundreds, thousands, someAmes even tens of thousands of
individual recordings to make them come to life.

‘ArAculaAon’ refers to the different ways of playing a note on an instrument.

So for example, on a string instrument you might play legato, arco, marcato,
spiccato, staccato, tremolando, pizzicato etc. There are dozens and dozens of
other arAculaAons and methods of playing, all of which can be found in good
sample libraries.

In order to make a line of music sound realisAc you have to use mulAple
arAculaAons: legato for the lyrical secAons, spiccato and staccato for the short
detached secAons, etc. Each one has to have its own specific place and
replicate as closely as possible the way a live musician would perform that line
of music. However, just using separate arAculaAons will not take you all the
way to your goal. You will need to adjust the Aming and the volume of each
individual notes to make them gel together so that they feel like a cohesive
line of music, not like the playback of individual sounds from a piece of
socware.

Changing Ar)cula)ons

One of the great challenges in orchestral

programming is finding a convenient way
of changing arAculaAons in real Ame.
There are a number of different
approaches to this. The most common is
to use something called Keyswitches. A
Keyswitch is simply a note outside the
range of the instrument that can be used
as a trigger to change arAculaAon, so for
example from a long note to a short one.
Most common keyswitches start at C0 (3 octaves below middle C) and then
ascend chromaAcally. Spitfire start at C-2, 5 octaves below middle C.

Separate Tracks

Another approach is to put each individual arAculaAon on a separate midi

track. This means loading individual arAculaAons into your sample player as
opposed to one mulA-instrument. Some libraries, like Spitfire’s BBC Symphony
Orchestra, only really support Keyswitches as all the sounds are loaded in a
single patch. Other libraries will give you individual arAculaAons as individual
presets. This allows you to load them and assign them to different MIDI tracks
within your sampler.
Ar)cula)on IDs and Expression Maps

A third way is to play in the notes then use ArAculaAon IDs (Logic) or
Expression Maps (Cubase) to assign arAculaAons to individual notes in the
piano roll editor. This can work well, but doesn’t allow you to play the notes in
and hear the arAculaAons as they will be in the final piece at the same Ame.
Also, se@ng up ArAculaAon IDs and Expression Maps can be extremely fiddly.
Most people will buy a third-party library of ArAculaAon IDs or Expression
Maps as presets, but this is an addiAonal expense.

Conclusion

Think like a musician. A sample program is an instrument just like a flute or a

violin and needs to be performed, not just programmed. Imagine you are a
musician performing a parAcular phrase of music; what arAculaAons would you
need in order to be able to make that phrase sound effecAve? Analyse the
music and then work out what sample presets and arAculaAons you need to
use to bring that phrase to life.

Now go and try this for yourself!

Day 2: Dynamics and Controllers

A line of music without dynamics is like food without seasoning. Changes in
volume, subtle and carefully considered, inject emoAon into music. Musicians
and performers know this, and a lot of the Ame will add these dynamic
contrasts completely automaAcally without conscious decision. As an
orchestral programmer working with samples you need to think like a
performing musician and work out how you can best use dynamic changes to
add expression and emoAon to your music.

Most sample developers use the

modulaAon wheel (shortened to
‘mod wheel’) to control not just
volume but also the Ambre of the
instrument; a French horn played
pianissimo has a very different tone
to one played forAssimo, for
example. So when you push the
mod wheel forward, you are not
just increasing the volume, you are
also changing the Ambre of the
instrument to match the dynamic. With string instruments ocen the vibrato
will also increase with the volume. So the mod wheel (CC1) is ocen a one-stop
shop for being able to inject greater emoAon into your music. You should be
able to play with your right hand while moving the mod wheel with your lec
and, in real Ame, be able to create an emoAonal and convincing performance.

Here is a handy spreadsheet of all the ConAnuous Controllers and what they
do.

MIDI CC PURPOSE MIDI CC PURPOSE

0 Bank Select (MSB) 76 Sound Controller 7

ModulaAon
1 77 Sound Controller 8
Wheel (MSB)
Breath Controller
2 78 Sound Controller 9
(MSB)

3 Undefined (MSB) 79 Sound Controller 10

4 Foot Pedal (MSB) 80 General Purpose

Portamento Time
5 81 General Purpose Controller
(MSB)

6 Data Entry (MSB) 82 General Purpose Controller

7 Volume (MSB) 83 General Purpose Controller

8 Balance (MSB) 84 Portamento CC Control

9 Undefined (MSB) 85 – 87 Undefined

10 Pan (MSB) 88 High ResoluAon Velocity Prefix

11 Expression (MSB) 89 & 90 Undefined

Effect Controller 1
12 91 Effect 1 Depth
(MSB)
Effect Controller 2
13 92 Effect 2 Depth
(MSB)
14 Undefined (MSB) 93 Effect 3 Depth

15 Undefined (MSB) 94 Effect 4 Depth

General Purpose
16 – 19 95 Effect 5 Depth
(MSB)

20 – 31 Undefined (MSB) 96 (+1) Data Increment

LSB Controller for

32 – 63 97 (-1) Data Decrement
0-31

Non-Registered Parameter
64 Damper Pedal on/off 98
Number LSB (NRPN)

Non-Registered Parameter
65 Portamento on/off 99
Number MSB (NRPN)

Sostenuto Pedal on/ Registered Parameter Number

66 100
off LSB (RPN)
Registered Parameter Number
67 Soc Pedal on/off 101
MSB (RPN)
102 –
68 Legato FootSwitch Undefined
119

69 Hold 2 120 All Sound Off

70 Sound Controller 1 121 Reset All Controllers

71 Sound Controller 2 122 Local on/off Switch

72 Sound Controller 3 123 All Notes Off

73 Sound Controller 4 124 Omni Mode Off

74 Sound Controller 5 125 Omni Mode On

75 Sound Controller 6 126 Mono Mode

76 Sound Controller 7 127 Poly Mode

AIaching a CC Controller

If you want to aCach a controller to a control inside a sample player, right click
or control click the control in quesAon. This brings up the MIDI learn box. Click
on “MIDI Learn” and wiggle the controller you wish to aCach. The controller
should now be aCached.

Now go and try this for yourself!

Day 3: Legato
The legato funcAon inside most sampled instruments allows players to
simulate the effect of a musician moving smoothly from one note to another.
Normal sample programs contain just a single note, whereas legato programs
include a Any recording of the sound of the musician transiAoning between
the two notes. This is very complicated to program and with every generaAon
of virtual instruments the effect becomes more sophisAcated and easier to
use.

Legato programs are complicated, and if you are used to using the purge
funcAon in Kontakt for example where you can empty the computer memory
of samples and just stream sounds from an SSD, this does not work as well
with legato samples.

Not every line of music needs to be played legato, so don’t overuse them.
Frequently you need to make use of sforzando, marcato, as well as just normal
sustained notes. Overusing legato makes the music sound less effecAve and
lose some of its edge because you’re not taking full advantage of a clean
aCack, which many live musicians would use more frequently. Legato is ocen
parAcularly overused with brass samples.

More recent legato samples take account of the speed of the transiAon. If you
are playing fast you would expect to hear a blurred legato sound, very much
like what you would hear from a musician playing a fast arpeggio or run.
However, if you play slowly on a string instrument for example you might
expect to hear something closer to portamento. SophisAcated virtual
instruments are able to detect the speed at which you are playing and adjust
the speed of the legato accordingly.

The transiAon sample is triggered when the player overlaps two notes and
conAnues to hold down the outgoing note while pressing the next note. If the
two notes don’t overlap, no legato effect is heard.

Now go and try this for yourself!

Day 4: Sec)on Size
Not every phrase or piece of music is best suited to having the maximum
number of players play every single note. SomeAmes a full symphonic violin
secAon of 18 players will be the best sound for the music, but ocen a chamber
secAon of say six or eight players may be more suitable, or even just a solo
player. Choosing the size of the secAon and the number of players required is
an important part of making your music sound more realisAc.

The Pipe Organ Effect

In the real world, if you had three French horns required to play a triad, you
would assign one instrument to each note. In the sampled world, if you have a
sample of three French horns playing together and you play a chord, you will
end up with three horns playing each note, so nine horns in total. The result
can be very unrealisAc and someAmes leads to a weird ‘phasey’ sort of sound.
This is known as the ‘pipe organ effect’.

Layering

While many instruments are beCer suited to being performed only by solo
players, like many brass and woodwind instruments, strings combine very well.
If you wish to layer a number of different samples together you can achieve
some interesAng creaAve effects; for example if you mix a solo instrument
with a larger secAon, and get the balance just right, you can decrease the
overall size of the secAon in the mind of the listener. Ge@ng the size of the
string secAon right is parAcularly important when the music is harmonically
complex; if you have chords with more than 3 or 4 notes, you would not want
16 violins playing each note. The correct use of a chamber string secAon can
greatly enhance the realism of the music.

Layering is also very useful when you wish to overlap different arAculaAons.
For example you may want to put a marcato sample overlaying a sustain
sample to give a stronger start to the note. You may equally choose to use a
sample dynamic movement, like a crescendo decrescendo, which you can lay
over the end of a sustained sample.

Conclusion: don’t assume that the largest secAon is always the best for your
sampled mockup. Ocen working with solo or chamber secAons is a far beCer
choice.

Now go and try this for yourself!

Day 5: The Room, Ambience and Reverb
The sound of your sample library is very much the sound of the room in which
it was recorded. If you buy a very ambient library like Spitfire Audio’s
Symphonic CollecAon, it will have the sound of the Air Lyndhurst Hall. If you
invest in EastWest Hollywood strings, it will have the much Aghter sound of
the EastWest studios in Hollywood. The choice of room and the choice of
ambience is a crucial and ocen deciding factor when it comes to selecAng the
right sample library for you. While it is more than possible to combine different
libraries together recorded in different locaAons, in my experience it ocen
works best if one main library is all recorded in the same space.

The Control Room at Air Lyndhurst Studio 1

Some Libraries and where they were recorded:

Spitfire Audio Symphonic Series - Air Lyndhurst Studio 1

Spitfire Audio Studio Series - Air Lyndhurst Studio 2
Berlin from Orchestral Tools – Teldex, Berlin
VSL Synchron – The Synchron Stage, Vienna
Early VSL – The Silent Stage Vienna
East West Quantum Leap Hollywood Series – East West Studios LA
Cinesamples – The MGM Stage

Adding Reverb

I will frequently use two layers of reverb with most of my sampled

orchestraAons. The first layer is a convoluAon or sampled reverb which places
the instruments in the same room. The second is normally an algorithmic
reverb which I use as a finishing reverb to give a sheen to the final mix. Some
good convoluAon reverbs are AlAverb by Audio Ease and Spaces by EastWest.
They tend to take more CPU but provide a very realisAc representaAon of a
space. As such they are very useful for taking dry samples and placing them in
a more ambient se@ng. Algorithmic reverbs include Lexicon PCM and
CinemaAc Rooms by LiquidSonics.
There are dedicated room simulaAon plug-ins like MIR from VSL, which allow
you to place instruments at specific places in a room and simulate the
acousAcs you would hear. Personally I have not found them any more effecAve
than using normal reverb, but some composers swear by them.

Always use bus-based reverb, not Insert reverb, when applying the same
reverb to mulAple tracks. ConvoluAon reverb in parAcular can be very draining
on your CPU, and you do not wish to add this to every single instrument track.
You will however need separate reverbs for each stem if you are outpu@ng
your music as stem files.

Now go and try this for yourself!

Conclusion

This has been a very brief look at what is a very complicated and detailed
subject. I have been working with orchestral samples for well over 20 years
and I sAll have a great deal to learn. We are very fortunate that so many
sample developers conAnue to produce such amazing products for us. Not all
of them are astronomically expensive, and some of the older and cheaper
samples are every bit as good as the latest and most expensive ones. Acer all,
a sample is simply a recording of a musician playing a note; if that musician had
a parAcularly good day in 1998 it is sAll a great recording today. However, the
technology marches on and becomes ever more sophisAcated, which makes it
ever more easy for us to create realisAc mockups of live musical performances.

The one thing you should remember above all is to think like the musician who
would perform that piece of music live. Every note, every arAculaAon, every
dynamic movement must contribute towards the expressive and emoAonal
impact of the music. Look at each phrase in turn, each line, and then bring all
the parts together in one saAsfying whole.

Orchestral programming is complicated, difficult and Ame-consuming. In a live

performance musicians listen to each other: they synchronise their Aming and
their intonaAon perfectly. In the sampled orchestral universe, that is not
possible. You as the musician and programmer have to do the thinking for
every single musician, making sure that their performance is perfectly aligned
to give the impression of a live orchestra playing music. When this all comes
together it is an incredibly saAsfying experience, but I remember how
frustraAng it is and how long the journey is in order to get to that point. If you
work hard at it, spend the necessary Ame and listen carefully to what a real
live orchestra actually sounds like then you will eventually produce beCer and
beCer sampled mockups of your orchestral composiAons.
Want to learn more?
If you want to conAnue your studies, you should check out our course
Sampled Orchestra)on in a Weekend.

Whether you are working in Logic, Cubase or any other Digital Audio
WorkstaAon (DAW), this course will unlock the techniques used by top
working composers and song-writers.

• A comprehensive guide to working with sampled instruments.

• Over 6 hours of video tutorials in both Logic and Cubase.
• A course text to support the video tutorials.
• Lifelong access and 30-day money back guarantee.

The detailed videos explain a key concept like the use of mulAple controllers,
layering, and arAculaAon switching. Watch live scoring examples as we bring
all those techniques together. Learn in both Logic and Cubase, but the
techniques can be applied to any DAW. You can also download the DAW file
or MIDI and see exactly how it was done.

• Learn to breathe life into your sampled orchestraAons so they don’t sound
flat and arAficial.
• Discover techniques for ge@ng the best results from inexpensive libraries.
• Understand how sampling technology works to get the most out of your
computer’s resources.
• Learn alternaAve ways of working you may not have considered.