A film composer writes a single melody. Two hours later it is played by a hundred people, and it matters which of them: the same line on a clarinet sounds warm, on a trumpet bright, and on a flute in its lowest octave it can barely be heard over the strings. A pop arranger faces a smaller version of the same problem. The chorus must feel bigger than the verse, and nobody wants to add more notes, so the arranger adds people.
That is the craft of this subject. Composition is deciding what the notes are. Arranging is re-casting music for a particular group. Orchestration is choosing the instruments and registers that carry each idea. They overlap, and they rest on three habits: know the ranges, know the transpositions, and listen to how a chord is spaced.
This explains how writing for instruments works. Real ranges and balance depend on the instrument, the player and the room; a teacher or a player is the final authority.
Every instrument has a range, from its lowest to its highest playable note, and within it registers that sound different. The B♭ clarinet is the textbook case: its low register (the chalumeau, written E3 to B♭4) is dark and hollow, its middle (the clarion) bright and singing, and the notes above that thin and shrill. A flute is breathy and soft in its lowest octave and piercing in its highest. Strings can sustain and play several notes at once; wind players need to breathe, so a phrase needs room for air. Brass players need to rest their lips.
The chart shows sounding ranges (what the audience hears) as numbers and bars. Sources differ by a tone or two, particularly at the top, and the very highest and lowest notes are for soloists, not orchestra parts. Treat the chart as the safe middle of what these instruments do.
Approximate practical sounding ranges, rounded, assembled from standard references and instrument articles (they differ by a tone or two); the English horn and piccolo are derived from their transpositions. C4 = middle C = MIDI 60 = 261.6 Hz. The tone you hear is a plain synthesised one, not the instrument. Letters (C D E) and fixed-do syllables (do re mi) name the same notes.
Many instruments are transposing: the note the player reads is not the note that comes out. On a B♭ clarinet, a written C sounds as the B♭ below it. There are three reasons. A family of instruments (clarinets of different sizes, saxophones) can share one set of fingerings if each is written as if it were in the same key. Brass instruments have historic crooks and slides that changed their home key. And some instruments are written an octave away from what they sound: the double bass sounds an octave lower than written, the piccolo an octave higher, to avoid a thicket of ledger lines. The score a conductor reads may be in concert pitch (everything as sounding) or transposed (each part as the player reads it); orchestral scores mostly print transposed parts.
The practical rule for the writer: to make a B♭ trumpet or clarinet sound a concert C, write a D; to make a horn in F sound a C, write a G a fifth higher. Mistakes here are common and wonderfully audible.
Top staff: the note the player reads. Bottom staff: the note the audience hears. Note spelling follows a simple flats-for-black-keys rule; a real part would be spelled for its key.
A voicing is how a chord's notes are spread across registers. The same three letters, C E G, can be a bright triad on a violin or a smear on a tuba. The reason is acoustic: every real note carries overtones (see Acoustics, Tuning & Timbre, 35/68), and the ear can resolve two close partials only if they are far enough apart. Low notes are close together in frequency, so their overtones clash and beat. Writers use a rule of thumb: wide at the bottom, close at the top, the way the harmonic series itself is spaced, and keep thirds out of the deep bass. Where exactly the limit falls depends on the source and the instrument; the model below shows why the limit exists.
The model is a published dissonance curve of the Plomp and Levelt type (in William Sethares's parameterisation), applied to a note with six overtones each at one over its number. It computes the roughness between every pair of partials from different notes of the chord. It models roughness, not beauty, and the numbers are in arbitrary units.
Each note has partials 1 to 6 with amplitude 1/n. Roughness d = a1a2[e−3.51 s Δf − e−5.75 s Δf] with s = 0.24 / (0.0207 fmin + 18.96), summed over partial pairs from different notes. Open triad = root, fifth, then the third an octave up.
Most composers do not invent a long tune; they invent a small idea, a motif, and ask what it can become. The standard moves are old and few: transposition (repeat it higher or lower, called a sequence when it steps along a scale), inversion (turn the intervals upside down), retrograde (play it backwards), augmentation and diminution (double or halve the note lengths). Beethoven's Fifth Symphony (1808) is the classic demonstration: its four-note opening is reshaped through the movement until a listener hears it everywhere. The same toolkit sits under fugues and under twelve-tone writing, where inversion and retrograde are the rules rather than the options.
An original five-note motif, written for this page. Inversion is about the first note; the sequence is a whole step up (two semitones); augmentation doubles the lengths. Note lengths are given in the readout rather than drawn.
Texture is how many lines there are and how they relate. A single line is monophony; many players on the same line with small differences, as in much gamelan and folk music, is heterophony; a tune over supporting chords is homophony; independent interweaving lines are polyphony. Density is how many notes sound and how widely they spread. Arrangers treat both as dials. A verse might be a voice, a bass and a quiet pad; a chorus adds an arpeggio, drums and a higher octave, and a final chorus often adds a countermelody. Subtraction works as well as addition: a sudden break in density is one of the strongest effects in an arrangement, and many orchestral climaxes work because everything before them was thin.
Four bars of I-vi-IV-V in C (C, Am, F, G) at 100 beats per minute, looped. The melody is original to this page. Toggle layers on and off while it plays to hear the density change; the readout counts notes per bar and the register span of the active layers.
When two instruments play the same line, the result is not just louder: the sound is a new colour. Orchestrators double at the unison, at the octave, and sometimes at stranger intervals. The most famous example is Ravel's Boléro (written in 1928 for the dancer Ida Rubinstein and first performed at the Paris Opéra on 22 November 1928): over a snare-drum rhythm repeated 169 times, two 18-bar melodies alternate, and in the middle of the piece the tune is given to a horn and, at the same time, to a celesta and two piccolos playing in other keys (G and E major). Those other-key lines land on the overtones of each melody note, so they fuse with the horn into one bright, strange instrument, as an organ stop adds partials to a pipe. The tuning is not exact: the equal-tempered G is about 2 cents flat of the true third harmonic, and the E about 14 cents sharp of the true fifth.
The tune is the page's own motif in low register; each added line is a quiet tone at that equal-tempered interval above, the way an orchestrator doubles. Interval sizes in cents are computed against the exact harmonic.
A full orchestral score lists the instruments in a standard order from top to bottom: woodwinds, brass, percussion, then strings. A good page is mostly decisions about what not to write. Check the range and the register. Give wind players a place to breathe and brass players a rest. Remember that loudness is not equal: a trumpet or a trombone can cover a section of strings, and the lowest octave of a flute disappears behind a string section. Match the technique: strings can play two notes at once and change strings, but fast wide leaps across strings are awkward; a harp cannot easily play every chromatic note quickly. Many composers mock up the piece on a computer first, which is useful, but a mock-up forgets breath, bow and balance, so the best test is to ask a player to read it.
Writing for instruments is a set of physical facts (range, transposition, spacing) and a handful of habits (develop a small idea, vary the texture, double with intent). Learn the facts so your choices are real choices.
Try this: listen to a piece for orchestra or a pop chorus and count the layers at the start and at the end of one section. When the sound grows, can you say which layer entered? When it feels thin, what was taken out?
Sources. Ranges: instrument articles (flute, oboe, clarinet, bassoon, horn, trumpet, trombone, tuba, violin, double bass, cello), rounded; transposition values: the transposing-instrument article and standard orchestration texts. Dissonance model: Plomp and Levelt (1965) in Sethares's parameterisation (constants checked on Sethares's page). Boléro facts: encyclopaedia entry; interval sizes computed. Motif transformations are standard theory. Beethoven's Fifth premiered in 1808. Not a substitute for a teacher or a player.