← Guide / Genres / Classical & cinematic
Classical & cinematic
Orchestral and score-adjacent writing where instrumentation, dynamic shape and room sound carry the emotion instead of a drum loop.
Sound and palette
Classical and cinematic prompts live or die on instrument naming. A model will happily give you a vague "epic orchestra" wash, but the results get specific the moment you name sections and registers: low strings, divisi violins, solo cello, woodwind doubling, horn section, timpani. Think in families first — strings, woodwinds, brass, percussion, keyboard — and then pick one or two voices inside a family to carry the melody. A prompt that says solo cello over sustained low strings produces a far more coherent arrangement than one that says full orchestra, because you have told the model which line matters.
The second palette decision is era and idiom. "Baroque" implies harpsichord continuo, terraced dynamics and busy counterpoint. "Romantic" implies large string sections, rubato and lush chromatic harmony. "Modern film score" implies sustained low strings, ostinato figures, taiko or sub-hits, and often a hybrid layer of synth pads and processed percussion. These labels do useful work in three or four words, so spend them deliberately rather than stacking all of them at once.
- Strings: sustained pads, tremolo, pizzicato, col legno, harmonics — each is a different texture, not a synonym.
- Woodwinds: flute and clarinet for lightness, oboe and cor anglais for plaintive melody, bassoon for weight.
- Brass: horns for warmth and heroism, trumpets for fanfare, trombones and tuba for menace.
- Percussion: timpani rolls for tension, taiko and gran cassa for impact, suspended cymbal swells for transitions.
- Hybrid colour: felt piano, granular pad, processed choir, sub-bass hit — modern score signatures worth naming explicitly.
Rhythm and tempo range
Orchestral music is less grid-locked than pop, but a BPM number still steers the model usefully. Reflective cues sit around 55–72 BPM, where a slow harmonic rhythm lets each chord register. Processional and heroic writing sits near 80–100 BPM. Action cues — ostinato strings, driving low brass — run 120–150 BPM, and the perceived speed often comes from subdivision rather than tempo, so a 120 BPM cue with sixteenth-note string ostinato feels faster than a 150 BPM cue in half notes.
Metre is worth naming when you want it. Asking for 3/4 produces waltz-like lilt; 6/8 produces a rolling, maritime feel; 5/4 and 7/8 produce unease that suits thriller and sci-fi material. If you leave metre unspecified you will generally get common time.
| Cue type | Tempo | Rhythmic signature |
|---|---|---|
| Elegy / reflection | 55–70 BPM | Slow harmonic rhythm, sustained strings, little percussion |
| Processional | 80–96 BPM | Steady timpani pulse, dotted brass figures |
| Adventure | 100–126 BPM | Eighth-note ostinato, horn calls over moving bass |
| Chase / action | 130–152 BPM | Sixteenth-note strings, taiko hits on downbeats |
Vocal behaviours
Most cinematic prompts are instrumental, and saying so matters — add instrumental or no vocals or the model may introduce a lead singer you did not ask for. When you do want voice, treat it as an instrument with a role. A wordless soprano vocalise sits above the orchestra as melody. A mixed choir singing sustained vowels functions as a pad. Latin or neutral-vowel choir keeps attention on the arrangement rather than the lyric.
If you want a sung lyric, keep the vocal simple and let the orchestra breathe between phrases — long, unhurried lines with space at the ends. Dense orchestration plus a busy vocal usually collapses into mud.
Production character
Room is the single biggest production variable. A scoring-stage sound means a long, warm tail, distant mic perspective and plenty of pre-delay; a chamber sound means close mics, audible bow noise and a short decay. Say which you want — large scoring stage reverb or intimate chamber recording, close mics — because the default is usually somewhere in between and sounds generic.
Dynamic language also shapes results: starts sparse, builds to full tutti gives the model an arc to follow. Terms like wide stereo image, warm analogue tape, or clean modern digital, high headroom steer tone. For hybrid scores, name the processing directly: granular textures, reversed swells, sub-bass impacts.
Common pitfalls
- Stacking eras — "baroque romantic modern hybrid" cancels itself out. Pick one.
- Asking for "epic" without naming instruments; you get a generic trailer wash with no melody.
- Forgetting instrumental and getting an unwanted vocal over your cue.
- Requesting full tutti from bar one, which leaves the piece nowhere to build.
- Over-listing: twelve instruments in one prompt means none of them is featured.
The fix for most of these is subtraction. Name one lead voice, one supporting texture, one rhythmic engine, one room. Anything beyond that is decoration, and decoration is what the model is best at inventing on its own.