Skip to content

Cue sketcher — unsettling score cues

Generate original, unsettling TV and film score cues and export WAV and MIDI

Open the tool Download zip

Cue sketcher — unsettling score cues with its default settings

Screenshot of the tool as it opens; it is a real capture, not a mockup.

Thanks, your report was sent.

Report a problem or a newer version
What is it?

No name or address needed, and please leave personal details out of the note. Only I read reports, and each is deleted after 90 days; the privacy page has the details.

What it makes

An original score cue, 30 seconds to 4 minutes long, generated from a seed and built in sections: Unease, Ostinato, Build, Peak, Drop-out and a final Sting. The parts are a chanting choir and percussive vocal chops made by pushing a buzzing source through vowel formant filters, pitch-bent wails and breathy whispers, an ostinato on inharmonic prepared-piano, detuned-pluck or music-box keys, slow dissonant string clusters with glissandi, a low drone, frame drums, toms and a tuned log drum, and claps, tongue clicks, shaker and body slaps. Meters run from 4/4 with 3+3+2 accents to 5/8, 7/8 and 9/8, in Phrygian, Locrian, Hungarian minor, whole-tone, octatonic or a major scale whose notes slip. Deliberate silences and tape stops cut through the cue. Export a 44.1 kHz 16-bit stereo WAV, a format 1 MIDI file with one track per part plus tempo, time signature and section markers, or the settings as JSON.

How to use it

  1. Pick a mood preset, then press New cue until one gets under your skin, or type any word into Seed. The same seed and settings always make the same cue.
  2. Move Dread and Absurdity to set the feel: dread is low, slow and clustered; absurdity bounces, slips a semitone and grins. Voice, Warp and Density control the vocals, the tape wobble and how full the grid gets.
  3. Choose the key, tonality, meter, tempo and target length, and the sound of the keys.
  4. Click a section in the timeline or the structure list to change its type, length and which parts play, or press Vary to re-roll just that section. Hatched blocks in the timeline are deliberate silences.
  5. Press Play (or the space bar), balance, mute or solo parts in the mixer, then export a WAV to drop under picture or MIDI to orchestrate in a DAW.

Build notes

  • All sound is synthesized with the Web Audio API: oscillators, noise, formant filters, a convolution reverb with a generated impulse, a modulated tape delay and a compressor. There are no samples, and nothing is fetched or sent.
  • Motifs, chants and rhythms are generated from the seed with odd meters and off-kilter scales, so each seed gives a new original cue rather than a copy of an existing score.
  • The default one-minute cue renders to WAV in a few seconds; a four-minute cue takes proportionally longer.
  • MIDI tracks use General MIDI programs as placeholders; drums and hand percussion are on channel 10. Pitch glides, vowel shapes and tape warp are audio-only.