Real-time procedural creature and monster voice synthesizer. Every growl, roar, and howl is generated fresh on the fly. 50+ presets, infinite variation. Pure C#, no plugins, no middleware.The demo scene is built using URP but the core synthesis is render pipeline agnostic, it will work in all render pipelines.Stop recycling the same monster roarsEvery game with creature audio has the same problem: a library of recorded growls that players hear three times and recognise forever. No matter how many variations you record, loops repeat. Immersion breaks. Your dragon sounds like every other dragon.Procedural Monster Synth fixes this permanently. It generates creature vocalisations in real-time, directly inside Unity's audio system. Every sound is synthesised fresh, every time. No two roars are ever identical.One engine. Every creature.A single unified synthesis engine models a complete creature vocal system, from the larynx to the lips. A massive kaiju and a skittering insect both come from the same synthesizer, just at different settings. Adjust a handful of parameters and your wolf becomes a cave troll, your dragon becomes a banshee, your alien becomes a demon.MIDI pitch is a creative tool, not just a tuning knob. Play any preset at different notes and discover entirely new creatures. A Dragon Roar played two octaves higher becomes a wyvern screech. A Goblin Cackle dropped an octave becomes a troll. Every preset contains dozens of hidden voices waiting to be found.50+ ready-to-use presetsShip creature audio today with over 50 curated presets across seven categories:Hero / Animal - Lions, tigers, bears, dragons, kaiju, banshees, cave stalkersAction / Combat - Rage bellows, beast charges, impact grunts, war pulses, quick snarlsAdventure / Fantasy - Goblins, trolls, wyverns, giant spiders, wraiths, bog creaturesSci-Fi - Robot voices, alien comms, mech servos, energy drones, xenomorph hissesHorror - Demon whispers, possessed voices, wendigo howls, abyssal leviathans, swarm buzzCinematic / Atmosphere - Trailer braams, tension risers, sub-drones, cathedral drones, wind effectsGeneral Game / Utility - Power downs, alarm klaxons, magic surges, steam vents, acid sizzlesEvery preset is a ScriptableObject asset. Duplicate any preset and make it your own.Three ways to design soundsThe Sound Stage editor gives you three powerful approaches to creature design:Parameters Mode - Full access to every control in the synth, organised by domain. For when you need surgical precision over a specific formant, a distortion curve, or an envelope shape.Macro Mode - 17 high-level sliders grouped into Anatomy macros (Larynx Pitch, Cord Tension, Chest Size, Throat, Maw Open) and Character macros (Size, Aggression, Breathiness, Instability, Cavernousness). Load a preset and push a few sliders to reshape the creature in seconds. Macros stack additively, so Chest Size and Aggression both contribute without fighting each other.Grid Mode - A 2D XY pad that blends between four presets in real time. Assign a Lion Roar to one corner, a Banshee Howl to another, a Kaiju to a third, and a Giant Spider to the fourth. Drag across the pad and hear creatures that would be impossible to reach by adjusting sliders alone. When you find a blend you like, bake it into a new preset with one click.Real-time parameter control from gameplayEvery parameter can be updated on a sounding voice while it plays. Tie creature vocalisations directly to gameplay state:A monster that growls louder and more aggressively as its health dropsA dragon whose roar deepens as it charges a fire breathA creature whose voice morphs as it transforms between formsSmooth pitch glide when a voice shifts between notesNatural breath exhale on releaseWhen a creature stops vocalising, it does not just cut to silence. The Release Tail captures a reservoir of the creature's own breath material while the note is sounding, then plays it back as a natural exhale after the voice stops, through the same vocal tract, so it sounds like the same creature breathing out. A small detail that makes a surprisingly big difference to realism.Polyphonic and layerableThe engine supports configurable polyphony with intelligent voice stealing. Layer multiple presets simultaneously for rich, complex creatures: a deep sub-heavy preset for body, a mid-range preset for vocal character, and a noise-heavy preset for breath texture. Trigger all three on the same note event.Platform SupportWorks on all platforms supported by Unity except WebGLStandalone (Windows, macOS, Linux), Mobile (iOS, Android), ConsoleNo native plugins, no platform-specific binaries, no middleware dependenciesRequirementsUnity 2022.3 LTS or newer (URP recommended)No additional packages or dependencies requiredArchitecturePure C# - the entire synth runs as managed C code inside Unity's audio thread via `OnAudioFilterRead`. Nothing to compile, no DLLs to manage, no plugin compatibility headachesZero-allocation audio thread - no garbage collection pauses, no audio glitches from GC spikes. The audio thread is completely allocation-free by designScriptableObject-based - engines and presets are ScriptableObject assets. Drag-and-drop workflow, full Inspector support, easy serialisationLock-free architecture - game thread and audio thread communicate through a lock-free command queue with sample-accurate scheduling. No mutexes, no blocking, no priority inversionWhat You Get1 unified procedural synthesis engine (`MonsterEngineSO`)50+ curated preset assets across 7 categoriesFull Sound Stage editor UI with Parameters, Macro, and Grid modes17 macro controls (8 Anatomy + 9 Character)2D Grid pad for real-time preset blendingConfigurable polyphony with 4-strategy voice stealingReal-time parameter morphing APISample-accurate note schedulingBuilt-in multiband compressionNatural release tail (breath exhale)Stereo Haas widening and equal-power panningDemo scene with working examplesFull source code includedAudio CapabilitiesPhysically-inspired vocal tract modelling (glottal source, body cavities, throat tube, 5 formant filters)Two oscillator models: LF glottal pulse and exponential decaySub-octave and detuned oscillator layersFM feedback for harmonic complexityBand-limited aspiration noise with formant trackingPre-tract and post-tract saturation with asymmetric biasSpectral tilt controlPitch sag, pitch AHD envelope, growl LFO, Perlin pitch wanderTremolo from gentle breathing to audio-rate epiglottal flutterFormant wander for evolving sustained vocalisationsJitter and shimmer for micro-instabilityMetallic inharmonic layer for alien/insectoid voicesRelease tail with tape-delay exhale reservoirCode IntegrationClean public API: `TriggerOneShot()`, `TriggerHeld()`, `TryTriggerNote()`, `ReleaseNote()`, `UpdateVoiceParameters()`, `SetVoiceNote()`Parameters struct for full state serialisation and interpolationVoice index and token system for tracking individual voicesDSP time scheduling for precise audio event timingIt wrote some of the documentation
