U-he Zebra 3

Zebra 3’s Generator Rack is a modular synthesis playground. Four main Oscillators morph between hand-drawn spline waveforms and can render as classic wavetables with up to 16-voice unison or as additive engines with up to 1024 sine partials. Each Oscillator has two serial FX slots offering 20 processes — from spectral filtering and phase distortion to animated spectral decay. Four FM Oscillators (comprising a modulator and carrier pair) chain freely — each accepts any rack signal as a third operator, so you can build complex FM algorithms without a fixed topology.

Noise generators, Exciters and Modal Resonators are designed to work together: drive a marimba or bell profile with a shaped noise burst or a patterned pluck excitation, then route the result through one of over 105 filter responses — including some of the finest analogue-modelled filters we have ever made — a Comb in reverb mode, or straight into an FM Oscillator as a modulator.

Dedicated Wavefolder, Ring Modulator and Distortion modules add harmonics anywhere in the signal path.

Vector mixing lets you crossfade or XY-morph between four sources, and almost every parameter across the rack — filter cutoff, fold depth, comb detune, resonator position — is a valid modulation target. The result is a single voice architecture that can go from pristine additive pads to gnarly physical-modelling textures without ever leaving the Grid.

Modulation Rack

Zebra 3 treats modulation as seriously as sound generation. Four ADSR Envelopes offer velocity sensitivity and a One Shot mode, with two Vary parameters adding options like delay, curvature (self-modulation), stretching or key scaling. Four LFOs cover sine, triangle, saw, square, sample-and-hold and smooth random — with tempo sync, phase offset and a delay ramp for gradual onset. Four MSEGs take things further: up to eight morphable curves along a shared timeline with freely positioned loop points, using the same spline editor as the oscillator. Attack, loop and release macros dynamically scale the speed and amplitude.

Beyond the Rack, on their own page, four dedicated Pitch modules give detailed control over tuning, glide and quantisation, independently managing up to four pitched parts within a single voice. And critically, many modulators double as triggers: an MSEG can drive an Exciter for patterned physical-modelling sequences, bringing rhythmic and melodic structure into the voice itself. In Zebra 3, pitch and gates live on equal footing with regular modulations.

Shaping and combining modulation sources is a particularly interesting area: Four Mappers let you reshape any modulation signal through a custom transfer curve — turning a linear ramp into a logarithmic sweep or a smooth fade into a stepped pattern — and trigger signals can transform a Mapper into a step sequencer. Four Mod Math modules combine, multiply and transform sources before they reach their target, so you can ring-modulate two LFOs against each other or gate randomness with an envelope — all without touching the signal chain.

Effects Rack

Zebra 3’s FX Grid receives audio from the four voice lanes via four global buses, Main plus three auxiliary sends. The send levels from the voice grid are modulatable per voice, so effects depth becomes part of the polyphonic expression even though the effects themselves are global.

The rack is built for impact. ModFX units cover chorus, flanging and phasing with their own internal LFOs, going well beyond typical stage counts — e.g. up to eight-voice chorus and 24-stage phasers with a choice of digital or analogue feel. Delays offer mono, stereo and ping-pong modes with tempo-synced note divisions, feedback filtering, modulation and a Diffusion (smearing) option, available in both analogue-modelled and digital flavours. Two Reverb models — Classic and Lush — offer everything from tight rooms to vast halls, while the TextureVerb effect scatters granular micro-fragments of audio for shimmering, abstract ambiences.

Compressors placed anywhere in the grid can glue a reverb tail, tame the dynamics after deep modulation effects, or produce sidechain-style pumping entirely within the patch. EQs, Filters, Distortion, Ring Modulator and Wavefolder modules are the same as their namesakes in the Generator Rack, so the tonal options you know from voice-level processing are also available for post-mix shaping. Two stereo mixers, a 4-in-1 mixer with vector and scan function, plus a complex 4-in-4 mixer are also available in the FX section.

Features

• Support for MPE, microtuning, MTS-ESP (realtime tuning)
• Adaptive interface: Modules only appear when added to the patch, organised via drag and drop
• Delays, multitap delay, reverbs, compressors etc. in global FX grid with per voice send amounts
• Available as CLAP, VST3, AU and AAX on macOS, Windows and Linux
• Over 1,200 factory presets
• Unified spline editor for oscillator waveforms and MSEG curves with geometric morphing
• Two oscillator engines: Wavetable (up to 16× unison) or additive (up to 1024 partials, can be inharmonic)
• 20 oscillator effects including spectral decay, sync, phase remapping and more
• Filters with 13 models, both classic and original (Ladder, Cascade, SVF etc) each offering up to 12 responses (e.g. LP, HP, BP with 6/12/18/24 dB/Oct)
• FM oscillators with 2 operators plus an audio input as a third operator
• Modal resonators and comb filters with detailed control over feedback and damping for physical modelling synthesis
• Noise and Exciter modules for realistic transient and physical modelling sources
• Vector & Scan mixers: Blend 4 sources with XY pad (loopable) or scan through them sequentially
• Pitch, gate and trigger signals as first-class modulation — modular-style routing freedom
• Mod Math modules for combining and transforming modulation signals
• Mappers that act as step sequencers or as freely drawable transfer curves for other modulators
• Ring modulators, distortions, wavefolders, equalizers (as global effects and per voice)

Requirements