KAGARIBI : Synthesizer

KAGARIBI is a free synthesizer for iPhone and iPad that can explain the sound you are playing. Most synths stop at making a good noise. Open Sound Anatomy while a preset is sounding and KAGARIBI writes out, in plain sentences and with the real values, what makes it work: "Unison ×3 — detune 12 cents, spread 25%. Stacking slightly detuned copies creates thickness and stereo width." "Filter env +30% — the attack opens then closes; decay 220ms sets the contour." You read the sound instead of guessing from a waveform. The rest of the instrument is built around what actually gets in the way on iOS. TUNING, INSIDE THE INSTRUMENT Sixteen tunings are built in: just intonation, Pythagorean, quarter-comma meantone, Werckmeister III, Kirnberger III, the Japanese pentatonics (miyako-bushi, ritsu, min'yō, ryūkyū, in-sen), 19-, 24- and 31-tone equal, the non-octave Bohlen–Pierce, and a harmonic-series scale. Scala .scl and .kbm files import from Files. The tuning is stored inside the preset, so it opens at the same pitch on another device or inside a DAW — no separate tuning plug-in in front of the synth. A chart shows how far each note sits from 12-tone equal temperament, and glides, bends and detune stay smooth in every tuning. ASSIGN A CONTROLLER ONCE MIDI assignments are stored on the device and survive preset changes and app restarts, so you never learn the same knob twice. After a preset change nothing moves until the physical control catches up with the value, so the sound never jumps mid-performance. Current values are sent back on the assigned CCs, so LED rings and motorised faders follow. Relative encoders, ranges and inversion are supported, and you can assign preset next/previous and panic as well as any of 96 parameters. YOUR SOUNDS STAY YOURS The app and its AUv3 plug-in share one library, so a sound made in the app appears in the preset list of every host, and the other way round. Writes are atomic, deleting moves a sound to a trash you can undo, and a file that cannot be read is set aside rather than discarded. BUILT NOT TO DROP OUT KAGARIBI shows how many more voices are safe rather than a CPU percentage. As the load approaches the audio deadline the voice cap drops automatically and recovers when there is room again, so playing continues instead of stuttering. Live Safe and Studio modes trade polyphony against oversampling, and Offline HQ rendering writes a 48kHz / 24-bit WAV at maximum quality without interrupting playback. MPE AND MIDI 2.0 Per-note pitch, pressure and timbre are handled throughout the engine, including MIDI 2.0 (UMP) per-note pitch bend. An ordinary keyboard and an MPE controller travel the same path, with no zone setup to configure. THE ENGINE Two oscillators with five analogue waveforms and six band-limited wavetables of four morphing frames each, plus phase bend, wave folding, 1–7 voice unison, FM, ring modulation, hard sync, a sub oscillator and noise. Six filters including a non-linear ladder. Three envelopes, two tempo-syncable LFOs, an eight-slot modulation matrix and four macros. Arpeggiator. Effects chain of drive, tone, phaser, chorus, tempo-synced ping-pong delay, reverb, compressor and limiter. Up to 32 voices. Multi-touch keyboard with velocity by strike position and glissando, USB and Bluetooth MIDI, and an AUv3 instrument with 96 automatable parameters and full state saving for Logic Pro, GarageBand, AUM and other hosts. FREE, WITH NOTHING ATTACHED KAGARIBI is free and complete. No in-app purchases, no advertising, no account, no analytics, no tracking. It does not use an internet connection to run, and your sounds and settings never leave the device. Thirty factory sounds are included across bass, lead, pad, pluck, keys, arp and texture.