Five releases worth paying attention to this week. Cherry Audio has revived one of the most obscure polysynths in history. Baby Audio's HyperWarp takes granular processing into graphical sequencing territory. Tech Audio ships a sampler that generates MIDI by simulating collisions in a 2D physics world. Liljeholmen DSP adds a full sequenced drum machine to their drum synth line. Floodo's Audio Labs brings Volterra Series saturation at an introductory price for Windows producers.
Cherry Audio Pentaphonic
$59, VST3 / AU / AAX, Mac and Windows
The original Gleeman Pentaphonic was designed in 1981 by brothers Bob and Al Gleeman in a Silicon Valley garage. Fewer than 70 units were ever produced, making it one of the rarest polysynths in existence. Cherry Audio's Pentaphonic is a recreation of that instrument.
The hardware was a five-voice polyphonic synthesizer using three digitally controlled oscillators per voice, each with eight waveforms, running through Curtis CEM 3320 four-pole filters. The Gleemans later offered a version in a clear acrylic housing with the circuitry lit against green PCBs visible through plexiglass panels, which became the instrument's visual signature. Production wound down as the market shifted toward digital synthesis in the mid-1980s.
Cherry Audio's plugin scales the voice count to 16, adds a second LFO, extends the filter with a 12 dB two-pole mode, and adds velocity and aftertouch support alongside the original's polyphonic sequencer. The architecture stays close to the source material. All eight original waveforms are present. A 22-effect processing chain is included. Seek out recordings of the original Gleeman before buying. The character is distinct in a way that does not translate well to description.
Baby Audio HyperWarp
$49 intro (regular $99, through November 30), VST3 / AU / AAX, Mac and Windows
HyperWarp is a granular multi-effect built around a graphical time sequencer. Incoming audio is broken into short grains and reassembled according to a user-drawn automation graph, producing stutters, speed changes, reverses, and pitch-shifted textures. Seven processing profiles shape the character of the granular engine. A random patch generator is included.
Baby Audio's approach frames granular processing as a performance and arrangement tool rather than a deep editor. Parameters respond smoothly to automation, preset browsing is fast, and the CPU overhead is low enough to run on multiple tracks. This sits in the same territory as their Crystalline and Spaced Out plugins: simpler than the technology could be, deliberately.
At $49 through November 30, this is priced below where granular effects from most established developers sit. A free trial is available.
Liljeholmen DSP Trumsynt-KIT
$49 (includes Trumsynt-KIT and Trumsynt-16), VST3 / AU, Mac 12+ (Apple Silicon and Intel) and Windows 11
Liljeholmen DSP is a small Swedish developer. Trumsynt-KIT is their new drum machine: eight independent synthesizer voices, each with a 16-step sequencer, per-step parameter locks, and separate audio outputs. The factory library ships with 253 voice presets and 12 kits.
Each voice is a complete synthesizer. The engine combines two oscillators with FM and hard sync, a noise layer, a click layer for transient shaping, and a self-oscillating four-pole ladder filter. Pre- and post-drive are available per voice, and the stereo mix channel has glue compression and drive. Eight 16-step patterns per voice with pattern chaining. The architecture is closer to a modular drum synth than a sample player, and kicks, snares, hats, bass, and pitched percussion all use the same voice type.
Trumsynt-16, the single-voice version with its own modulation sequencer, is included in the same $49 license. A 14-day trial is available on their site.
Tech Audio Shrapnel
$49, VST3 / AU / AAX / CLAP + Standalone, Mac and Windows
Shrapnel replaces the sequencer with a physics sandbox. The interface is a 2D canvas where you place obstacles and configure a particle spawner. Particles fire into the scene, collide with obstacles, and each collision triggers a sound. Hit speed, particle size, color, and collision position shape the output. You do not program notes. You design a physical environment and let it generate events.
Each obstacle supports up to eight sound layers with four output modes: sampler (sample groups with random selection and pitch/volume variation), granular (collisions trigger grains), MIDI (sends note data to drive other instruments), or audio input (captures live audio and replays it on collision). The physics simulation handles timing and variation, producing patterns that are non-repeating in a way step sequencers cannot match.
For sound design, ambient texture work, or anything requiring non-deterministic rhythmic material, Shrapnel opens up approaches that are difficult to replicate by other means. A free demo is available.
Floodo's Audio Labs Cascade
$39 intro (regular $129), VST3, Windows 10/11 only (macOS in signing pipeline)
Cascade is a dual-stage saturation plugin built on Volterra Series modeling. Where standard harmonic distortion describes how a system generates individual harmonics, Volterra modeling captures how those harmonics interact across successive stages: each stage alters the input in a way that affects how the next stage responds. Floodo's Audio Labs uses this to model ten saturation personalities, from British consoles and classic compressors to vintage tubes and tape, as cascading stages where stage order changes the result.
The practical effect is saturation behavior that responds differently to transients than to sustained material, and where running two stages produces more complex harmonic content than either stage alone. This is Windows-only at launch, with macOS in the code-signing pipeline. At $39 intro, it is worth trying if you are a Windows producer and want to hear what Volterra-based modeling sounds like in a saturation context. A fully functional trial is available.