This issue: 11 new finds across 8 devices in the dspm ecosystem, including 2 new connection ideas.
New connections
Connection: PiPitch (dspm-box) -> Octatrack MIDI track via H12MIDI Pro
Pitch-track a live audio source on the dspm-box's Pi 5 into MIDI notes with PiPitch, route them through the CME H12MIDI Pro to an Octatrack MIDI track — turns any acoustic input into a live-playable trigger for the OT's sample machines.
Kastle 2 FX Wizard as a MIDI-clocked generative voice through the CME host
Now that the Kastle 2 FX Wizard takes USB MIDI, route it through a CME USB-MIDI host to clock its CV/gate pattern generator off the rest of the rig instead of running it free-running.
norns
norns 3.0.0 merges matron and crone into shared memory
norns 3.0.0 unifies matron and crone into one process, replacing their internal OSC IPC with shared memory — the biggest architecture change since 2.0.
The change is to matron↔crone internal IPC only — it does not touch norns' external OSC listener on :10111, which is what the Node bridge and phones actually talk to for dspm_archive.lua param control. So the bridge's OSC path should be unaffected. But "should be" is doing work in that sentence for a from-scratch rewrite of the process model: script-side osc.event handling, param dispatch timing, and anything touching crone/softcut directly are exactly the surfaces that regress first after a merge like this. Don't blind-upgrade the physical shield or the norns-docker-debian twin to 3.x without testing OSC param control and softcut behavior on one side before mirroring the fix to the other — same discipline as evidence-first-debugging.
Source: monome/norns GitHub releases
Bastl
Bastl Kastle 2 FX Wizard gets USB MIDI in/out via firmware update
The patchable Kastle 2 FX Wizard multi-effects box joins Wave Bard in getting free USB MIDI I/O, on top of its existing tempo-synced CV/gate pattern generator.
Another Kastle 2 unit that can now sit on the same USB MIDI bus as the rest of the rig instead of needing its trigger input hand-patched from a separate clock source. Since the FX Wizard is an effects processor rather than a sound source, the useful move is feeding it MIDI clock from the OT or bridge so its tempo-synced effects (delay, etc.) stay locked to the set without a patch cable — freeing that trigger input for a CV modulation source instead. Combined with the Wave Bard's MIDI note input, a single USB-MIDI host port can now drive both Kastle 2 units together as one MIDI multi-timbral pair.
Source: gearnews.com
Kastle 2 FX Wizard gets USB MIDI in/out via firmware
Bastl's patchable multi-effects Kastle now takes and sends USB MIDI, joining its built-in CV/gate pattern generator and analog sync in/out.
The pattern generator's gate/CV output can now be MIDI-clocked or driven by incoming MIDI notes/CCs instead of only its internal LFO/envelope source, which makes it usable as a small generative MIDI/CV modulation source inside the rig — e.g. clocked from norns or the bridge over USB MIDI, patched to modulate its own crusher/slicer depth for an unpredictable, semi-autonomous effects layer on whatever's running through it. Pairs naturally with the existing Bastl CV connection idea (2026-06-21-connection-bastl-cv-orac-organelle-modulation) — now that path can originate from MIDI instead of only CV.
Source: gearnews.com
Nord Micro Modular
Nord Modular Editor adds arpeggiator, randomizers, morph groups (July 2026 update)
The free browser-based Nord Modular G2 editor gains module/LFO/sequencer randomizers, an arpeggiator, morph groups, and dark mode — a substantial update to the tool already on the roster.
Randomizers + morph groups turn the browser editor into a real generative patch-design tool rather than just a visual editor — you can now sketch a chaotic patch, randomize it toward something usable, and export it to load onto the Micro Modular over sysex, all without touching the hardware's tiny screen. The new MIDI controller mapping is the more directly useful bit: it means patch parameters (including any vocoder/filterbank chain used as a processing stage) can be tied to incoming MIDI CC, so the bridge's MIDI output could drive Nord patch parameters live during a set, the same pattern already used for the Rytm and OT.
Source: matrixsynth.com
Pure Data
Pd vanilla 0.56 ships native OSC objects, retiring the mrpeach dependency
Pure Data vanilla now ships oscparse/oscformat + netreceive/netsend natively — no external library needed for basic OSC in or out.
Both norns and the Organelle are Pd-based under the hood, and any bridge-side or patch-side Pd work (talking OSC to either, or prototyping an OSC patch before porting it to Lua) no longer needs an external dependency — one less thing to install/version-match when moving a patch to plugdata (2026-06-28-plugdata-v0-9-3-pure-data-plugin-on-arm64-and-windows-asio) or a different machine mid-tour. Lowers the floor for a quick Pd-side OSC test patch to sanity-check what the Node bridge is actually sending on :10111 before touching the norns script.
Source: pd.iem.sh / puredata.info
MIDI router
CME H12MIDI Pro brings 6-port USB MIDI hosting with router/filter/mapper
CME's flagship USB MIDI host adds 6 host ports and 6 TRS I/O with software-configurable Type A/B, merge/split routing, filtering, and remapping.
This is the class of box the roster already earmarks for the planned CME USB-MIDI host in the rig. 6 host ports covers Octatrack + Analog Rytm + norns (via a class-compliant USB-MIDI adapter) + Kastle 2 (now USB MIDI, see 2026-08-05-kastle-2-fx-wizard-gets-usb-midi-in-out-via-firmware) simultaneously, with the client port left free to also hand a merged MIDI stream to the laptop running the Node bridge — so the bridge could inject phone-triggered MIDI CCs into the same merged stream the hardware sees, rather than needing its own USB-MIDI interface per device.
Source: cme-pro.com
CME H12MIDI Pro — next-gen USB host MIDI hub, 6 host ports, 480 channels
CME's successor USB-MIDI host: 6x USB-C host ports (16-in/16-out expandable), 6x TRS MIDI I/O, 480 channels, near-zero latency, pocket-sized — the leading candidate for the planned CME USB-MIDI host box.
This is the more capable option for the CME USB-MIDI host box already on the roadmap (see gear-integration) — 6 USB-C hosts plus 6 TRS I/O beats the H4MIDI WC's single expandable USB-A host, meaning norns, the babayaga bridge, the Octatrack, and the Analog Rytm could each get a dedicated host port instead of sharing a hub. That removes a failure point at gigs (USB hub power/enumeration issues) and gives every device its own routing rule in HxMIDI Tools. TRS-A/B switchability also covers both Elektron's and Bastl's MIDI cable conventions without adapters. Worth comparing directly against the H4MIDI WC (see cme-h4midi-wc-8-port-usb-host-midi-router-and-remapper) before buying either.
Source: cme-pro.com
dspm box
PiPitch — low-latency guitar/audio-to-MIDI LV2 plugin for Zynthian on Pi 5
Open-source ONNX/RTNeural pitch-to-MIDI LV2 plugin, now autoroutable in Zynthian chains — turns any audio input into a MIDI note source on the same Pi 5 stack dspm-box is built on.
This turns any audio input on the dspm-box's Pi 5 stack into a MIDI note source — feed it room mic, a guitar DI, or even the norns' own audio out, and get MIDI notes back that can drive the Analog Rytm, the midi-control-surface PWA/WS surfaces, or a synth chain inside Zynthian itself. For dspm-box specifically (which already plans a Zynthian synth mode alongside the TX-6-style mixer mode), PiPitch is a way to let acoustic or performer input control the electronic side without a dedicated pitch-to-MIDI pedal. It's also a candidate audio→MIDI front-end for the barcode looper: pitch-track a live take and use the resulting MIDI to trigger parameter changes via the bridge.
Source: discourse.zynthian.org
Zynthian on Pi 5 is fast but still pre-alpha
Zynthian loads snapshots in half the time on a Pi 5, but the Bookworm build is pre-alpha — audio driver overlays, DAC routing, and WebConfig are all still in flux.
This is direct research for the invented dspm-box build (Pi 5 + RaspiAudio 8x8 HAT + knob panel, boots into TX-6-style mixer mode or Zynthian synth mode — see 2026-06-27-dspm-box-multimode-mixer-synth-invention). It confirms the Pi 5 has plenty of headroom for the synth-mode side, but flags the real blocker: Zynthian's own Pi 5 audio driver support isn't settled yet, and the RaspiAudio 8xIN+8xOUT HAT's device tree overlay would need to coexist with whatever Zynthian expects — exactly the kind of card-handover conflict the roster's own dspm-box search angle ("systemd Conflicts audio card handover") anticipates. Don't build the synth-mode half on the assumption Zynthian-on-Pi5 is turnkey yet.
Source: discourse.zynthian.org
Auto-compiled by the dspm gear feed — an offline research archive of the gear behind the dspm performance ecosystem.
