Personal and Professional are one-time lifetime licences
The Trial / Free edition allows large-folder indexing, analysis and review, with up to 100 lifetime unique verified Organized Library copy outputs. Personal and Professional are one-time lifetime licences, with future major upgrades such as 2.x and 3.x included under the current licensing policy. The current available application version remains 1.3.3.

Downloads and platform information:
https://musicnotehaven.ethercomm.eu/downloads/

Release notes:
https://musicnotehaven.ethercomm.eu/release-notes/

I'd welcome feedback from MIDI collectors, archivists and keyboard users about file identity, KAR/MIDI metadata and handling distinct arrangements of the same piece.

First off, I am incredibly grateful for the pioneers of MIDI and acoustic modeling. From the early wild west and east days of creating standards, to the incredible minds pushing things forward today, thank you. It’s this foundational work which provides the freedom and ease to tinker, hack, and rethink how to interface with sound design and musical inventions of all sorts…

Lately, I’ve been obsessed with how we, as humans, drive expression but also that which “expresses” upon us, a form of cohabitation. The brilliant EaganMatrix acoustic modeling engine and the way it handles organic, fluid responses is magical or perhaps elegantly mathematical. An equally brilliant mathematical generative sandbox is KARMA. Created entirely to express human intent, breaking free from the rigid midi landscape of yesteryear. Dedicated MPE controllers are also beautiful, the QuNexus was a great step in the journey and the current offering of fantastic controllers is growing, but I started wondering about all sorts of friction we create which could translate into musical expression.

How to expand on these foundational implementations as well as learn, by looking back to go forward?

How to achieve a similar deep, independent per-note expression which is musically aware of itself by using music theory as a variable in the model?

Eventually being able to introduce many new variables and modes of expression…

Currently, a functioning rough draft of something akin to a voice allocation scheme I am calling: HARMONIC_mpe. This might be the long lost keytar cousin of 6 string allocation. 🤪

I have been able to build a routing bridge that leverages KARMA v3.3 for the Motif XS with off the shelf MIDI 1.0 gear to unlock a “musically aware” style of multi-dimensional, acoustic-modeled expression. The goal isn’t to automate my music, but to create a setup where I’m generating the dynamics, using all forms of inputs to express multi-dimensional articulations on individual notes within a chord or phrase. And that chord or phrase also musically lives and breathes.

The Canvas: a multi-channel setup (Song mode) on the XS, and or a large mix of multi-timbral synths to create the voices specifically tuned for “organic, acoustic-style modeled” types of responses or easily re-geared towards a yet to be discovered soundscape.

The Brain: I am using a 6-module setup within the KARMA software. Instead of using it to generate traditional tracks, it’s acting as a “musically aware” voice allocation router.

The Splitter (Note Series Index Filters): play a chord, KARMA’s filters instantly recognize the chord structure and cleanly separates the notes —sending the indexed intervals to distinct modules. Such that the Root note goes to Module 1, the 3rd to Module 2, and so on. Each module gets its own dedicated MIDI channel. And by extension, when the sound module has multi-channel output, separate audio chains for post processing.

The Human Driver: By using KARMA’s DynMIDI matrix, I can mimic axis movements thru standard physical gestures—like moving individual faders, pedals or pads from MIDI 1.0 controllers—and map them directly to those specific modules. This extends to non human organic drivers of expression, using nearly any “midi controller” becomes possible. Weather, time, density, proximity, any sensor, any data which can be translated to midi CC.

Because the indexed notes are already separated by the filters and routed to distinct midi channels, “wiring” a traditional midi data stream thru DynMIDI cleanly reshapes and modulates the root of the chord. While another data stream breathes life into another interval of the chord. Each interval distinct but always perfectly or as imperfectly harmonious as expressed… much is yet to figure out

A Few Thoughts on Musical Intent & The Road Ahead

What excites me most about this setup, even as a legacy project, is how it unlocks a musical intent for both generative midi and acoustic modeling. In this sandbox, a chord is “understood” rather than simply transmitted. Because the routing relies on KARMA’s Index math rather than being strictly pitch-based, the musical structure itself becomes an active participant in how data flows. The system inherently knows the relationship of the notes within the harmony, changing how interactions are mapped.

It also challenged me to expand the horizon of what expressions might actually be possible. True musical articulation is more than just what a physical finger touches. We must respect the physical “cohabitation” we share with an instrument, but also remember that we already play countless instruments through an intermediary—sticks for a drum, bows for a violin, or complex mechanical hammers for a piano. And certainly math…

Should that intermediate layer not naturally extend to all possible interactions? After all, a physical guitar slips out of tune in humid or dry conditions. Saliva accumulates and alters the resonance of a brass horn. So many wild, chaotic environmental variables enter into how our ears ultimately perceive a sound—well beyond the neat, direct tactile intent of its original creator.

By decoupling the note-triggering surface from the modulation source, I am simply leaning into that acoustic chaos, introducing entirely new macro-gestural dimensions into physical modeling. Ultimately, I look at this setup as a transition.

As the saying goes, “a bridge gets more people to the promised land than a catapult.” For myself, heavily invested in the fun and exploration of both new and “legacy” gear, this offers a healthy, functioning pathway forward. It allows me to keep some beloved instruments relevant within a framework like MPE—which, if I am being completely honest, resides as one of truly beautifully formulated, brilliant hacks.

The fun truly starts when playing the mpe controller of your choice to modulate the over 400 parameters within each KARMA module routed back to an mpe enabled sound module!

Expressing a musical intent in real-time from each axis of the controller.

MIDI 2.0 is to me the new wild frontier, the promised land!!!

A work in progress…

PEACE
_
_
_

Yours hopefully

Chris.

MIDI Manager 1.3.0 is now publicly available for Windows and Linux.

Windows is available through Microsoft Store as package 1.3.0.0. Linux is available as Debian package 1.3.0-1, built and accepted on Ubuntu 26.04 LTS amd64.

Version 1.3.0 focuses on safer and more explainable library organization:

• Organization analysis uses filename clues together with embedded MIDI/KAR evidence and checks the offline music reference first.
• If there is no usable local result, the optional Online Artist & Song Discovery fallback sends only candidate artist/title text.
• Long-running analysis reports clearer phase, elapsed-time and heartbeat progress.
• Organized Library rename and Re-evaluate have stronger collision, identity and rollback protections.
• Manual Artist, Title and Filename corrections stay better aligned with the actual organized path, including case-only renames.
• Personal and Professional Version 1.x licences remain lifetime for the 1.x generation and remain usable offline after successful local verification.

Trial / Free can still index and analyse large source folders and can copy up to 100 lifetime unique verified files into Organized Library. Preview, Review and failed operations do not consume that limit.

Downloads:
https://musicnotehaven.ethercomm.eu/downloads/

Release notes:
https://musicnotehaven.ethercomm.eu/release-notes/

Manual:
https://musicnotehaven.ethercomm.eu/manual/

• sigboost Inc.: 00H 48H 00H
• Tsukuba Scientific Co., Ltd.: 00H 48H 03H
• Sonicware Co., Ltd.: 00H 48H 00H (Current typo)
• Poppy Seed Studio: 00H 48H 05H

I'd love to hear if anyone still finds this useful, it is browser-based so you don't need old drivers.

Thanks

--Andy

The current releases are:

Windows 1.2.1 via Microsoft Store
Linux 1.2.2 as the verified Debian package

Since the earlier versions, quite a lot has changed in the workflow.

The Source Index is now more durable, Technical MIDI Scan can resume and reuse current analysis where possible, and the Organizer can work incrementally with files that are already fully analyzed.

Recognition 2.0 has also become much more evidence-driven. File Detail, Human Review and Re-evaluate Library make it easier to inspect uncertain identifications rather than simply accepting filename-based guesses.

The Organizer workflow remains deliberately copy-first:

• Organization analysis
• Plan safe copies
• Review uncertain items
• Plan corrected copies
• Execute and verify
• Archive and cleanup as an explicit separate step

Library Health, MIDI Player, Playback Queue, Floppy Image Studio and IMG Viewer are all part of the current workflow as well.

The application has also been tested with libraries approaching 200,000 MIDI/KAR files, which has helped drive a lot of the recent work around incremental processing and responsiveness.

I made a new complete workflow video for 1.2 here:

https://youtu.be/5q9JugTzUL0

Main website:

https://musicnotehaven.ethercomm.eu/

Feedback from people working with larger MIDI collections, older keyboard media, GM/GS/XG files or mixed MIDI/KAR archives is still very welcome.