To All Corporate Members, Software Architects, and Engineering Leads,

Following recent communication with the MIDI Association Technical Standards Board regarding idea contribution policies, I need to draw attention and opinions on two foundational digital signal processing (DSP) and psychoacoustic anomalies directly to individual member companies and developers in the public domain.

I hope companies with engineering access to the active MIDI 2.0 Working Groups can help implement these enhancements. This data addresses two structural bottlenecks that, (1.) currently degrade digital synthesis and, (2.) outlines the cross-disciplinary scientific proof requiring an expansion of the protocol's pitch architecture.

(1.) The Multi-Channel Sample Summing Power Error
The internal mixing buses and software voice accumulators of modern synthesizers and samplers default to legacy linear instantaneous amplitude summation (V_out = V1 + V2 + ... + Vn). Assuming a constant digital reference load, doubling the instantaneous voltage equivalent of phase-aligned harmonics scales the output power by an un-physical factor of 4 (P = V^2 / R), forcing an unnatural +6 dB surge across shared frequencies.

In physical acoustics, independent, non-coherent instruments add their powers linearly in open air (P_out = P1 + P2 + ... + Pn), yielding a power factor of 2 (+3 dB). Standard digital summing violates this conservation of energy law, forcing an aggressive, un-controlled "fortissimo" volume bloat across polyphonic passages. To achieve real-world physical compliance and restore natural dynamic balance, your mix accumulators should implement a Power-Conservative RMS Summing Mode at the core voice-summation gate:

V_mixed = SQRT( X1^2 + X2^2 + ... + Xn^2 )

Processing polyphonic channels through this root-sum-of-squares vector eliminates the un-physical power explosion, allowing stacked instruments to preserve their individual timbral integrity, spatial boundaries, and character naturally without requiring aggressive post-mix EQ carving or master gain damping.

(2.) The Biophysical and Empirical Mandate for Stretched Octave Architecture
While MIDI 2.0 provides an exceptional 32-bit data infrastructure to manipulate pitch via Per-Note Pitch Bend Change messages, implementation software remains tethered to a rigid, un-stretched 2:1 (1200-cent) octave envelope.

The restriction of the octave to a strict 2:1 ratio is an industrial artifact that ignores human neurophysiology. Seven decades of independent, peer-reviewed psychoacoustic and neurological literature consistently demonstrate that human peripheral hearing does not track a flat 2:1 integer grid:

* Melodic Octave Enlargement: Ward (1954) established that human listeners, when matching consecutive pure tones, systematically adjust the upper frequency (fH) higher than twice the lower fundamental (2fL), proving a cross-cultural biological preference for melodic "octave stretching."
* Constant Stimuli Verification: Cuddy, L.L. and Dobbins, P.A. (1988), in "Octave Discrimination: Temporal and Contextual Effects" (Canadian Acoustics, 16, 3-13), confirmed that the subjective musical octave is stretched approximately 20 cents sharper than the physical 1200-cent boundary.
* Mistuning Asymmetry: Demany et al. (1991, 2021) proved that human sensitivity to octave mistuning is highly asymmetric; the auditory system exhibits a significantly wider tolerance for positive (stretched) mistuning than negative (compressed) alterations.
* Brainstem Circuit Modeling: Alain de Cheveigné (2023), in "Why is the perceptual octave stretched?", isolates this asymmetry straight to the neural cancellation and coincidence-sensitive circuits of the auditory brainstem, proving the stretch is hardwired into human sensory biology.
* Orchestral Realization: Institutional research out of the University of Helsinki (Jaatinen, Pätynen, and Alho, 2019/2023) tracks this exact "octave enlargement phenomenon" across real orchestral instruments, showing that stretched tuning layouts optimize acoustic clarity and expressiveness.

My own published research builds upon this foundation, isolating the hard memory encoded within biological ROM of two distinct, unequal diatonic semitones: Wit's 1.055 resonance measurement (135/128, approx. 92.18 cents) and Burns’ 1.066 emission tracking cluster (16/15, approx. 111.73 cents). These coordinates sum to the 9/8 whole tone, expanding the scale architecture to a definitive biological stretched octave coordinate: (9/8)^6 = 1223.46 cents. Compounding these precise offset frequencies allows upper partials to cross at varying geometric angles—mimicking the natural phase distribution of a live concert hall and completely eliminating digital phase-locking.

3. Structural Proposal: Open-Ended Per-Note or Per-Octave Pitch Coefficients
To exploit the true capacity of MIDI 2.0, the implementation standards must look past the 2:1 container. I propose that Member Companies advocate for a standardized Per-Note Tuning Profile facility (or Per-Octave Profile, stretched octaves allowed) inside the MIDI-CI (Capability Inquiry) framework.

Rather than enforcing rigid 1200-cent templates, software instruments should expose an open-ended architecture where every individual note index, or a repeating 12-note octave cell, can be assigned its own custom pitch-stretch coefficient. This architecture would allow creators to program any desired macro-octave magnitude—whether aligning with the 1223.46-cent biological model or other historical/cultural tuning profiles—natively on a single shared channel. Exposing this open facility provides developers with an automated environment to realize authentic, context-dependent vocal and orchestral inflections that current samplers flatten.

I have processed open-source MIDI data grids through a specialized Tuning Parser Application running this stretched configuration across reference-grade sound libraries. Complete 16-file audio verification matrices isolating these acoustic behaviors across selected musical examples—including Mussorgsky's "Gnomus," Wagner's "Flying Dutchman" Overture, and Tchaikovsky's Piano and Violin Concertos, among others—are hosted on my open-access repository for public evaluation.

Precise microtonal tuning on my provided demonstrations is made fully operational within the current legacy MIDI environment by programmatically re-assigning an independent, dedicated MIDI channel for every single note. A maximum polyphony of 60 simultaneous notes is achieved across 4 concurrent MIDI software instances utilizing CoolSoft VirtualMIDISynth. Twelve alternating semitones of said sizes (16/15 and 135/128) arrive at the definitive 1223.46-cent octave envelope, resulting in emotionally rich musical renderings. Detailed mathematical analysis explains how this structural arrangement natively initiates the sought-after vertical "ringing" effect found in elite a cappella vocal quartets.

The complete biophysical equations, neuro-mechanistic deductions, and data sets backing this framework are available via my open-access monograph:
Liu, C.-F. (2026) The Resonance of Emotion: A Mechanistic Theory of Musical Intonation and Affect. Open Journal of Acoustics, 14(2), 21-81.
Direct Publisher Portal:
removed link
Audio Demos open site:
removed link