Nultra Multiband Ecosystem

Nultra Multiband Ecosystem is a stereo audio processor built around the Nultra Operator, a user-created identity-gated state operator:

S_next = S + A(t) * (f(S) - S)

In this repo:

• S is the current band-limited audio state
• f(S) is a transformed version of that state
• A(t) is a shaped, thresholded aperture that decides how much transformation is allowed

The important idea is the true null plateau. When A(t) = 0, the signal is preserved exactly while time still advances. That makes this feel different from ordinary tremolo, choppers, or always-on filter sweeps.

This repository contains the Canonical Version 1.0 Baseline of the processor and the interactive menu runner used to audition it on real tracks.

Why It Sounds Different

Most modulation effects constantly touch the signal. Nultra does not.

The aperture is built as:

A_base(t)   = base oscillator or chaos trace
A_gamma(t)  = A_base(t)^gamma
A(t)        = 0,                  if A_gamma(t) < eta
              A_gamma(t),         otherwise

That hard threshold creates spans where the audio is left completely alone, followed by spans where the signal is pulled toward a transformed state.

In the multiband ecosystem build:

• the low band stays structurally stable
• the mid band carries the underwater motion and residue memory
• the high band carries the metallic diffusion cloud
• the low band can modulate the high band threshold, so the spectrum behaves like an interdependent system instead of three silos

Canonical Version 1.0

The frozen baseline is:

• src/nultra_multiband_ecosystem/processor.py
• src/nultra_multiband_ecosystem/menu.py

Locked architectural interlocks:

• 3-band split at 180.0 Hz and 2400.0 Hz
• numba acceleration on the hot loops
• high-band safety interlocks:
  • slew limit 0.0040
  • zero-cross search 96 samples
  • edge ramp 36 samples
• explicit stereo offsets preserved:
  • mid pi/2
  • high 2pi/3

Canonical presets:

sweet-spot

high_parasitic_eta_alpha = 0.14
high_gamma               = 7.3
high_eta                 = 0.31
mid_eta                  = 0.29
mid_feedback_beta        = 0.28
mid_feedback_delay       = 1024
high_diffusion_gain      = 0.92

Perceptual target: subtle, fluid, organismic textural breathing

stronger

high_parasitic_eta_alpha = 0.10
high_gamma               = 6.9
high_eta                 = 0.28
mid_eta                  = 0.27
mid_feedback_beta        = 0.34
mid_feedback_delay       = 1408
high_diffusion_gain      = 0.98

Perceptual target: intense matrix cross-modulation, wider metallic phase clouds, longer ghost trails

Future experiments should be forked from this baseline instead of rewriting it in place.

Installation

python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt

Or install as a package:

pip install .

Quick Start

Run the processor directly:

python -m nultra_multiband_ecosystem --mode file --input path\to\song.wav

Use the interactive menu runner:

nultra-menu

The menu runner:

• scans a root directory recursively for .wav files
• lets you choose sweet-spot, wetter, stronger, or custom
• prompts for key settings in custom mode
• renders a timestamped output file

By default it scans:

E:\4trax

But you can override that:

nultra-menu --root D:\Audio

Project Layout

src/nultra_multiband_ecosystem/
  __init__.py
  __main__.py
  processor.py
  menu.py

docs/
  nultra_operator.md

tests/
  test_smoke.py

Notes

• No source music files are committed here.
• Rendered .wav files are intentionally ignored.
• The Nultra Operator described here is a project-specific mathematical and DSP concept, not a standard audio term.

License

MIT. See LICENSE.