01 / SEED
Melody
A short motif carries a rhythmic shape, interval pattern and contour — something a listener can recognize.
01 / ORIGIN
Gekibana: where the idea began
Gekibana explores algorithmic visual creativity. A flower has its own structure; several flowers can form a bouquet; specialized tools reshape the objects while retaining the relationships that make the composition recognizable.
Dynamic mandalas demonstrate how a complex visual world can arise from a remarkably small number of actions. In the workflow demonstrated by Gekibana, seven clicks can create and animate a mandala — not by manually drawing every part, but by invoking generative and transformational procedures.
The musical question follows naturally: what if motifs were built and transformed like visual objects — but along the dimension of time?
3Clicks on an empty canvas to establish the starting elements
1Click to produce the mandala
3Clicks to configure and launch special animations
The video above shows a Gekibana animation, not an automated music composition system.
Stem / support→Tonal center, pulse, supporting bass
Flower→Recognizable motif or theme
Bouquet→Ensemble of interacting voices
Waves / symmetries→Variation, inversion, rhythm and orchestration
02 / PRINCIPLE
Melody → Multiplicity → Harmony
Music of Forms begins not with a random stream of notes and not with the imitation of an existing musical style. It begins with an original melodic seed, which is developed through transformations while its identity and the direction of the composition remain intelligible.
02 / TRANSFORM
Multiplicity
New lines and versions emerge through inversion, transposition, shifts in duration, register, timbre and limited mutation.
03 / HOLD
Harmony
Constraints guide consonance, movement, phrasing and the larger journey from an opening to a resolution.
The rule: variation is welcome while at least one recognizable feature of the original motif survives. Novelty, repetition, pauses, cadence and controlled chaos become resources for composition.
03 / ENSEMBLE
The “little composers”
The next idea is not one universal algorithm that does everything. It is an ensemble of small, specialized creative procedures, each with a recognizable task and character. The concept describes seven proposed “masters” whose interaction could make the musical form richer and more controllable.
Punctuation
Pauses, endings, phrasing and a sense of breath.
Wave
Rise, fall and the flow of musical movement.
Spiral
Repeating a thought while changing its course.
Symmetry
Mirrors, balance and ordered relationships.
Golden Ratio
Distribution of accents and larger proportions.
Little Chaos
Surprise and deviation within boundaries.
Inverse Function
Reversal, reflection and musical dialogue.
One ensemble
Not seven independent pieces — several functions helping a single form evolve.
Motifs may changeVoices may enter or leaveForm must remain intelligible
The seven-master ensemble and evolutionary behavior are research concepts, not implemented features of the existing prototype.
04 / FIRST RESULTS
From a concept to working code
This is more than a manifesto, but it is not yet an autonomous composing system. The first experiments in August 2025 established a simple, reproducible workflow: motif → sections and constraints → programmatic construction → MIDI → listening and revision.
4
“Flowers”
Four short MIDI melodic motifs with bass support, produced with Python.
A–B–A′
Theme, change, return
A small three-part composition derived from the fourth motif, with a variation and a return.
JSON
Structure → MIDI
A parameterized description and code path for producing and revising a MIDI composition.
Current status. The core concept and an initial MIDI-generation prototype exist. The first musical results were treated by their authors as sketches, not finished compositions. The proposed ensemble of “little composers”, autonomous musical evolution and demonstrated artistic quality remain future work.
Open R&D · Creative systems
Can the rules of form become the tools of composition?
We are interested in developing this idea with partners in algorithmic creativity, composition tools, interactive art and generative systems — turning the structural concept into a fuller creative environment.
The ambition is not to copy music, but to build a new language for composing it.
From form in space to form in time.