GENRE_SPAN_ANCIENT_TO_TECHNO.md

music/GENRE_SPAN_ANCIENT_TO_TECHNO.md

THE GENRE SPAN — ANCIENT THROUGH MYTHOLOGICAL AND MODERN TO TECHNO

CANON SUBORDINATION — this document is PROPOSAL-TIER: it serves canon and never outranks it.

Canon served: the composition floor as re-graded by Josh on 2026-08-07 night —

docs/spine/DECISIONS_PENDING_JOSH.md:5600 (ROUND 2 GRADED), specifically the clause

*"Remember all of the culture and regions we are gonna be working with from ancient history,

mythological, and modern day as well up to even technological and techno music."*

Derived from: registries/T0_Chapter_Index [ACTIVE v1.2]/ (the era and region_primary

columns, 79 rows) · registries/T0_Theme_Registry [DRAFT v0.1]/ (38 rows, the

instrumentation_substrate / head_cell_spec / allow_register_list cells) ·

docs/spine/CH_*.md # music_mood lines · the CVD · docs/DOC_MAP.md § 0 for authority order.

READ THAT CANON FIRST. **If this document disagrees with canon, CANON WINS and this document

is the defect.** Nothing here is applied until it is ratified.

Tier: RESEARCH SYNTHESIS, proposal-tier. This document contains no canon and rules on nothing.

It is the seventh lane under docs/proposals/music/MUSIC_CRAFT_DOCTORATE.md and it does not

restate the six that exist. Where a claim is craft consensus rather than evidence it says so;

where this document PROPOSES rather than reports, the proposal is marked.

0. Derivation — what canon was read, and what canon had already decided

The span is not a guess: it is readable off the era column of T0_Chapter_Index, which covers all

79 nodes. Read at HEAD, the arc runs from Ch 10's Olduvai (era cell: Paleolithic, ~1.8 million

years) through Ch 31's Altamira, the composite antiquities of Ch 14-28, medieval and early-modern

Europe and the Atlantic sailing era at Ch 24-37, a hard hinge into PRESENT DAY at Ch 39, twenty

modern nodes, four atemporal dimensional nodes at Ch 66-69, five mythological otherworlds at

Ch 70-74, and the fairy realm at Ch 75-77. That is not a stylistic preference; it is the structure

of the game.

Four things canon had already decided, which this document derives from rather than proposes:

• THE ELECTRONIC REGISTER IS ALREADY ASSIGNED, AND IT IS NOT ASSIGNED TO THE REALMS.

T0_Theme_Registry's VIMANA_ACTIVATION row carries instrumentation_substrate =

"orchestral with electronic-acoustic hybrid for special-asset cues plus ceremonial choral overlay;

Vimana propulsion SFX layer per §10", and its allow_register_list repeats the hybrid clause. The

wonder-and-vril identity is the hybrid one. Its motif_class is harmonic, not tune, so it

takes the progression-recognition test and never the hummability checklist.

• THE PURE-ELECTRONIC CUES ARE NAMED IN THE SPINE, IN MODERN NODES.

docs/spine/CH_44.md:220 asks for "a spare digital ambient for the metaverse" beside "a modern

financial-thriller tension for the heist" and "a 1910 period register for Jekyll Island" — three

idioms in one node. docs/spine/CH_46.md:207 asks for "a taut surveillance-thriller tension for

the facility". docs/spine/CH_39.md:279 opens the modern arc on "the megacity pulse".

docs/spine/CH_13.md:246 — inside the Josh-gate slice — already reads "mood

reverent-to-industrial".

• THE SEAM PROBLEM IS ALREADY CANON, UNDER A DIFFERENT NAME. JOURNEY_WORLD's

transformation_plan reads: "the defining surface is PER-CULTURE REGISTER MIGRATION -- the same

tune re-scored into an entirely different civilisational voice and still recognisably the same

tune (the Nerevar Rising to Dragonborn precedent)", and its head_cell_spec carries the reason

for its shape: "4-note head, triad-outlining so it survives re-scoring into ensembles with no

shared tuning system". Section 5 extends that constraint to the electronic half and finds it

holds, with one clause to add.

• THE REALMS ARE NOT A GENERIC ETHEREAL WASH. FAM_THE_OTHERWORLD's allow_register_list reads

"each realm's own invented register, per the realm design directive -- these are REAL otherworlds

and take grandest-display treatment, not a generic ethereal wash", and its deny list forbids

building a realm register by exoticising a living human tradition. Section 4 is written against

that clause.

One measured fact about our own comparator corpus, because the competitive read rests on it.

build/audio/exemplars/CORPUS.json carries 182 rows. Counted at HEAD, NINETEEN are electronic-led

rather than orchestral- or band-led — Deus Ex (EX_124), both Doom rows (EX_046, EX_047), Ultrakill

(EX_055), Jet Set Radio (EX_136), Streets of Rage 2 (EX_145), Ridge Racer Type 4 (EX_104), both

Silent Hill rows (EX_076, EX_122), Max Payne (EX_010), Perfect Dark (EX_175), Transistor (EX_116),

both Mass Effect rows (EX_081, EX_091), both Minecraft rows (EX_017, EX_018), Chemical Plant Zone

(EX_128), Escape from the City (EX_151) and F-Zero X (EX_061). That is 10.4% of the roster; of the

THIRTY rows whose audio is in hand (18 ACQUIRED plus 12 CARDED), SEVEN are electronic-led — 23%. So

the idiom sits at roughly a quarter of the measurable pool of Josh's own favourites, and not one

formula card drives an electronic grammar, because every melodic field on the card

(the_hook.pitch_contour, motif_economy.interval_3gram_repetition) is pitch-based and a filtered

bassline is not primarily a pitch event. Named again as measurement debt in section 6.

1. The idiom map for a 79-node game

1.1 How the map was derived, and what kind of claim it is

The map below is a PROPOSAL. Its INPUT is canon — the era cell of each T0_Chapter_Index row plus

that node's # music_mood line in the spine — and the mapping from an era to an idiom is this

document's recommendation, not a ruling. Where a spine line already names a register, the map quotes

it rather than choosing for it. Where it does not, the map proposes and says so.

The organising principle is that an idiom is chosen for what it DOES dramatically, never for what

period it decorates. Three of the ten below are period-signalling idioms; the other seven are

function-signalling and can appear in any era.

1.2 The map

1.3 The three claims inside the map that are load-bearing

• THE ERA DOES NOT CHOOSE THE IDIOM; THE FUNCTION DOES, AND THE ERA CONSTRAINS THE PALETTE. Ch 44

is one modern node and its spine line asks for three idioms — thriller, digital ambient, and a

1910 period register. Ch 30 is one medieval node and asks for bardic harp, drone, and plainsong,

which are three distinct historical practices. An idiom-per-era map would have produced one cue

per node and would have contradicted the canon it claims to serve.

• MODERNISM IS A HORROR AND BOSS VOCABULARY, NOT A PERIOD. Penderecki's Threnody used tone clusters,

quarter-tone bands, bow-tap and behind-the-bridge playing, and timings notated in seconds rather

than bars; film took the vocabulary wholesale, with Friedkin's The Exorcist using Polymorphia

directly. Game scoring inherited it — Jason Graves' Dead Space and Garry Schyman's BioShock are

described in the interactive-music literature as fundamentally aleatoric, with adaptive systems

triggering short segments of atonal chaos on game events. This is why the vocabulary is licensed

in Ch 2 (a Pleistocene node) and Ch 70 (a mythological one) with no anachronism claim: nobody

hears a cluster and thinks 1960.

• THE CYCLIC-TRADITIONAL ROW IS OWNED BY ANOTHER LANE AND IS NOT RE-DERIVED HERE.

craft_research/ORGANOLOGY_SLICE_CULTURES.md is authoritative for Ch 2-13 and

craft_research/REPETITION_AND_VARIATION_FORM.md § 5 for cyclic traditions as engineering. This

document's only addition to that row is made in section 3.3.

2. What is actually entering game scoring now

This is the competitive read behind Josh's earlier "set the next bar" floor: a report on the last

three years of released work, with titles and composers named so a later lane can check it. Where a

trend claim is an inference across several titles rather than a stated fact, it says so.

2.1 The turn toward chamber and solo writing

The prestige direction has moved away from the full-orchestra-always default toward intimate chamber

and solo writing — string quartet, solo guitar, solo voice, small ensembles with long rests between

entries. Coverage of the 2024 and 2025 release years repeatedly describes minimalist chamber

writing, slow-burning atmospheric drones with reserved individual interventions, and cellos moving

in and out of the soundscape as the characteristic sound (VGC's composer-chosen best-of features for

both years; a closer listen's Press A surveys). The precedent under it is older: Santaolalla's The

Last of Us (EX_005 in our corpus) is a solo-instrument score for a AAA flagship. What it means for

us: our principle 2 sets 24-to-40 parts for full cues and 3-to-8 for the sparse nature class, and

the chamber turn says the SPARSE class is where the prestige is — our 3-to-8 band is a first-class

deliverable rather than a reduced one.

2.2 The vocal-led score

The strongest single data point in the current window is Clair Obscur: Expedition 33 (2025), music

by Lorien Testard with vocalist Alice Duport-Percier elevated to co-composer, roughly thirty other

musicians and a nine-person choir, 154 tracks at base-game release, spanning classical opera to

heavy-metal rock and making substantial use of leitmotifs. It took Best Score and Music at The Game

Awards 2025 and the World Soundtrack Award for Game Music. The pattern — the voice carrying the

identity rather than decorating it — is the clearest "next bar" signal in the set.

Ours already has a row for this and it is not a tune. VOICE_COLOUR_SIGNATURE is

motif_class = timbre, carrier_voice = "wordless choir", judged on the atmosphere bar and

explicitly never on the hummability checklist. The competitive read says that row is

under-prioritised, not over-specified — and the doctorate's honest-limits section already names a

CC0 voice set as the highest-leverage acquisition the program lacks. This is the second independent

argument for the same purchase.

2.3 Folk revival and historically informed performance

Two titles define the poles. Pentiment (Obsidian, 2022) commissioned the early-music ensemble

Alkemie, who curated, composed and arranged nearly all of the music over three years on shawms,

hurdy-gurdies and other period instruments, with the game director stating the music is either

strictly historical or historically inspired, and with Hildegard, Isaac, Josquin and Machaut

suffusing the score. Kingdom Come: Deliverance and its 2025 sequel (Jan Valta with Adam Sporka) took

the opposite and equally deliberate route: a film-like orchestral score with medieval instruments

and choir inside it, on the stated reasoning that a 15th-century story is being told to a

21st-century player. Valta's useful observation for us is that the period's music was dominated by

the spoken word with instruments in a background role, and that its metre ebbed and flowed far more

fluidly than modern ears expect.

The finding: THE FORK IS EXPLICIT AND BOTH ARMS ARE RESPECTED. There is no consensus that

authenticity wins; there is a consensus that the choice must be MADE and declared. Our canon already

made it, one node at a time, in the phrase that recurs across the spine's music_mood lines — "at

reference register", "referenced-not-reproduced". That is the Valta arm with a care clause attached,

and it is the right arm for a game that visits seventy-nine places rather than living in one.

2.4 Drone and dark ambient

Akira Yamaoka's Silent Hill work is the canonical case and is in our corpus twice (EX_122, EX_076):

dark ambient synthesiser soundscapes, electric guitar, industrial percussion, with noise and radio

static functioning as threads in the gameplay and narrative rather than as decoration — Yamaoka

worked as sole composer AND sole sound designer, which is why the boundary between score and SFX is

not visible in that game. Jason Graves' Dead Space is the orchestral wing of the same idea.

The finding for us: THE SCORE/SFX BOUNDARY IS A PIPELINE ARTEFACT, NOT A MUSICAL FACT. Our

music_mood lines and T0_SFX_Registry are separate surfaces, and several of our nodes ask for a

register that only exists across the join — Ch 68's "vast, patient, pre-verbal register (the sea

heard, not scored as a foe)" (CH_68.md:194) is a sound-design cue written in a music field. Recorded

as a finding for canon rather than applied.

2.5 The metal and orchestral hybrid

Mick Gordon's Doom (2016) and Doom Eternal (2020) are the reference, and both are in the corpus

(EX_046 ACQUIRED, EX_047). The method as Gordon has described it: throw away the previous project's

approach and relearn from the beginning for each score; design synth pads from recordings of

chainsaws; recruit a live choir by open call to heavy-metal vocalists singing in fictional

languages; put heavy riffs at very low tunings at the centre rather than at the edge. The

implementation is equally load-bearing — instead of playing pre-recorded tracks, Doom uses a modular

system of layers and loops combined on the fly, with the score written as "verse-like" and

"chorus-like" blocks the game selects and jams between according to combat intensity.

Two things transfer. Our endgame already asks for this register — Ch 75's music_mood names "a

marquee war register for the peopled front... and a terminal eschaton-dragon register... intensity

peak" (CH_75.md:269). And Gordon's per-project relearning is our principle 1, compose the structure

before the sound, reached from a completely different direction.

2.6 Generative and adaptive-first writing

The honest state of the art, per the 2025 survey literature and the interactive-music textbooks: the

professional pipeline is middleware-mediated (Wwise, FMOD) with vertical remixing and horizontal

re-sequencing as the standard techniques, and PROCEDURAL generation in 2025 is most common for

ambient layers, exploration textures and background soundscapes rather than for melodic themes.

Research systems generating continuous adaptive streams across mood and tension exist — the

Progressive-Adaptive Music Generator at the 2024 ACM FARM workshop is current — but they are

research rather than shipped practice for thematic material.

The finding is a warning rather than an opportunity: THE FIELD'S OWN CONSENSUS IS THAT GENERATIVE

SYSTEMS DO NOT YET CARRY THEMES. That is the conclusion our own melody A/B reached by measurement —

authored composition delivers the brief's licensed leap 100% of the time against 12.5% from

sharpened text-only generation — and it is why the melody-first ruling reads as industry-aligned

rather than as a local preference.

2.7 The diegetic instrument

The player performing music inside the fiction is a durable rather than new trend, and its power is

the loop between the two layers: in Ocarina of Time the songs Link learns and plays diegetically

recur as non-diegetic score in the same or later areas, so the player's own performance becomes the

world's theme. Genshin Impact's Windsong Lyre extends the idea into a free-play instrument.

This one is owed by our own canon and is unpaid. T0_Theme_Registry carries a diegetic_surface

column, minted 2026-08-05 and DELIBERATELY UNPOPULATED on every row, and

LEITMOTIF_ARCHITECTURE.md § 5 records the outstanding item: a plant node and a diegetic surface

for each Tier-A theme, PROTAGONIST_THEME first, with the note that docs/spine/CH_PROLOGUE.md's

# music_mood describes a register with no diegetic performance surface in it at all. The

competitive read raises the priority of that debt; it does not change its owner, which is the lane

that reads the node.

2.8 What the competitive read says the next bar actually is

Stated as a proposal. The bar is not "more orchestra". Across the 2023-2025 window the scores that

won COMMITTED to a specific human sound and built the whole work out of it — one vocalist, one

ensemble, one guitar, one choir of metal vocalists — and let the adaptive system serve that

commitment rather than generate around it. A seventy-nine-node game cannot make one such commitment.

What it can do, and what section 5 argues it must, is make ONE MOTIVIC commitment and re-commit it

in every idiom it enters. That is harder than any single-idiom score and it is the only version of

"next bar" available to a game with our geography.

3. The electronic and techno volume

This is the gap. Volume 1 of the doctorate and its six lanes cover orchestral and cyclic-traditional

structure thoroughly and say almost nothing about the electronic half. This section is longer than

the others for that reason.

3.1 The grid, and what four-to-the-floor actually is

The base pattern of house and techno is a kick on every quarter of a 4/4 bar — four to the floor —

with an open hi-hat on the offbeat eighths and a clap or snare on two and four. Its function is not

rhythmic interest. It is a METRONOMIC REFERENCE that frees every other part to be metrically

ambiguous without the listener losing the beat, which is the point Mark Butler's Unlocking the

Groove: Rhythm, Meter, and Musical Design in Electronic Dance Music (Indiana University Press, 2006)

is built on: EDM's rhythmic interest lives in the interaction between a fixed reference pulse and

parts that contradict it. Butler is the scholarly anchor for this section.

The variants carry different affect:

• Four-to-the-floor — the reference pulse. Forward motion that costs no attention.
• Broken beat and breakbeat — the kick displaced off some downbeats. Instability held under a

groove; the body still moves but cannot settle.

• Half-time — the backbeat moved to beat three so the tempo reads at half its actual value. Weight.

The single most transferable device to boss scoring, because it doubles apparent scale without

changing the tempo the systems are synchronised to.

• Two-step and the shuffled grid — swing applied to the sixteenths. Humanity inside a quantised

texture.

• No kick at all — the breakdown state, whose power is structural rather than rhythmic (3.3).

3.2 The phrase grid, and why it is a GRID rather than a phrase

Dance music is built from phrases of 8, 16 or 32 bars, and arrangement changes land on those

boundaries; a change on the wrong bar loses the floor. This is the strongest structural difference

between the electronic idiom and everything else in section 1. In an orchestral cue a phrase is a

MELODIC unit whose length is chosen by the tune. In techno the 8/16/32 hypermetric grid EXISTS

BEFORE THE MATERIAL and every element is placed against it. The composer does not decide how long a

phrase is; the composer decides what changes at bar 17.

Two consequences, both proposals:

• THE GRID IS A NOVELTY SCHEDULE WITH THE TIMES ALREADY FILLED IN. Our principle 4 asks for a tide

of small treatment changes every four to eight bars under about three landmarks. In techno that

tide is not a target, it is the form: something changes at every 8- or 16-bar line by convention,

and our own corpus derivation found the loved tracks changing something every 5.79 bars. The

techno grid and our measured corpus band are the same number reached from two traditions — the

second such convergence this program has found, after the 3-to-5-cycle law and the game-audio

loop-fatigue literature.

• THE GRID MAKES THE PLAN FALSIFIABLE FOR FREE. A cue whose plan says "filter opens across bars

33-48, kick returns at 49" is checkable against the render by an instrument that knows nothing

about melody. Our pass-2 architecture demands exactly that comparison and has to work much harder

to get it from an orchestral cue.

3.3 The arrangement grammar, and its DJ-tool origin

The canonical section order is intro, build, breakdown, drop (or main), and outro, and the reason it

looks the way it does is that the music is made to be MIXED. Dance records are arranged so that

another DJ can beatmatch in, ride the groove, and blend out; house and techno tracks routinely open

with 32 to 64 bars of drums and bass with no melodic commitment for precisely that purpose, and

close the same way. A classic techno intro of 64 bars is around two minutes at 130 BPM, introducing

one new element every 16 or 32 bars.

The two facts under that which matter to a game score:

• TECHNO IS LINEAR AND EVOLVING RATHER THAN DRAMATIC. Unlike the mainstage EDM shape, techno's main

section typically evolves slowly rather than delivering a sudden drop, with the producer

introducing small variations, removing elements and modulating sounds to hold interest. That is

our principle 5 (the law is non-decay, not escalation) restated by a different tradition, and it

is why techno rather than EDM is the right electronic reference for a score whose cues loop.

• THE DJ-TOOL INTRO IS THE SAME OBJECT AS THE ADAPTIVE ENTRY POINT. A 64-bar intro exists so a

track can be ENTERED from another track. A game cue's loop-in and transition stems exist so a cue

can be entered from another cue. The idiom has spent forty years engineering the exact problem our

middleware layer has, and it solved it by making the entry and exit regions melodically empty and

rhythmically complete. PROPOSAL: our cue table should carry an explicit entry_region_bars and

exit_region_bars for every looping cue, and the composed loop seam that loop_seams.py already

writes is the orchestral case of the same device.

• AND THE THIRD FACT, WHICH IS THE ONE THIS DOCUMENT MOST WANTS ON THE RECORD. The techno

arrangement grammar and the Javanese irama ladder are the same engineering. Irama re-tiles a fixed

cycle at separable density levels; the techno arrangement re-tiles a fixed loop at separable

element counts. Both change the DENSITY of a fixed structure rather than the structure. Our

doctorate's principle 7 already takes the density-level shift as the default landmark on the

strength of irama and Carnatic laya; this section adds a third independent tradition reaching the

identical ladder, and it is the tradition our endgame nodes need. Two of our slice chapters are

gamelan cultures and several of our modern chapters need techno, and the SAME structural device

serves both. That is not a coincidence to note; it is the seam solution of section 5 discovered in

the rhythm layer.

3.4 The tension devices, named one by one

These are the idiom's working vocabulary, and their orchestral equivalents are given so the score

can cross between them without changing what it means.

• FILTER AUTOMATION. A low-pass cutoff swept open across a build, or closed to strip a texture.

Orchestral equivalent: a gradual mute-to-open, or entries added from the bottom of the register up.

Our principle 6 classifies a bare filter sweep as CHEAP, and that classification stands — a sweep

that adds no followable voice satisfies nothing. What redeems it in techno is that the sweep is

usually the vehicle for an ARRIVAL, not a substitute for one.

• THE RISER, ALSO CALLED THE UPLIFTER. A rising noise or pitched sound that grows across a build.

Function: telegraphing an arrival so the arrival can be bigger. Orchestral equivalent: the

ascending string run, the cymbal roll.

• THE DOWNLIFTER. The same device reversed, placed at the START of a build to mark that the build

has begun. Orchestral equivalent: the descending harp gliss into a new section.

• THE IMPACT. A single loud broadband hit at a section boundary, often a kick, crash and reverbed

noise together. Function: masking the seam so the transition reads as fluent rather than as an

edit. Orchestral equivalent: the tutti accent at a double bar.

• THE REVERSE. Any sound played backwards to lead into its own onset. Function: an anacrusis that

costs no new material.

• THE BEAT-STOP. The drums removed for one to four bars, everything else continuing. This is the

strongest device in the idiom and it is a SUBTRACTION, which is our own principle 9 stated in a

different vocabulary. The corpus finding that removal creates more energy than addition is the

same fact that makes a breakdown work.

• THE DROP. A build, then a removal held for at least a bar, then a re-entry that restores what was

taken. Our own doctorate § 6 item seven names this as the highest-value derived feature available

from data we already hold, because the card carries every component (raises[].end_s,

drops[].at_s, bars_held_s, re_enters_at_s) and nothing computes the PAIRING. The techno

literature's definition and our card's field set describe the identical object.

3.5 The bassline as the melodic voice, and what the 303 actually does

In most electronic idioms the lowest sustained voice carries the tune. This is not a stylistic

inversion for its own sake; it follows from the four-to-the-floor kick occupying the lowest

transient band and leaving the 60-250 Hz sustained region as the most audible melodic slot in a

mix whose mid-range is given to percussion and texture.

The Roland TB-303 is the specific instrument that made this a compositional method, and its

mechanism is worth stating exactly, because "acid" is usually described by its sound rather than by

what produces it:

• One oscillator, sawtooth or square. The sawtooth is the acid tone.
• A four-pole (24 dB/octave) resonant low-pass filter. At high resonance the filter self-oscillates

near the cutoff, which is the squelch.

• A single decay envelope driving both the amplifier and the filter, so every note's timbre is a

function of its own length.

• ACCENT, per step: it raises the note's amplitude AND widens the filter envelope and cutoff

together. Roland's own term for the circuit was the "gimmick circuit". This is the single most

important fact in this subsection: accent is a JOINT dynamic-and-timbral operation, so in this

idiom loudness and brightness are not separable parameters.

• SLIDE, per step: the note glides to the next pitch instead of retriggering, producing legato lines

that slither between pitches. Combined with resonance this produces the screaming acid line.

• A 16-step sequencer, with accent and slide assignable per step.

What this means compositionally: the 303's expressive surface is a 16-step PATTERN plus two boolean

lanes (accent, slide) plus two continuous performance parameters (cutoff, resonance). The composer's

material is a short pitch cell; the composer's DEVELOPMENT is the placement of accents and slides

and the trajectory of the two knobs across the arrangement. That is a variation technique with a

finite, enumerable operation set, which is exactly the shape our program's colour_change

enumerated-device requirement already has.

3.6 Timbre as melody, and where the tune goes when there is no tune

In IDM and much ambient and dub techno the melodic function is carried by timbral evolution rather

than by pitch succession. The IDM lineage — Aphex Twin, Autechre — is described in the literature by

its rejection of standard rhythms, timbres and song structures in favour of intricate sound

manipulation and detailed beat programming, with headphone detail and algorithmic structure

displacing the dancefloor. Autechre's stated method of over-using a new instrument until its

capabilities are exhausted BEFORE composing with its sounds is the practice statement of the

principle: the sound is the material.

Dub techno is the clearest case because it is the sparsest. Basic Channel (Moritz von Oswald and

Mark Ernestus) built the idiom out of Robert Hood's minimal techno slowed down, drenched in reverb

and processed with Jamaican dub techniques. In a dub techno track the melody is usually a single

chord stab whose filter, reverb send and delay feedback change continuously across eight minutes.

Nothing transposes. The listener follows a line anyway, because a continuous change in spectral

centroid IS a contour.

The connection to the orchestral half is exact and already in our own material. The orchestration

lane's texture list describes Ligeti's micropolyphony as polyphony that is real and deliberately

inaudible as counterpoint, heard as a woven texture with the individual lines hidden — Ligeti's

clouds carrying etched internal detail in constant motion where Penderecki's blocks stay eventless.

A dub techno track is a micropolyphonic texture with a beat under it, and both hold attention for

eight minutes for the same reason: SUBTLE MOTION INSIDE A STATIC OBJECT.

3.7 The sub-genres that matter to us, and what each is FOR

• DETROIT TECHNO (Juan Atkins, Derrick May, Kevin Saunderson; then Underground Resistance, Robert

Hood and Jeff Mills as the second wave). Function: futurism with melancholy in it. The second wave

distilled the music to hypnotic loops, stark textures and a sense of alienation. This is the

correct reference for our modern-arc megacity register, not mainstage EDM.

• MINIMAL TECHNO (Robert Hood). Function: process made audible; almost nothing happens and the

nothing is the content. The reference for the atemporal dimensional nodes.

• ACID (the 303). Function: an organic, screaming, living line inside a mechanical texture. The

reference for anything that is alive but not human.

• INDUSTRIAL. Function: the machine as an antagonist; distortion and metallic percussion as

material. Yamaoka's Silent Hill is the game-scoring case, and our own Ch 13 music_mood already

reads "reverent-to-industrial" (CH_13.md:246).

• IDM. Function: intelligence displayed as complexity; the score thinking visibly.
• DUB TECHNO (Basic Channel, Maurizio). Function: depth and fog; space as the subject. The

reference for submerged and drowned nodes.

• SYNTHWAVE. Function: a remembered future; nostalgia for a technology. Handle with care — its

signifiers are the strongest cliché in the modern electronic palette.

• AMBIENT TECHNO. Function: a pulse you can ignore; presence without demand. The reference for

long modern-arc traversal.

3.8 What MELODIC INTELLIGENCE means in an idiom whose melody may be a filtered bassline

Josh's round-2 complaint was that the melodies are too simple and that intelligence has to go into

the melodies and into how everything orchestrates. That complaint has to be answerable in this

idiom, and it is. Stated as this document's central proposal:

• FIRST, THE MELODIC OBJECT IS A CELL PLUS AN OPERATION SET, NOT A TUNE. In an orchestral cue,

melodic intelligence shows as development — the subject inverted, augmented, fragmented,

reharmonised. In techno the cell is 4 to 8 pitches on a 16-step grid and the development happens

in a DIFFERENT PARAMETER SPACE: accent placement, slide placement, step-mute placement, octave

displacement of single steps, and the cutoff and resonance trajectory. A track that states one

16-step cell and runs thirty distinct, ordered operations on it across seven minutes is doing

exactly what a Beethoven development does, in a space where pitch is held and timbre moves. A

track that states one cell and loops it unchanged is the same failure as an orchestral cue that

restates its theme verbatim, and it fails for the same reason.

• SECOND, THE SPECTRAL CONTOUR IS A CONTOUR AND SHOULD BE GRADED AS ONE. Our nostalgia rubric grades

contour singability on a pitch contour. In an idiom where a filtered line's spectral centroid rises

across sixteen bars and falls across eight, that trajectory has a shape, a range, a peak position

and a deviant in exactly the sense our card already measures pitch contours. PROPOSAL: add a

spectral_contour block to the identity card computed from the spectral centroid of the isolated

bass band, with the same fields the pitch contour carries, and let an electronic cue be graded on

it. This is cheap — the STFT is already computed for every card — and it converts the largest

measurement blind spot in the program into a graded axis.

• THIRD, HARMONIC INTELLIGENCE SURVIVES INTACT AND IS THE PART MOST OFTEN SKIPPED. Detroit techno's

melancholy is harmonic: minor sevenths and ninths, modal mixture, chord stabs whose voicings imply

more than three notes. Nothing about the idiom forbids a real progression, and the tracks that are

remembered thirty years later are the ones that have one. Our own head-cell law and harmonic

signature fields apply unchanged. The intelligence deficit Josh named is therefore NOT excused by

the idiom in any of its three parameters — cell development, spectral contour, or harmony — and

the correct response is to grade all three rather than to lower the bar for the electronic cues.

4. The mythological and otherworld register

Our five mythological nodes are Hades and Elysium (Ch 70), Xibalba (Ch 71), Naraka and Meru (Ch 72),

Annwn (Ch 73) and the Nine Worlds of Yggdrasil (Ch 74), with the fairy realm at Ch 75-77 and four

atemporal dimensional nodes at Ch 66-69. Their spine music_mood lines have already chosen the

affects: elegiac and grave with a luminous Elysium border (CH_70.md:200); taut, ludic,

dread-and-play together for the rigged game (CH_71.md:222); rising and liberatory for the release of

the waters (CH_72.md:204); warm and honourable turning grave at the ford (CH_73.md:196); grave, vast

and threatened for the world-tree under assault (CH_74.md:204). Every one of those lines ends with a

protected-liturgy exclusion.

So this section answers a narrow question: HOW DOES A SCORE SIGNAL THE NON-HUMAN WITHOUT REACHING

FOR THE DEFAULTS, when the affects are already assigned and the obvious material is forbidden.

4.1 The five devices, and what each actually signals

• EXTENDED AND PREPARED TECHNIQUE. Sul ponticello, col legno, behind-the-bridge, bow-tap, and the

prepared piano's felt, metal and rubber insertions. What it signals: a FAMILIAR instrument

behaving wrongly. The strongest and cheapest otherworld signal available, because the listener

recognises the instrument and cannot place the sound. Penderecki invented notation for most of it

in Threnody and film has used it as the sound of the disturbed imagination ever since.

• MICROTONALITY AND JUST INTONATION. Any system outside twelve-tone equal temperament, and

specifically just intonation's derivation of every interval from small whole-number ratios — the

ratios of the overtone series of any vibrating object. Orchestral strings and winds play

microtonal gradations with technique rather than with special instruments. What it signals is the

interesting part: just intonation reads as MORE natural than equal temperament because its

intervals beat less, so the signal is not strangeness but an uncanny PURITY — a world whose

physics are more consonant than ours. Equal-tempered music arriving after a just passage sounds

slightly sour, which is a homecoming device nobody has to explain to the player.

• CHOIR TREATMENT. The signal is not the choir; it is what the choir is doing. Wordless vowels,

whispered and spoken layers, extreme registers, divisi into clusters, men's and women's sections

used as separate timbres rather than as an SATB block. VOICE_COLOUR_SIGNATURE already forbids

liturgical text, quoted chant and reproduction of any living sacred vocal practice, which removes

the lazy option and forces the interesting one.

• INVENTED LANGUAGE. Established practice with real precedents: Skyrim's thirty-voice choir in the

game's Draconic language; Ar tonelico's Hymmnos, constructed by Akira Tsuchiya with emotion as its

explicit design emphasis; Doom Eternal's fictional-language choir. It buys a vocal-led cue that

cannot be mistaken for anyone's real liturgy, which is exactly the constraint our five

mythological nodes impose. PROPOSAL: ONE constructed vocable set for the otherworld family,

invented once and reused across the realm nodes with per-realm phonetic colouring — a different

invented language per realm buys nothing a listener can hear and costs five times as much.

• GRANULAR AND SPECTRAL PROCESSING. Spectral music (Grisey, Murail, the Groupe de l'Itinéraire)

derives harmony, timbre and form from acoustic analysis of sound spectra rather than from melodic

or rhythmic structure, and Grisey's instrumental synthesis writes an orchestra to reproduce an

analysed spectrum. Granular processing shatters a source into short grains and rebuilds it at a

different rate, pitch or density. Both signal a sound whose INTERIOR is visible — the most literal

available musical statement of a realm where the substrate of things is exposed, which is what our

dimensional nodes are.

4.2 The rule that keeps this honest

FAM_THE_OTHERWORLD's deny list already carries it: a realm register may never be built by

exoticising a living human tradition. The failure mode is specific and common — reaching for a duduk,

a shakuhachi or a throat-singing sample because it sounds "otherworldly", which converts a living

people's music into a signifier for the non-human. The devices in 4.1 are listed in the order they

are BECAUSE none of them requires borrowing anyone's instrument: extended technique, microtonality,

choir treatment, invented vocables and spectral processing are all operations on material we already

have licence to.

One gap found and recorded rather than applied: FAM_THE_OTHERWORLD carries no head_cell_spec

interval specification at all — its cell field holds prose explaining why the row is deliberately

not cardinal. Every other family head carries a shape. If the realm family is going to be scored

microtonally, the absence of an interval spec is the field where that decision would live, and

section 5.4 gives the constraint it would have to satisfy.

5. The seam problem

A player travels from a Paleolithic node to a technological realm inside one continuous game. Ten

idioms is not a feature; it is a threat to the work's identity. This section is the answer.

5.1 Why it is the hardest problem in the score

Because the failure is invisible per cue. Every individual cue can be excellent in its own idiom and

the whole can still be a compilation rather than a score. Nothing in our instrument set measures

"does this feel like one work", and nothing in any instrument set does. The only defence is

STRUCTURAL: build the coherence into the material so that it survives regardless of who realises it.

5.2 The four cohering devices, ranked by strength

• A SHARED MOTIVIC CELL THAT SURVIVES RE-IDIOMISATION. The strongest device by a wide margin, and

the one our canon already selected. The public precedent our own registry names is the Elder

Scrolls line from Nerevar Rising to Dragonborn. The purest demonstration in the medium is

Civilization VI (Geoff Knorr with Roland Rizzo, Griffin Cohen and Phill Boucher), where each

civilisation's theme exists as four arrangements of one folk melody across the Ancient, Medieval,

Industrial and Atomic eras — a single-instrument melody in the Ancient era becoming a full modern

composition by the Atomic. That is our exact problem solved at scale in a shipped AAA title, and

it is the single most useful comparator this lane found.

• A PERSISTENT TIMBRAL SIGNATURE. One sound present in every idiom, doing nothing but being there.

Ours exists and is already a registry row: VOICE_COLOUR_SIGNATURE, motif_class = timbre,

wordless choir, allowed in "each node's own vocal practice where its register card admits one".

A human voice is the one carrier that crosses all ten idioms without anachronism, because voice

predates every instrument in the map and postdates none of them.

• A HARMONIC LANGUAGE THAT BENDS RATHER THAN SWITCHES. Modal writing that can be read as Dorian by

a medieval cue, as a raga-adjacent set by a South Indian cue, as a minor-seventh-heavy Detroit

palette by a techno cue, and as a pentatonic subset by a gamelan cue. This is achievable because

those four descriptions can all be satisfied by ONE pitch-class set, and choosing that set once is

cheaper than reconciling four.

• TRANSITIONAL AND HYBRID CUES AT THE SEAMS. The travel seams are already canon —

memory death-choice-resurrection-and-travel-rulings records travel as compressed playable

segments across 78 seams. A seam cue that starts in the departing idiom and ends in the arriving

one, on the same cell, converts the biggest coherence risk in the game into its most

characteristic feature. PROPOSAL: the seam cue is a cue CLASS, not a per-node decision.

5.3 The head-cell constraint — what must be true of a cell to survive a gamelan, an orchestra and a 303

This is the section the lane was chartered for. The constraint is derived by taking one cell and

asking what each of the three carriers destroys.

A GAMELAN destroys equal temperament. Slendro and pelog are not 12-TET and no two gamelan are

identically tuned; intervals shift by tens of cents from ensemble to ensemble. It destroys harmonic

context — the balungan is a single line and the elaboration is heterophonic, so nothing supports the

cell from underneath. And it constrains compass: a saron's range is limited and octave displacement

is not free.

AN ORCHESTRA destroys nothing but forgives everything, which is its own hazard — an orchestral

statement can make a weak cell sound strong through orchestration alone, and the doctorate's

principle 10 already names the test (if it needs its orchestration to be recognisable it is not a

motif; test it through three carriers before spending anything else on it).

A 303 destroys polyphony — it is monophonic. It destroys precise interval perception at high

resonance, because the self-oscillating filter adds a pitched component that competes with the

fundamental. It destroys free rhythm — the sequence is 16 steps. And it destroys sustain — every

note's amplitude and brightness decay together on one envelope, so a cell that depends on a long

held note has no held note.

The intersection gives SEVEN constraints. Marked as this document's proposal, offered to the

head-cell law rather than applied to it:

• ONE. THE CELL'S PITCH SET MUST BE EXPRESSIBLE AS A FIVE-NOTE ANHEMITONIC SUBSET WITHOUT LOSING

ITS IDENTITY. A cell whose identity requires a semitone cannot be stated in slendro. Triad-

outlining already satisfies this, which is why JOURNEY_WORLD's existing spec is correct and this

constraint confirms rather than changes it.

• TWO. IDENTITY MUST BE CARRIED BY CONTOUR PLUS ONE INTERVAL, NOT BY THE INTERVAL SERIES. Contour

survives retuning; a series of exact intervals does not. The one interval doing identity work must

be one whose SIZE reads even when mistuned by 30-50 cents — an octave, a fifth, or a leap of a

minor sixth or wider. FAMILIAR_BOND's ascending sixth into the third note satisfies this

exactly and would survive all three carriers unchanged.

• THREE. THE ESSENTIAL SHAPE MUST FIT INSIDE A FIFTH, WITH ANY WIDER SPAN CARRIED BY OCTAVE

DISPLACEMENT OF SINGLE NOTES RATHER THAN BY THE CELL'S OWN RANGE. This is what lets a saron and a

monophonic bass line state the same figure.

• FOUR. THE RHYTHM MUST BE A SUBSET OF A SIXTEEN-STEP GRID. No triplet, no rubato, no unnotatable

placement in the IDENTITY. Triplets and rubato remain fully available as VARIATIONS, and in the

orchestral idiom they should be used; they simply may not be the thing that makes the cell

recognisable.

• FIVE. THE CELL MUST BE RECOGNISABLE UNACCOMPANIED. It may not require a functional bass or a

specific harmonisation to read as itself, because neither the balungan nor the 303 line has one.

This is checkable today: state the cell as a solo line with no accompaniment and run the existing

hook-presence instrument on it.

• SIX. THE CELL NEEDS A FOOT — a returned-to pitch that acts as the anchor. Both non-orchestral

carriers establish identity by return-to-tone rather than by cadence, because neither has cadential

harmony available.

• SEVEN, AND THIS IS THE CLAUSE THE EXISTING HEAD-CELL LAW DOES NOT CARRY. A CELL REQUIRED TO

MIGRATE ACROSS TUNING SYSTEMS MUST TAKE A METRIC OR REGISTRAL DEVIANT, NEVER A CHROMATIC ONE. The

head-cell law licenses exactly one deviant per cell. A BORROWED-DEGREE deviant — a chromatic

alteration — does not exist in slendro, is not reliably audible through a resonant filter, and is

the one deviant class that cannot cross. A METRIC deviant (an entry displaced off the downbeat) and

a REGISTRAL deviant (an unusual leap) both survive every carrier in the map.

5.4 What this feeds back into how we write head cells

Checked against the live registry, and the result is mostly a confirmation, which is worth saying

because it would have been easy to manufacture a defect:

• JOURNEY_WORLD — 4-note, triad-outlining, one unusual leap as the deviant. PASSES all seven. It

is the migrating motif and it was specified correctly, for the stated reason.

• FAMILIAR_BOND — 3-note, triad-outlining, ascending sixth as the deviant, and the row already

states that every per-species variant keeps that interval. PASSES all seven, and the constraint

explains WHY keeping that interval is the right rule rather than an arbitrary one.

• HOME_AND_LOSS — 5-note, purely stepwise, with a METRIC deviant (begins on the second beat, never

resolves onto a downbeat). Passes constraint seven exactly. Its stepwise construction makes

constraint one tight rather than failing — a five-note stepwise cell is diatonic, so it needs a

pelog-subset check before any gamelan statement. Recorded as a check to run, not as a defect.

• PROTAGONIST_THEME — 4-note, purely stepwise, with a BORROWED-DEGREE deviant on the third note.

That deviant is chromatic, so by constraint seven it does not cross tuning systems. THIS IS NOT A

DEFECT, and the row itself says why: its instrumentation_substrate is "orchestral with solo

instrument leitmotif evolution by age stage", and its migration axis is AGE, not region. The

protagonist's theme is not the motif that visits every register — JOURNEY_WORLD is, by design,

and that division of labour is exactly what makes the chromatic deviant affordable. The finding is

a positive one: the roster's two migrating motifs both carry crossing-safe deviants and the one

that carries a chromatic deviant does not migrate.

• FAM_THE_OTHERWORLD — no interval spec at all, per section 4.2. This is the one row where the

constraint has nothing to check, and it is also the family that will need a microtonal register.

Owed before the realm cues are composed.

The general rule this produces, offered to the head-cell law: THE DEVIANT CLASS SHOULD BE CHOSEN

FROM THE MOTIF'S MIGRATION AXIS. A motif that migrates across CULTURES takes a registral or metric

deviant; a motif that migrates across TIME or SCALE inside one register may take a chromatic one.

That is one sentence and it is checkable from two columns the registry already has.

6. Pipeline hooks — what our stack renders today, what it cannot, and what it would cost

6.1 What is actually there

Verified at HEAD rather than quoted. The realisation chain is authored MIDI from

harness/music_gen/theme_compositions.py and author_head_cell.py, rendered by

harness/music_gen/orchestrate.py through sfizz 1.2.3 (BSD-2-Clause, not redistributed) driving

VSCO 2 Community Edition and VCSL (both CC0), with the permitted instrument list closed in

harness/music_gen/sfz_palette.py. That file's header records the care reason the world-instrument

half of VCSL is deliberately absent: none of the realised artifacts is a region cell, so a family

head realised on an mbira would be scoring a living tradition against a card that has not authorised

it. ACE-Step is available for BEDS ONLY under the melody-first ruling. Below the sampler sits an

additive synthesis engine inside author_head_cell.py — a harmonic sine bank with ADSR and vibrato,

kept as the honest sketch tier.

The libraries present in the environment, checked directly: numpy 2.4.6, scipy 1.18.0, librosa

0.11.0, soundfile 0.14.0, pretty_midi, mido. All BSD-3 or MIT.

6.2 The per-idiom hook table

6.3 The synthesis gap, stated exactly, and the four options

THE GAP, STATED PRECISELY. Our pipeline contains no subtractive synthesis. Checked directly:

author_head_cell.py's engine is an additive sine bank with an ADSR envelope and vibrato, and

grepping the music_gen tree for a filter, a noise source, a cutoff parameter, resonance, distortion

or saturation returns nothing. A techno cue is MOSTLY filter, noise, resonance and saturation, and

the TB-303's entire identity is a resonant four-pole low-pass filter with a shared decay envelope

and per-step accent and slide — we can produce none of those five things. We can produce the pitch

sequence and nothing that makes it the idiom. This is the largest single capability gap in the music

pipeline and it is larger than the orchestral one, because the orchestral gap is a QUALITY gap

(chamber samples instead of a large library) and this is a CATEGORICAL one.

The four options. The licence reads follow the standing law — face value, only explicitly triggered

prohibitions flagged.

• OPTION A — BUILD A SUBTRACTIVE VOICE IN NUMPY AND SCIPY, INSIDE THE EXISTING ENGINE. Buys:

oscillators (saw, square, pulse with width), a resonant state-variable or ladder low-pass whose

cutoff and resonance automate per step and per bar, a noise source, an ADSR shared between

amplitude and filter as the 303 shares it, per-step accent and slide, and a saturation stage —

the complete 303 mechanism plus the analogue-drum primitives (a pitched sine with a fast pitch

envelope is a kick; filtered noise with a decay envelope is a hat). Licence: numpy and scipy are

BSD-3 and both are already installed, so nothing is downloaded and no new licence is read. Cost:

engineering only, on the order of one focused sitting. Limit: not a boutique analogue emulation,

and wavetable or FM depth is further work.

• OPTION B — ADOPT SURGE XT. Already surveyed in docs/pipeline_review/tech_research/PIPE_AUDIO_MUSIC_2026-07-29.md

§ 8.1 option D, with the licence read recorded: GPL-3.0, the copyleft binds the software and not

the audio a user produces with it, so there is no output-licence question and no triggered

prohibition. Surge XT gained a headless command-line build at 1.3, which is what makes it usable

by a battery at all. Cost: a binary on the box (not redistributed), plus the real problem — a

Surge patch is not an artifact we AUTHOR from a plan, so plan-versus-render falsifiability gets

weaker rather than stronger.

• OPTION C — CC0 DRUM-MACHINE SAMPLES THROUGH THE EXISTING SFIZZ PATH. Genuinely CC0 TR-808 and

TR-909 sets exist (the free-drum-samples repository under CC0 1.0, with samples derived from

Edward Loveall's CC0 TR-808 recordings; Producer Space's CC0 packs). Cost near zero, and the sfizz

path that plays VSCO 2 plays these unchanged. What it does NOT buy is the idiom: a 909 hit with no

filter movement over it is a drum sound, not techno.

• OPTION D — GENERATE TECHNO BEDS WITH ACE-STEP. Available under the existing bed law and REFUSED as

a melodic source by the melody-first ruling and the measured A/B. Stays where it is: beds only.

THE RECOMMENDATION, this document's proposal. OPTION A AS RUNG ONE, WITH OPTION C ALONGSIDE IT FOR

THE PERCUSSION LAYER, AND OPTION B HELD AS RUNG TWO ON EVIDENCE. The deciding argument is not cost,

though A is cheapest — it is FALSIFIABILITY. Under option A every filter automation point, every

accent, every slide and every element entry is a SCHEDULED EVENT WE WROTE, so the cycle-law

instrument, the layer census and the novelty schedule grade an electronic cue on exactly the terms

they grade an orchestral one, and the pass-2 contract that the plan must be falsifiable by the

render survives into the idiom. Under option B the patch is opaque and the comparison degrades to

listening. The strongest objection to A is that hand-rolled DSP will sound worse than a mature

synthesiser; the answer is the one the orchestral rung already gave — the deliverable is the SCORE

and the plan, every renderer consumes the same event list, and changing the renderer changes nothing

upstream of it.

TWO MEASUREMENT DEBTS THIS SECTION ADDS, both cheap and both unpaid:

• THE SPECTRAL CONTOUR BLOCK of section 3.8, from the spectral centroid of the isolated bass band,

with the same field shape as the existing pitch contour. Without it the card cannot see a melody

in any electronic cue, and seven of our thirty measurable exemplars are electronic-led.

• THE DROP PAIRING of doctorate § 6 item seven, which this section's arrangement-grammar reading

independently confirms as the highest-value derived feature available from fields we already hold.

7. Honest limits of this document

• IT IS RESEARCH, NOT A LISTENING TEST. Not one claim here has been checked against a human ear.

Josh is the first ear and remains so.

• THE IDIOM MAP IS A PROPOSAL AND ONLY ITS ERA INPUT IS CANON. Where the map assigns an idiom to a

node whose spine music_mood line does not name one, the assignment is this document's

recommendation, marked as such in 1.1. Nothing in the map has been written to any registry.

• THE ELECTRONIC LITERATURE IS PRACTITIONER-DOMINATED. The one scholarly anchor found for EDM rhythm

and metre is Butler; the arrangement-grammar and transition-FX material is production-community

consensus rather than peer-reviewed evidence, and is labelled as consensus where it appears. The

303 mechanism is the exception — manufacturer-documented and hardware-analysed, evidence tier.

• THE SEVEN HEAD-CELL CONSTRAINTS ARE DERIVED BY REASONING FROM ORGANOLOGY, NOT MEASURED. They are

cheap to falsify — state one cell through the three carriers and run the existing hook-presence

instrument on each — and that test has NOT been run. It should be, before the list is treated as a

law rather than a proposal.

• THE OPTION-A COST ESTIMATE IS AN ENGINEERING JUDGEMENT, and DSP that has to sound good has a way

of costing more than DSP that has to work.

• NOTHING HERE MEASURES WHETHER THE MUSIC IS BEAUTIFUL, which remains outside every instrument in

this program.

8. Sources

Web sources, all read 2026-08-07.

• EDM arrangement, phrase grid and DJ-tool origin — https://edm-ghost-production.com/dj-knowledge-base/electronic-song-structure · https://universeoftracks.com/the-ultimate-guide-to-techno-track-structure/ · https://www.mastrng.com/song-structure-arrangement/ · https://quadrophone.com/composition/arranging-dance-music/ · https://edmtips.com/edm-song-structure/ (consensus tier)
• Transition FX vocabulary — https://www.ujam.com/tutorials/how-to-create-huge-edm-transitions/ · https://www.pointblankmusicschool.com/blog/designing-unique-risers-and-fx-for-transitions-to-level-up-your-tracks/ · https://optoproductions.com/production-basics-22-transition-effects/ (consensus tier)
• TB-303 mechanism — https://www.musicradar.com/news/producers-guide-to-the-roland-tb-303-and-clones · https://articles.roland.com/beyond-acid-pushing-the-tb-303-into-new-sonic-territory/ · https://electronicmusic.fandom.com/wiki/TB-303 · https://www.eddybergman.com/2025/03/TB303-VCF.html (evidence tier)
• Detroit techno, minimal, dub techno — https://en.wikipedia.org/wiki/Detroit_techno · https://timeline.carnegiehall.org/genres/detroit-techno · https://techno-music.com/techno-subgenres-guide/ · https://www.normanrecords.com/features/genre-guides/best-techno
• IDM and timbre-led composition — https://genrematic.com/post/intelligentdancemusic/ · http://econtact.ca/14_4/ramsay_acousmatic-idm.html · https://en.wikipedia.org/wiki/Aphex_Twin
• Scholarly anchor for EDM rhythm and metre — Mark J. Butler, *Unlocking the Groove: Rhythm, Meter, and Musical Design in Electronic Dance Music*, Indiana University Press, 2006 — https://archive.org/details/unlockinggroover00butl (evidence tier)
• Micropolyphony, sonorism and the horror inheritance — https://en.wikipedia.org/wiki/Micropolyphony · https://articulatesilences.wordpress.com/2012/07/27/threnody-to-the-victims-of-hiroshima-krzysztof-penderecki/ · https://the-artifice.com/penderecki-sound-horror/ · https://nofilmschool.com/how-exorcist-changed-sound-horror
• Aleatoric technique in game scoring — Michael Sweet, *Writing Interactive Music for Video Games: A Composer's Guide*, ch. 16 — https://www.oreilly.com/library/view/writing-interactive-music/9780133563528/ch16.html
• Spectral music — https://en.wikipedia.org/wiki/Spectral_music · Grove Music Online, "Spectral music"
• Microtonality and just intonation — https://theproaudiofiles.com/microtones/ · https://open.lib.umn.edu/musiccomposition/chapter/microtonality/ · https://newmusicusa.org/nmbx/intonation-and-microtonality/
• Constructed language in game vocal writing — https://en.wikipedia.org/wiki/The_Elder_Scrolls_V:_Skyrim_(soundtrack) · https://conlang.fandom.com/wiki/Hymmnos
• Current-window game scoring — https://www.videogameschronicle.com/features/best-game-music-2024/ · https://www.videogameschronicle.com/features/this-is-the-best-video-game-music-of-2025-as-chosen-by-the-composers-behind-it/ · https://acloserlisten.com/2025/12/11/press-a-2025-the-years-best-videogame-soundtracks/ · https://daily.bandcamp.com/best-of-2025/the-best-video-game-soundtracks-of-2025
• Clair Obscur: Expedition 33 — https://en.wikipedia.org/wiki/Music_of_Clair_Obscur:_Expedition_33 · https://www.expedition33.com/post/composing-the-music
• Historically informed game scoring — https://www.alkemie.org/pentiment-album · https://en.wikipedia.org/wiki/Pentiment_(video_game) · https://www.earlymusicamerica.org/web-articles/early-music-rocks-the-video-game-universe/ · https://www.nme.com/features/gaming-features/kingdom-come-deliverance-2-jan-valta-interview-warhorse-3902073 · https://magneticmag.com/2025/02/kingdom-come-deliverance-2-soundtrack-an-exclusive-interview-with-composer-jan-valta/
• Metal-orchestral hybrid and adaptive implementation — https://doomwiki.org/wiki/Doom_Eternal_soundtrack · https://ensigame.com/articles/gaming/secrets-of-dooms-music-mick-gordons-adaptive-soundtrack-and-its-impact-on-gameplay-2
• Dark ambient and drone in horror scoring — https://www.soundoflife.com/blogs/people/akira-yamaoka-silent-hill-music · https://gamemusic.net/drones-and-ambient-music-in-horror-games/
• Jazz in game scoring — https://www.siliconera.com/persona-5s-soundtrack-comes-acid-jazz-influences/ · https://daily.redbullmusicacademy.com/2014/09/yuzo-koshiro-interview/ · https://en.wikipedia.org/wiki/Music_of_the_Streets_of_Rage_series
• Generative and adaptive practice — https://www.gamedeveloper.com/audio/deep-dive-generative-music-in-video-games · https://dl.acm.org/doi/10.1145/3677996.3678291 · https://www.sciencedirect.com/science/article/abs/pii/S1875952119300795
• The one-melody-across-eras precedent — https://civilization.fandom.com/wiki/Soundtrack_(Civ6) · https://www.geoffknorr.com/sid-meiers-civilization-vi
• Diegetic instrument practice — https://www.popmatters.com/legend-zelda-music-embroidered-link · https://tvtropes.org/pmwiki/pmwiki.php/DiegeticSoundtrackUsage/VideoGames
• Ancient-music reconstruction and its limits — https://en.wikipedia.org/wiki/Hurrian_songs · https://en.wikipedia.org/wiki/Seikilos_epitaph · https://en.wikipedia.org/wiki/Armand_D%27Angour
• CC0 drum-machine samples — https://github.com/Boochi44/free-drum-samples · https://producerspace.com/
• Open-source synthesis, headless — https://github.com/surge-synthesizer/surge · https://surge-synthesizer.github.io/faq/ · https://cdm.link/free-surge-xt-1-3/

Repository sources, read at HEAD 2026-08-07: docs/spine/DECISIONS_PENDING_JOSH.md:5600 ·

registries/T0_Chapter_Index [ACTIVE v1.2]/T0_Chapter_Index DRAFT v1.1.csv (79 rows, era,

region_primary, music_mood_tags) · registries/T0_Theme_Registry [DRAFT v0.1]/T0_Theme_Registry.csv

(38 rows, 45 columns) · docs/spine/CH_02.md:234 · CH_10.md:224 · CH_11.md:237 · CH_13.md:246 ·

CH_15.md:263 · CH_17.md:248 · CH_21.md:248 · CH_23.md:249 · CH_28.md:252 · CH_30.md:247 ·

CH_38.md:224 · CH_39.md:279 · CH_42.md:257 · CH_43.md:245 · CH_44.md:220 · CH_46.md:207 ·

CH_47.md:209 · CH_53.md:243 · CH_55.md:207 · CH_56.md:212 · CH_60.md:227 · CH_66.md:205 ·

CH_67.md:202 · CH_68.md:194 · CH_69.md:213 · CH_70.md:200 · CH_71.md:222 · CH_72.md:204 ·

CH_73.md:196 · CH_74.md:204 · CH_75.md:269 · CH_76.md:212 · CH_77.md:237 ·

CH_EPILOGUE.md:182 · docs/proposals/music/MUSIC_CRAFT_DOCTORATE.md §§ 3, 5, 6, 8 ·

docs/proposals/music/LEITMOTIF_ARCHITECTURE.md §§ 3, 5 ·

docs/proposals/music/craft_research/ORCHESTRATION_TEXTURE_AND_DENSITY.md § 4 ·

docs/proposals/music/craft_research/REPETITION_AND_VARIATION_FORM.md §§ 4, 5 ·

docs/pipeline_review/tech_research/PIPE_AUDIO_MUSIC_2026-07-29.md § 8.1 ·

harness/music_gen/sfz_palette.py · harness/music_gen/author_head_cell.py ·

build/audio/exemplars/CORPUS.json (182 rows; 18 ACQUIRED, 12 CARDED, 152 MISSING) ·

build/audio/exemplars/formula_cards/EX_124.json.

Generated by harness/site/structure_site.py — the URL path is the repo path. review root