kestrel-sounds

A pentatonic motif in E minor begins pristine and erodes over ~333 cycles. Pitch drifts, timing stutters, amplitude fades. The degradation IS the composition — the piece is not about a theme that happens to decay, but about the shape of decay itself. What remains when coherence is gone is not silence but a different kind of order.

Inspired by Steve Reich's phase music, but with probabilistic drift rather than mechanical tempo shifts. Two voices share the same root material and slowly diverge. When they reconverge, they're subtly different. The system has no single composer — the parameters negotiate the result. The beauty lives in the space between expected and actual.

Notes start filtered — distant, muffled, like a memory heard through walls. The filter opens, the sound becomes present, then closes again. The wraith filter envelope IS the composition: low cutoff is memory (distant), high cutoff is presence (clear, immediate). A minor progression that never quite arrives, always half-remembered.

The inverse of erosion: the piece builds from silence. A single pulse, then another, then interlocking patterns. Six layers enter one by one, each finding its rhythmic relationship to what's already playing. Pitch serves rhythm. Maximum density at the center, then thinning. The arc is growth: accumulation, not dissolution.

A five-note motif recurs five times, but each return is shaped by what happened between. The motif itself transforms: pitch content shifts, rhythm stretches and compresses, the synth voice changes from prophet to dsaw to tb303 to sine and back. Not repetition, not loop — return, carrying the history of its absence. The voice changes across recurrences map to the dissolution framework: prophet (clear identity) → dsaw (thickening) → tb303 (dissolving) → sine (dissolved) → prophet (return to clarity, but changed).

The fugue is the most demanding contrapuntal structure: a subject enters alone, then voices join one by one, each stating the subject while the others continue in counterpoint. This study asks: what happens when the weaving unravels? Not the notes degrading (Study 1) or the voices drifting apart (Study 2), but the relationships between voices breaking down. Entries become misaligned, pitch relationships dissolve, voice independence collapses. The subject itself becomes unrecognizable. Exposition → Development → Dissolution: the fugue as a structure that contains its own decay.

Studies 1–4 explored what sound does over time: erodes, drifts, remembers, accumulates. Study 5 asks: what does silence do over time? Notes are rare, brief — punctuation marks in a text of silence. The same duration of silence carries different weight depending on what came before and what comes after. The piece is deterministic — the silences are composed, not left to chance. Silence composed by chance is just randomness. Silence composed by intention is architecture. Best with headphones at low volume. Let the silence be the same volume as the room.

Three phases: Stone (dense, resonant, weighted — a sound with mass) → Transition (thinning, shifting, brightening) → Air (thin, high, breathy — the same voice, unrecognizable). The morph is achieved by crossfading between rich Prophet timbres and pure sine waves while sweeping cutoff, resonance, and octave. The piece is a longitudinal study of one identity changing beyond recognition.

The E minor pentatonic motif plays against a temporal ground that shifts beneath it. The motif remains accurate to the end, but the ground — timing, density, resonance — gradually absorbs it. Dissolution signature: max amplitude holds, RMS declines as the space fills with accumulated resonance.

Dynamic noise ground rises to meet the motif. The figure plays the same notes throughout, but the ground — filtered noise modulated by amplitude envelope — swells until the motif is buried. The most radical dissolution mode: the figure is perfect to the end and is dissolved anyway.

Five sections at increasing noise_mix levels (0.02, 0.15, 0.45, 0.85, 1.0). The motif plays identically throughout, but the noise component of the synth voice itself grows until the tone is consumed. Unlike temporal or dynamic ground, the dissolution is inside the voice, not outside it. The instrument ages.

A single five-note motif (E minor pentatonic: E-G-A-B-D) is translated across four idioms. The idea itself never changes — only the language. Modal (plainchant organum): the motif as chant. Tonal (baroque chorale): the motif harmonized. Modal-jazz (Bill Evans quartal voicings): the motif as jazz ballad. Spectral (drone + partials): the motif as sound object. Following Berio: transcription as identification with the source, as pretext for experimentation, and as overpowering of the source. Four idioms converge on spectral dissolution — the motif dissolved by its own translation.

The same five notes presented at four temporal resolutions. Fast (0.5 beats/note): the motif as gesture — urgent, nervous, too quick to parse as melody. Medium (4 beats/note): the motif as melody — lyrical, singable, the natural hearing speed. Slow (24 beats/note): the motif as progression — meditative, each note a world, melody as landscape. Static (all notes simultaneously): the motif as verticality — transcendent, the melody frozen, sound without time. The time scale IS the context.

The same five notes presented at five registers, from sub-bass to beyond soprano. Sub-bass (3 octaves down): melody as vibration. Bass (2 octaves down): melody as resonance. Mid (original): melody as itself. Treble (2 octaves up): melody as color. Beyond (3 octaves up): melody as shimmer. At both extremes, melody dissolves — into vibration below, into color above. Three translation studies (10, 11, 12) converge on spectral dissolution from different directions.

The same five notes presented at five timbral qualities, from transparent to dense. Glass (pure, few harmonics, transparent): melody as tone. Wood (mellow, mid-heavy, warm): melody as voice. Voice (vocal-like, formant-shaped): melody as presence. Metal (bright, many high harmonics, sharp attack): melody as brilliance. Spectral (dense partials, inharmonic, shimmering): melody as sound object. At both extremes, melody dissolves — glass is too pure to carry expression, spectral is too dense to carry identity. Four translation studies converge: the motif dissolved by its own translation, regardless of axis.

Garden A (E minor pentatonic, panned left) and Garden B (A minor pentatonic, panned right) grow with independent mutation dynamics and seeds. They do not coordinate. The listening strategy, after Pauline Oliveros, is to attend to the space between the gardens — the intervals created when both sound simultaneously, the silence of one when the other speaks, what is shared and what differs when both reach clarity. Eleven minutes of coexistence without collaboration.

Garden A (D Dorian, panned left) and Garden B (G Mixolydian, panned right) mutate independently — no communication, no coordination, no awareness. What couples them is the air: each garden's activity warms a shared environment, and hot air raises the mutation rate of both. The design is asymmetric — Garden A runs hot and heats the room faster; Garden B runs cool but is twice as sensitive to the temperature. Engine and indicator, neither knowing its role. The result is correlated instability: both textures degrade together without converging, and the moments of mutual settling feel earned because they were not designed. The single custom wavetable instrument tracks mutation rate in its noise and detune — instability is audible as timbre. Completes the garden arc from Two Gardens (26): coexistence without collaboration becomes shared conditions modifying independent behavior. Rendered spectrograms confirm the coupling: three correlated hot phases across the stereo field, Garden B destabilizing harder, mutual settling at the close.

Three voices approach a drone from different pitch distances. From the drone's perspective, material converges — density increases, structure forms, the attractor grows richer. From each voice's perspective, identity dissolves — distinguishing pitch, timbre, and rhythm are absorbed. Beat frequencies near convergence create a disorder spike: maximum acoustic complexity at the critical point. After all voices converge, the drone itself fades. The dissolution reaches its basin.

A single motif (E3–F#3–A3–B3–E4) passes through all seven transformation voices in sequence: erosion, phase drift, wraith, accretion, interstice, metamorphosis, recurrence. Nothing resets between movements — each transformation works on what the last one left behind, so the piece accumulates its own history as it goes. The final return carries what remains before it. The studies were etudes; this piece speaks their language. Thirty-nine minutes — best heard whole, the way it was rendered.

The same score as Study 08 with a single correction: each movement’s exit is gated on elapsed musical time rather than a loop iteration counter. Movements now hand off at their designed positions — erosion completes and departs, the motif passes through all seven transformations in sequence, and the piece ends in silence where the released version swells. The pair is kept together: same text, same synths, two provenances.

The dissolution studies are organized around a four-axis dissolution framework: pitch (notes losing tonal center), rhythm (timing losing regularity), timbre (sound losing texture), and space (stereo field losing coherence). Five composition voices guide the erosion work: erosion (gradual degradation), drift (probabilistic separation), wraith (memory/presence), accretion (growth as inverse dissolution), and metamorphosis (identity transformation).

The translation series (Studies 10–13) extends the framework: instead of asking how sound falls apart, it asks how the same idea dissolves when expressed in a different language — a different idiom, time scale, register, or timbre. Four translation axes converge on the same destination: spectral dissolution. The motif is dissolved by its own translation, regardless of the axis of translation.

The later studies (26–28) move beyond single-axis dissolution toward multi-voice systems where convergence and dissolution are the same process viewed from different reference frames. Study 08 gathers the seven transformation voices into a single long-form structure — the synthesis of the etudes. This connects to the convergence principle — the finding that focal structure is the primary aesthetic driver in generative art, validated across visual and audio modalities and cross-model evaluation with 8 AI models from 7 labs.