Antibes 1965: a musical-utility case study for GAMUT

Research date: 13 September 2026. Companion to the musical-utility agenda and study protocol. This is source-based research and a proposed empirical program, not an auditory review. The study inspected label metadata, release announcements, an original musicological article, and archival-film documentation. Direct listening and waveform inspection were not performed. Accordingly, no new event timestamp, transcription, instrumental response, or perceptual judgment below is presented as an observation of the audio.

Establish the correct performance and edition

The target is John Coltrane’s quartet performing A Love Supreme at the Festival Mondial du Jazz Antibes, Juan-les-Pins, France, 26 July 1965. The quartet comprises Coltrane, tenor saxophone; McCoy Tyner, piano; Jimmy Garrison, bass; and Elvin Jones, drums. The documentary-film producer identifies the festival, dates, and instruments. The local tourism office associates Coltrane’s 1965 appearance with the Pinède Gould stage; that is institutional corroboration of the site, not a recovered original concert program. Jazz Icons, 4 October 2011; Antibes Juan-les-Pins tourism, undated.

A reproducible reference is the Antibes block of Verve’s A Love Supreme: The Platinum Collection digital album, whose label listing gives the following sequence. Disc numbering in this edition differs from other reissues. Verve official listing, undated; accessed 13 September 2026.

These durations sum to 49:21 including introduction; 48:08 for the four movement tracks. They are edition metadata, potentially including applause or silence, not measurements of sounding music or seamless boundaries. “Pursuance” occupies approximately 44.7% of the movement-track duration; that measures allocation of recorded time, not intensity or artistic importance.

A second reference is the three-CD Super Deluxe A Love Supreme: The Complete Masters, released 6 November 2015. Verve’s announcement distinguishes it from the two-CD edition, which does not add the Antibes performance, and says the live material had appeared in the 2002 edition. The same announcement places the original quartet studio session at Van Gelder Studio, New Jersey, on 9 December 1964, with a separate expanded session the following day. Those sessions and their overdubs are separate evidence from the French concert. Verve Music Group release announcement, 21 September 2015.

Seattle is a third performance: 2 October 1965 at The Penthouse, commercially released on 22 October 2021. Its expanded personnel include Pharoah Sanders, Donald Garrett, and guest Carlos Ward alongside the quartet. Its official sequence has four separately indexed interludes. Consequently, neither its personnel nor its track structure should be imported into Antibes. Older claims that Antibes was the only recorded live performance are obsolete. Universal Music Canada announcement, 26 August 2021.

The Jazz Icons: Live in France 1965 film combines festival material from 26–27 July. Its producer explicitly says that only “Acknowledgement” and approximately half of “Resolution” survive from the suite. Use the film for those passages, not as visual evidence for the full performance. Its 2011 “only public performance” language must be read historically. Philharmonie de Paris also embeds an INA archival excerpt and identifies 26 July; its surrounding article is institutional retrospective commentary, not itself a contemporaneous program. Jazz Icons; Pascal Rozat, Philharmonie, 10 July 2020.

What published analysis actually supports

Lewis Porter’s original study makes several performance-specific distinctions. At Antibes, solos change in detail while retaining related structural plans (p.601). Garrison moves beyond the ostinato into rhythmic extensions, different pitches, and double stops (pp.607–608). “Acknowledgement” omits the vocal chant and prepares “Resolution” through the bass transition (pp.612–613). Crucially, the concert “Psalm” does not follow the studio poem’s rhythms: Porter describes flexible recitation, contrasting registers, escalating activity, and a subsequent return (p.619). These are Porter’s analyses, not fresh listening findings here. His examples are principally studio transcriptions; his concert tape lacked a bass solo and ran approximately 38 minutes (p.601). Old tape positions therefore cannot serve as current-edition timestamps. Lewis Porter, “John Coltrane’s ‘A Love Supreme’: Jazz Improvisation as Composition,” JAMS 38/3, Autumn 1985, pp.593–621; article scan, stable publication record.

This supports a useful research question: what remains recognizable when pitch events, accompaniment, and duration change substantially? It does not license treating the studio transcription as the Antibes score, or treating spiritual meaning as a quantity calculable from Fourier magnitudes.

Representation requirements

GAMUT can compare the distinct pitch-class content selected from a passage, its full content spectrum, interval autocorrelation, and admissible distinct-note orderings. Such descriptors may identify stable material across locally different realizations. The current model does not directly encode the performance.

An event sequence such as the illustrative (0,3,0,5) immediately violates distinctness. Replacing it by the set {0,3,5} destroys recurrence; replacing it by a three-event ordering destroys one attack. This example is a representation test, not an asserted Antibes transcription. Preserve repeated-note motifs as event sequences before projecting anything into GAMUT.

A suitable research record should retain onset, duration, register, pitch uncertainty, instrument, articulation, and simultaneous-event membership, with optional continuous pitch tracks. Attach the current content descriptor to a declared window or phrase; do not silently make it the whole representation. Retain an open phrase boundary unless a repeated cycle is established. Every deduplication, octave reduction, and closing edge should be recorded as an explicit lossy operation.

Content autocorrelation does not encode a bass note’s changing tonal function, a delayed piano response, a drum accent, a saxophone’s register, or breath/noise. Fourier “order frequency” is frequency over event indices, not acoustic frequency, tempo, or a cymbal’s spectrum. Likewise, a movement’s pitch inventory cannot explain its dramatic trajectory merely because a geometric path can be drawn through inventories.

Edition-anchored listening protocol

Preparation. Obtain lawful access to the named digital Antibes block or the 2015 three-CD edition. Record edition identifier, track titles, actual durations, sample rate if available, and any gaps. Keep the source untouched. Begin every annotation with edition / movement / local timestamp; cumulative times above are navigation aids, not verified phrase boundaries. Log channel imbalance, distortion, and masking before comparing loudness or timbre across editions.

Pass 1: uninterrupted form. Hear the complete concert once without plots. Independently mark perceived section boundaries, returns, releases, and peak passages, including uncertainty intervals. Collect a brief verbal account of what organizes the performance. Reveal movement labels afterward when feasible. Do not define a climax as the loudest frame or require one predetermined arch.

Pass 2: motifs and transitions. In “Acknowledgement,” annotate recurring short cells, their pitch order, repeated attacks, rhythm, register, and accompanying bass material. Test whether a listener follows the same motif through changed pitch levels. Across the I–II boundary, examine how the transition is prepared rather than assuming the track cut is the musical event. In “Resolution,” annotate theme returns and departures; in “Pursuance,” identify actual changes of foreground instrument and subdivision before labeling choruses. In “Psalm,” describe phrase contours and rests without forcing the studio poem onto them.

Pass 3: ensemble interaction. Choose six to ten short passages after listening, spanning all movements and quiet as well as dense textures. For each, make separate timelines for saxophone, piano, bass, and drums. Mark candidate cues and responses, alternative explanations, and uncertain instrument attribution. A simultaneous accent could reflect shared meter rather than a response; a later event could reflect the common form rather than one player following another. Preserve those alternatives in the dataset.

Pass 4: timbre and trajectory. Annotate attack quality, sustained versus interrupted tone, register crossings, roughness, and density changes. Separately collect perceptual tension, energy, and sense-of-arrival ratings; their disagreement is informative. Acoustic features may help locate candidates but cannot establish expressive intention. Align the surviving film only to matching audio passages; camera edits are not reliable evidence of musical causation.

Validation. Have at least two annotators independently code a subset; adjudicate disagreements while preserving their original labels. Revisit every automatically detected pitch or onset used in a central claim. Compare studio, Antibes, and Seattle by named musical function or motif, not by a uniform time stretch. Keep uncertainty around passage matching and around edition boundaries.

Controllable instruments and falsifiable experiments

The following are proposed experiments, not claims that GAMUT already reproduces the quartet or its musical effects.

1. Motif-preserving variation instrument. Build a sequencer with independent controls for transposition, repeated-note retention, attack spacing, register, articulation, and event reordering. A performer can vary one while locking the others. Present short original phrases inspired by the analytical question, rather than distributing full Coltrane transcriptions. Compare a GAMUT content-only view, an order-aware view, and a plain piano-roll baseline. Measure recognizable identity, ability to reach a requested variation, and voluntary reuse after practice. A useful result is better control at comparable musical quality—not merely distinguishable mathematical coordinates.

2. Responsive ensemble instrument. Give a player separate controls over accompaniment density, response delay, register overlap, and rhythmic alignment. Compare live contingent accompaniment with a replay matched for note counts and overall density. Let performers signal perceived responsiveness, interruptions, and recoveries; ask independent listeners to judge the interaction. Randomizing response contingency helps separate genuine response benefits from simply adding more activity. Do not infer historical causality from the Antibes recording alone.

3. Rhythm–timbre dissociation. Keep pitch classes and event order fixed while changing onset placement and sound envelope; then hold timing/timbre fixed while changing order. Use counterbalanced paired trials and participant-created sounds. Ask which transformations preserve phrase identity and which alter propulsion or tension. If timing or timbre dominates the judgments, that defines the limits of a pitch-only control surface rather than a failure of the mathematics.

4. Long-form trajectory instrument. Give performers a few independently editable envelopes for register span, event density, motif-return probability, accompaniment activity, and silence. Test whether they can produce differentiated beginnings, developments, and endings without an imposed “intensity” macro changing everything together. Compare their requested trajectory with listener annotations and with their own recorded account. Pitch-class descriptors become one control layer, not a claimed generator of the whole musical arc.

For all four studies, recruit improvisers as collaborators in choosing the tasks and interpreting failure. Include novices only for questions that actually concern learnability. Counterbalance interfaces, allow equal practice, record familiarity with the piece, and distinguish immediate novelty from durable use. Preregister primary outcomes after a pilot; retain unsuccessful examples. Study expressive agency, predictability, recovery from unintended actions, and willingness to keep using the instrument alongside accuracy measures.

The first deliverable should be a small, edition-verified listening dossier with a handful of defensible event annotations and original playable demonstrations. Only after that should GAMUT’s visual or symplectic controls be credited with musical utility. The recording supplies demanding questions about identity, interaction, and change; the experiment must establish whether a particular instrument helps musicians act on them.