V4.2 Smart Device Setup...

ISLA v4.2.0 — Connected Listening

Smarter Device Setup · Vinyl-Style Slowdown · LIVE Reliability · Audio Control Center Improvements

Release Date: October 27, 2026
Author: Don Michael Feeney Jr.
License: MIT

Overview

ISLA v4.2.0 introduces a significant refinement of the listening experience, emphasizing simpler audio-device configuration, more reliable transport, and intuitive musical exploration.

This release improves how ALIVE connects to physical listening devices, strengthens LIVE audio-buffer handling, introduces a new Vinyl-style slowdown control, and transforms the Audio Control Center's visual feedback into a more informative representation of playback activity.

The underlying musical-processing engines and experimental Remastered research algorithms remain unchanged.

The objective is straightforward: make ISLA easier to operate, more dependable, and more enjoyable to explore.

What's New

1. Automatic ALIVE Listening-Device Setup

Opening and playing local music is now more straightforward.

• Open File → Play can automatically identify and select an unambiguous, compatible physical listening device.
• Existing compatible device selections are preserved.
• Recording sample rate and channel compatibility are checked before opening playback.
• When several eligible outputs are available, users are prompted to choose rather than risk an ambiguous connection.
• ALIVE and LIVE device menus synchronize with the confirmed selection.
• The interface displays Connected — ALIVE playing while the listening output is active.

ALIVE continues to operate independently of BlackHole and does not automatically activate LIVE or change macOS audio routing.

2. Vinyl-Style Slowdown — 50% Playback

A new sliding on/off control introduces traditional record-style slowing directly into ALIVE.

When activated:

• Playback slows to 50% of its normal speed.
• Musical pitch drops by approximately one octave.
• Original / Bypass, Velvet, Legato, and Melt remain selectable.
• Switching back to 100% restores normal playback speed and pitch.
• Playback-position continuity is preserved across speed changes.
• The Control Center's effect animation reflects the slower playback state.

This feature operates exclusively on the listening output. Original recordings, prepared audio buffers, terminal playback, LIVE processing, and exported mixes remain unchanged.

Additional safeguards address playback completion, final-frame handling, speed-switch timing, and pending mute adjustments.

3. Improved LIVE Audio Reliability

Version 4.2.0 introduces targeted improvements to the LIVE transport following native testing of BlackHole and Bluetooth audio configurations.

The updated application requests more robust native capture buffering and recommends supported listening-device sample rates.

A new single-producer/single-consumer ring-buffer implementation also addresses an application-level copy-lock collision that could terminate monitoring despite available buffer capacity.

The changes preserve:

• Existing audio-loss detection and recovery protections.
• Native device identity and routing verification.
• Timestamp and stale-audio safeguards.
• Fixed eight-block buffer capacity and genuine queue-full protection.
• Explicit failure reporting and safe shutdown behavior.

These improvements address observed transport problems without relaxing the existing protection mechanisms or modifying the DSP algorithms.

4. A More Understandable Audio Control Center

The Control Center now provides clearer guidance during device setup and audio operation.

Device setup improvements

• Find Connected Devices shortcut.
• Simplified three-step LIVE configuration.
• Clearer output names and device selection.
• More useful connection and recovery instructions.
• Automatic sample-rate suggestions without overriding deliberate user choices.

Playback visualization

• Improved effect/status display for Original, Velvet, Legato, and Melt.
• Clear distinction between selected effects and worker-confirmed processing.
• Visible switching and transport states.
• Animation that responds to confirmed playback and stops during pauses, recovery, or shutdown.

The effect display communicates application state rather than claiming to measure the audible waveform.

Engineering Verification

The PR reports the following local verification results:

1,024 passed · 1 skipped · 4 expected failures

The skipped test concerns optional screenshot capture, while the expected failures are associated with existing offline behavior.

Native Mac testing also recorded:

• A successful automatic ALIVE connection to Sony WH-CH720N headphones without changing macOS routing.
• A 150.28-second LIVE transport run involving 375 requested effect changes.
• Zero reported native directional loss flags, capture gaps, or ring-buffer failures during that run.
• Buffer high-water marks of 2 out of 8.
• Playback-speed signal verification at 44.1 and 48 kHz, demonstrating approximately 220 Hz output from a 440 Hz source at 50% speed.
• Regression coverage for cursor continuity, mode switching, mute ordering, completion handling, keyboard interaction, device setup, and transport recovery.

Tests also verify that original audio data and exported mixes are not altered by listening-only slowdown.

Scope and Limitations

ISLA remains an independent musical-processing research project.

The experimental Remastered engines retain their existing numerical safeguards, processing limits, and research status. The 30-second ALIVE Remastered admission limit remains in effect.

Native listening tests provide useful evidence of operational improvement but do not establish universal Bluetooth reliability, long-duration stability, certified real-time performance, or objectively superior sound quality.

More robust native buffering may introduce additional monitoring latency. Physical audio quality will continue to require listening-based evaluation.

Pupose

ISLA continues to develop around a simple idea: musical expression and engineering discipline can coexist.

Version 4.2.0 advances that goal by making the application easier to connect, clearer to understand, and more flexible to explore—while preserving the technical boundaries that support its research.

V4.1.2 Remastered Shortcut

ISLA v4.1.2 — Remastered-First Listening Shortcut

ISLA v4.1.2 is a small user-interface patch building on the published v4.1.1 release.

What's new

• Try Remastered First (up to 30s) now appears beside the ALIVE listening workflow after a file is prepared. For eligible short recordings, it opens the existing ALIVE Remastered workspace directly, without requiring the listener to select Original, Velvet, Legato or Melt first. Users may press Play first, but don't have to.
• Clearer eligibility text explains that ALIVE Remastered is an experimental short-excerpt research workflow, not a full-song processing mode.
• Existing research entry, background preparation, validation, cancellation, and 5 ms playback handoff remain unchanged.
• Version identifiers across Python, HTTP API, CMake/JUCE, citation and regression tests are aligned at 4.1.2 for this source candidate. The latest published release remains v4.1.1 until a separate release is deliberately published.

Important limitations

• The research renderer still accepts only nonempty mono/stereo, 44.1 or 48 kHz, at most 30-second prepared recordings. The shortcut correctly stays disabled for longer songs, invalid sample rates, unprepared files and incompatible LIVE state.
• Full-song source selection with an independently prepared 30-second audition segment is not implemented here. See Issue #85 for the proposed follow-up design.
• Native macOS LIVE capture interruptions remain under investigation; this GUI patch makes no driver, CPU-overload, audio-safety or scientific-benefit claims.
• Frozen Experiment 02 algorithms, research evidence, playback DSP, safety limits and published v4.1.0/v4.1.1 tags are untouched.

Verification before release

• Real Qt regression: short recording can go from ALIVE Play directly to Remastered entry; previous mode is restored after cancel; no production-mode selection is required.
• Ineligible full-length recording and invalid-rate recording remain blocked, with standard ALIVE controls unaffected.
• Full macOS/Linux ALIVE and Qt suites and strict Linux build must pass on the final PR commit; conduct an ordinary Mac listening sanity check before publishing.

Let the Music Flow. With More Choice. 🎹💙

V4.1.1 Telemetry & Stability

ISLA v4.1.1 — LIVE Efficiency, Telemetry & Interface Stability

Let the Music Flow. With Greater Clarity.

ISLA v4.1.1 is a focused engineering patch that builds on the Audio Control Center and playback improvements introduced in v4.1.0.

This update improves the lifecycle of LIVE Remastered processing, expands real-time diagnostic visibility, refines the operator interface, and strengthens regression testing.

Release Transparency: ISLA v4.1.1 improves processing efficiency and diagnostic observability. However, intermittent LIVE audio interruptions remain under investigation, particularly with native macOS audio capture and certain physical playback-device configurations. The changes in this patch should not be interpreted as a verified solution to all audio interruptions. Remastered processing remains experimental, and automated testing does not guarantee uninterrupted playback on every system.

---

What's Improved

1. More Efficient LIVE Remastered Processing

ISLA now retires the inactive Remastered processing bank after transitioning back to Production.

• Stops unnecessary Remastered processing after its output is no longer selected.
• Releases inactive Remastered processing history.
• Creates fresh processing history when Remastered is selected again.
• Preserves the existing warm-up behavior and 5-millisecond crossfade.
• Maintains Production processing history to support clean transitions.

This optimization reduces unnecessary work when Remastered is inactive without changing the underlying musical processing algorithms.

Important: Both processing paths may still perform work while Remastered is actively selected. Further optimization will be guided by measured performance.

2. New LIVE Processing Telemetry

A dedicated LIVE Processing — Engine & Budget category has been introduced within the Audio Control Center.

Developers and operators can now inspect:

• Requested versus worker-applied processing engine.
• Remastered processing-bank status.
• Nominal audio-block processing interval.
• Last and peak worker processing times.
• Separate Production and Remastered processing measurements.
• Total processed blocks and Remastered blocks.
• Processing-budget exceedances.
• Inactive Remastered-bank retirements.

These measurements provide additional evidence for diagnosing processing behavior and investigating potential performance bottlenecks.

Worker wall-clock measurements are approximate and must not be confused with actual CPU utilization or proof of native audio-driver failure.

3. Clearer LIVE Start, Stop & Failure Indicators

The LIVE monitoring interface has received targeted usability improvements.

• Light-green Start LIVE button when ready.
• Darker-green indication when LIVE monitoring is confirmed active.
• Light-red Stop LIVE button for clearer operator recognition.
• Dedicated monitoring-state indicator.
• More visible explanations when audio stops unexpectedly.
• Improved vertical spacing and control accessibility.

The interface reflects authoritative application state rather than assuming that clicking a button guarantees audio playback.

Safety preservation: ISLA does not automatically restart monitoring after a fatal audio failure. The operator must address the interruption and explicitly restart the session.

4. More Precise Audio Buffer Diagnostics

LIVE diagnostics now provide additional information about application audio-buffer failures.

Previously, certain queue failures appeared under a combined overflow/contention message.

The updated diagnostics distinguish whether a failed buffer write resulted from:

• A full application audio queue.
• Nonblocking lock contention.

Existing buffer high-water measurements are also surfaced.

This improved observability should help identify the origin of future interruptions without altering buffer capacities or weakening the established safety shutdown mechanisms.

5. Stronger Automated Regression Testing

Additional acceptance coverage strengthens confidence in:

• Remastered processing-bank retirement and reinitialization.
• Engine switching at 44.1 kHz and 48 kHz.
• Production audio equivalence following Remastered retirement.
• Crossfade continuity and processing history.
• Processing telemetry and workload counters.
• LIVE input-buffer failure classification.
• Accurate interface state indicators.
• Existing playback and audio-control interactions.

Earlier test-harness scheduling improvements also reduce nondeterministic failures involving simulated native audio callbacks.

The associated pull-request validation completed successfully across the required macOS and Linux test environments, including the strict Linux build.

---

Engineering Findings & Known Limitations

Physical listening tests conducted during development identified two distinct LIVE failure scenarios.

Application-buffer interruption: One session stopped following an application input-buffer overflow/contention event without a corresponding native device-overflow report.

Native capture interruption: Another session reported repeated native input-overflow events. ISLA completed an initial recovery before stopping safely when another capture discontinuity occurred within the recovery quiet interval.

In the latter session, Production processing remained below the nominal audio-block processing budget in the recorded measurements, and Remastered processing was inactive.

These observations do not establish Remastered computational overload as the cause of those interruptions.

Traditional ALIVE file playback demonstrated successful operation independently of LIVE system-audio monitoring. The experimental ALIVE Remastered preparation limit remains 30 seconds; longer recordings continue to use standard ALIVE playback or require a shorter research excerpt.

Future investigations will prioritize native audio-device synchronization, capture-buffer behavior, callback scheduling, and longer physical-device listening sessions.

---

Preservation & Compatibility

This patch preserves:

• The established Velvet, Legato, and Melt musical processing profiles.
• Original/Bypass behavior.
• Existing research algorithms and frozen research evidence.
• Audio-buffer safety limits and recovery policies.
• Manual Start/Stop control and audio-device ownership.
• Existing ALIVE playback and Remastered admission requirements.
• The MIT open-source licensing model.

No experimental research results are reclassified or represented as scientifically validated by this update.

---

Development References

• [PR #82 — LIVE regression-test reliability](#82)
• [PR #83 — Remastered processing lifecycle and telemetry](#83)
• [PR #84 — LIVE diagnostics and monitoring-state indicators](#84)

ISLA continues to evolve through reproducible research, careful engineering, transparent limitations, and community participation.

Let the Music Flow. 🎹💙

ISLA — Open-Source Audio Processing & Research | MIT License

V4.1 UI & Stability

ISLA v4.1.0 — UI & Stability

Let the Music Flow. With Greater Control.

ISLA v4.1.0 is a focused usability and stability release that builds on the experimental Remastered Audio capabilities introduced in v4.0.0.

Release Transparency: ISLA v4.1.0 introduces verified interface improvements and targeted playback-stability corrections. However, Remastered processing remains experimental, and intermittent LIVE audio interruptions are still under investigation. Automated testing validates specific behaviors but does not guarantee uninterrupted playback across all hardware configurations. These limitations are acknowledged as part of ISLA's ongoing development.

This update improves the desktop Audio Control Center, strengthens ALIVE playback interactions, hardens LIVE stream startup, and expands automated tests around audio continuity and application state.

The goal is simple: make ISLA easier to operate, more predictable during playback, and better equipped to identify failures—without changing its underlying musical processing algorithms.

---

✨ What's Improved

1. Refined Audio Control Center

The native Qt interface has been reorganized into a clearer, more responsive workspace.

• Improved navigation across ALIVE, Remastered, and LIVE controls.
• Added a compact, vertically scrollable layout for narrower windows.
• Improved button visibility and reachability without replacing existing controls or changing their audio behavior.
• Introduced dedicated ALIVE Telemetry and LIVE Telemetry panels for clearer diagnostic inspection.
• Added blue interface accents and improved visual organization.

The updated layout is designed to remain usable across different window sizes while preserving the established application workflow.

2. Animated Processing-Mode Indicators

A new animated footer signature visually represents ISLA's three musical processing profiles:

VELVET · LEGATO · MELT

The indicators respond to the processing mode confirmed by the ALIVE audio callback or LIVE worker, rather than merely highlighting a newly requested selection.

These animations represent application state; they are not measurements of the actual audio waveform.

3. More Predictable ALIVE Playback

Several audio-control timing issues were identified through targeted regression testing.

• Volume and playback synchronization: Pending ALIVE volume changes are applied before Play or processing-mode commands.
• Rapid mode-switch protection: The Control Center defers subsequent traditional mode selections until the current 5 ms crossfade has finished.
• Latest-selection behavior: Multiple rapid requests are consolidated so the newest pending mode takes precedence.
• State cleanup: Pending selections are discarded when playback is paused, stopped, reloaded, or the active workspace changes.

These changes reduce avoidable control-state inconsistencies and transition discontinuities without modifying the frozen isla/playback.py implementation.

4. Safer LIVE Audio Startup

LIVE monitoring now starts its DSP worker before activating the native duplex audio stream.

This closes a startup ordering gap in which the audio device could begin delivering callbacks before the processing worker was ready.

Additional safeguards protect against premature watchdog failures, incorrect Running-state reporting, and unsafe resource cleanup during failed startup.

Existing capture-recovery limits, device validation, audio-buffer capacities, and safety shutdown behavior remain intact.

---

🧪 Expanded Regression Testing

ISLA v4.1.0 strengthens automated verification across playback, interface interaction, and audio-stream lifecycle behavior.

New tests examine:

• Nonconstant audio sample continuity at 44.1 kHz and 48 kHz.
• Mono and stereo playback with varied callback lengths.
• Sample ordering, crossfade accuracy, and end-of-file handling.
• Rapid processing-mode changes and cancellation of pending selections.
• Volume changes immediately preceding playback commands.
• Stream ownership and resource cleanup.
• LIVE startup activation and early device-failure recovery paths.
• Real Qt mouse interactions and control reachability at multiple window sizes.

Latest main-branch acceptance: 9 of 9 GitHub Actions checks passed, including macOS and Linux ALIVE/Qt testing on Python 3.10 and 3.13, plus the strict Linux build.

Passing automated tests establishes the tested behavior, not universal hardware compatibility or perceptual audio quality.

---

🔬 Scientific and Architectural Preservation

This release retains ISLA's established processing foundation.

• Original/Bypass, Velvet, Legato, and Melt remain available.
• The experimental Remastered capabilities introduced in v4.0.0 are preserved.
• No changes were made to the frozen ALIVE playback engine or core DSP processing equations.
• Existing audio-buffer preparation, source preservation, and governed research boundaries remain in place.
• The command-line interface and offline audio workflows remain available.

v4.1.0 focuses on control, continuity, and operational reliability—not new harmonic-processing claims.

---

⚠️ Known Limitations

LIVE monitoring remains sensitive to native audio-device behavior, particularly when using BlackHole capture with a separate physical or Bluetooth output.

Isolated capture interruptions may recover successfully; repeated interruptions within the existing safety interval can still stop monitoring to prevent stale or discontinuous audio from being treated as valid.

Additional playback-continuity tests also document unresolved lower-level edge cases involving direct mid-crossfade buffer selection, premature stream-finished signaling, and playback-position recovery after unexpected device termination. These remain explicitly visible as expected test failures rather than being represented as resolved.

Native macOS listening and device-specific reliability validation remain important.

---

📦 Included Pull Requests

Interface and presentation: [#71](#71), [#72](#72), [#73](#73), [#74](#74)

LIVE startup stability: [#75](#75)

Responsive interface correction: [#76](#76)

ALIVE playback coordination: [#77](#77), [#79](#79)

Playback-continuity regression coverage: [#78](#78)

All nine pull requests have been merged into main.

Validated source commit: f7e80f4bf5bc6dc29bc51f0db557df377d627293

---

ISLA — Let the Music Flow.

Version 4.1.0 advances ISLA's usability and operational stability while preserving the mathematical foundation of its musical experimentation.

Open source · MIT License · Python · Qt · Audio DSP

[ISLA on GitHub](https://github.com/dfeen87/ISLA)

V4 Remastered Audio

ISLA v4.0.0 — Remastered Audio & Control Center

ISLA v4.0.0 introduces experimental Remastered audio processing, expanded playback and export capabilities, and a refined desktop Audio Control Center.

This release builds on v3.3.0 while preserving the established Velvet, Legato, Melt, and Original/Bypass processing modes.

What's New

v4.0 Remastered Audio

• Introduced experimental Remastered Velvet, Legato, and Melt processing for ALIVE local-file playback and continuous LIVE monitoring.
• Added adjustable processing depth from 0–100% in 5% increments, independent of listening volume.
• Added source validation, numerical safety checks, preparation progress, cancellation, and restoration of the previous ALIVE mode.
• Preserved the existing production DSP modes and historical research implementations.

Audio Control Center

• Organized the native Qt interface into three workspaces: ALIVE | v4.0 Remastered | LIVE.
• Improved mode selection, device configuration, playback controls, and diagnostics.
• Added independent ALIVE and LIVE listening-volume controls with mute functionality.
• Improved control-state synchronization, error reporting, and recovery behavior.
• Added full-screen launch support for the desktop interface.

Playback, Import, and Export

• Expanded ALIVE local-file support, including DRM-free AAC/M4A and Apple Lossless imports on macOS.
• Added Save Mix to export prepared ALIVE audio as lossless 32-bit floating-point WAV files.
• Added background export, cancellation, overwrite protection, and cleanup handling.
• Improved local audio-file discovery and Bluetooth output-device selection.

Fixes and Improvements

• Corrected several Qt control-state and mode-selection issues.
• Improved LIVE volume attenuation, device refresh, and monitoring recovery handling.
• Improved Remastered preparation, mode restoration, and failure recovery.
• Fixed stale interface warnings and rejected-setting synchronization.
• Excluded a specific PySide6 dependency version associated with a reproducible Qt runtime crash.

Testing and Verification

The final implementation passed its documented pull-request acceptance checks, including:

• Nine hosted CI jobs covering Python acceptance, Qt interface tests, and the Linux C++/VST3 build.
• Remastered intensity and LIVE processing-equivalence regression matrices.
• Audio output-gain and Save Mix regression tests.
• Additional GUI, playback, recovery, and state-management tests.

These automated results do not constitute comprehensive physical-device or perceptual listening validation.

Known Limitations

• Remastered processing remains experimental. Its scientific evaluation is REVISE / INCONCLUSIVE; improved perceived audio quality has not been established.
• Remastered ALIVE preparation is limited to supported WAV excerpts of up to 30 seconds at 44.1 or 48 kHz.
• LIVE monitoring requires compatible macOS audio routing and a supported output device. Bluetooth latency and device-dependent interruptions remain possible.
• Some reported native playback artifacts have not been conclusively resolved.
• A standalone macOS application installer is not included.

Getting Started

Requires Python 3.10 or newer. LIVE monitoring targets macOS 13 or newer.

git clone https://github.com/dfeen87/ISLA.git
cd ISLA
python3 -m venv .venv
source .venv/bin/activate
python -m pip install -e '.[gui]'
python -m isla window

For installation details, audio-routing instructions, listening examples, and research documentation, see the README.

Project Information

Author: Don Michael Feeney Jr.

License: MIT

Repository: https://github.com/dfeen87/ISLA

Previous release: v3.3.0

Release tag: v4.0.0

---

ISLA remains an open-source audio-processing and research project combining practical music playback with experimental digital signal processing.

V3.3 Control Usability

ISLA v3.3.0 — LIVE Control & Reliability

Release preparation · 24 October 2026 · eight merged post-v3.2.0 pull requests

Scope: User-facing LIVE control usability, reliable Qt mode selection, terminal guidance, safe audio-device selection, physical listening-output attenuation, and improved macOS Audio MIDI documentation. This release does not introduce a new harmonic algorithm or tune frozen scientific experiments.

Immutable baseline: v3.2.0 — Audio Control Center, commit fee185c5f1851c6b7fce547068d86b756ea12332, published October 24, 2026. The v3.3.0 release candidate includes the following subsequent merged PRs: #38, #39, #40, #41, #42, #43, #44, and #45. There are no unmerged PR changes intentionally included.

Publication status: This is the source release note, not proof of a published GitHub release. The v3.3.0 tag/release must be created separately from the final approved and green main commit. Do not publish this note as a claim of native physical-Mac certification or unattended playback reliability.

Highlights

1. Reliable LIVE effect selection and confirmations — #39, #41, #43, #45

• LIVE starts in Original / Bypass in the graphical Control Center; effect changes select Velvet, Legato, or Melt without restarting the running audio stream.
• Step 4 now uses a single accessible effect dropdown and a separate, explicit Reset to Original action. Re-selecting the same effect is idempotent, rather than toggling unpredictably.
• Corrected a user-reported bug where periodic state refreshes temporarily disabled the dropdown, dismissed its popup, or reverted a newly selected mode to Bypass. The GUI now preserves pending intent, reconciles asynchronous success/failure, rejects stale snapshots and distinguishes requested, confirmed, and DSP-applied mode.
• Regression coverage now uses real Qt popup mouse/keyboard interactions in addition to model/worker state checks; operator verification on real Mac hardware remains separate.

2. Dedicated LIVE Listening Volume — #41

• New 0–100% GUI listening-volume control with Apply Volume and Mute/Unmute, restoring the previously selected nonzero setting when unmuted.
• Applies attenuation only to ISLA's post-DSP physical output, not BlackHole capture, the system's master output, local ALIVE files or other apps. 100% is unity (no additional gain), and 0% mutes.
• Uses a short smoothing ramp (approximately 5 ms) in the existing worker path; changing the level does not restart LIVE. It does not provide universal loudness normalization or establish musical superiority of any effect.

3. Responsive Stop, safer device polling and diagnostics — #39, #40

• Avoids routine full CoreAudio/PortAudio enumeration while streaming, preventing UI polling from competing with Stop and mode commands. Stop releases audio before any subsequent optional recheck.
• Surfaces the actual LIVE fail-closed safety-stop reason near Step 4 while preserving detailed metrics and error counters.
• Enlarged, scrollable diagnostics retain the reader's scroll position during health updates, while selected/active effects and processing acknowledgment are visible.

4. Numbered, verified capture/output pickers — #43

• In isla>, after source system, typing bare input or output now opens a filtered numeric picker. Type only the displayed number at its selection prompt; cancel exits without mutation.
• The LIVE input picker allows the standalone BlackHole 2ch capture device; the output picker allows a separately verified physical stereo device, never the BlackHole feedback sink or an input-only microphone.
• PortAudio index and CoreAudio-native identity are rechecked before applying a picker choice, guarding against hot-plugged/stale selections. Prior setup audio flow, explicit numeric selectors and exact-name commands remain available.
• Device selection and rescans remain unavailable while a LIVE stream is active. ISLA does not change the Mac's default output automatically.

5. Contextual, professional terminal command center — #45

• A width-aware, ASCII-friendly DOS-inspired help screen shows the current source, running/stopped state, mode, devices, relevant commands and actionable next step.
• Topic help includes help live, help file, help devices, help modes, and help troubleshoot, plus ?.
• Mistaken play during LIVE, repeated monitor start, and attempted output changes during active monitoring now provide contextual, safe instructions. These command responses do not reopen or hijack a live stream.
• The existing diagnostic status output, file player commands, and shutdown behavior remain available.

6. Safer and clearer installation/onboarding — #38, #42, #44

• Repaired and embedded the real ISLA Control Center screenshot in the README.
• Added prominent warnings to lower source and physical listening volume before using any 100% output command. AppleScript output-volume commands affect whichever macOS output supports them; they are not dedicated BlackHole gain controls.
• Added two authentic Audio MIDI Setup photographs demonstrating BlackHole 2ch at 1.0 / 0.0 dB unity and a deliberately low MacBook speaker volume with centered balance. These are device-specific examples, not a promise that the speaker screenshot equals 25% macOS output.
• Documented independent macOS alert-volume adjustment (e.g. osascript -e 'set volume alert volume 25') and Sound Effects routing.

Compatibility and commands

• Python: 3.10+; distribution isla-audio, import isla, with compatible legacy phased_music_notes imports.
• Desktop UI: optional PySide6-Essentials, installed using python -m pip install -e '.[gui]'.
• Launch: isla or python -m isla for terminal; isla window or python -m isla window for Control Center; window from an active terminal session.
• LIVE: experimental macOS 13+; standalone BlackHole 2ch capture, separate supported physical stereo output and manually configured macOS system routing, typically 44.1 kHz or 48 kHz.
• ALIVE: local WAV playback and original offline export remain separate from LIVE.
• Public package, HTTP app and JUCE/CMake version metadata advance together to 3.3.0. No VST/AudioUnit certification, standalone Mac installer or production professional real-time guarantees are introduced.

Example LIVE workflow (device numbers depend on the current machine):

isla> source system
isla> input
Select input: 0
isla> output
Select output: 1
isla> monitor start
isla> mode velvet
isla> bypass
isla> monitor stop

Within the GUI, Start LIVE and Stop LIVE remain explicit operations; changing an effect or ISLA listening volume does not replace these controls.

Test evidence and remaining acceptance

The merged #45 implementation summary reports 559 passed with 1 existing expected failure for its full local Linux Python 3.12 / optional-Qt suite, and 90 focused terminal/Qt checks. These are reported results on that PR, not automatically evidence that a later release-version commit is green.

Release gates to verify against the exact final v3.3.0 candidate:

1. Full ALIVE Python acceptance across the repository's supported Ubuntu/macOS CI matrix.
2. Control Center Qt offscreen GUI acceptance across the supported matrix, including the actual popup and stale-state regression tests.
3. Linux/JUCE strict-build CI, version/identity tests, and all research preservation checks.
4. Explicit local/native macOS operator acceptance at supported rates: BlackHole-to-physical-output routing, no feedback, distinct Bypass/Velvet/Legato/Melt selection, pending-state persistence, Reset to Original, listening-volume and mute behavior, Stop/Close, and safe return of the macOS output setting.
5. Verify package/API/CMake/research identity at the approved exact release commit, and preserve all historical v3.0.0, v3.1.0, v3.1.1 and v3.2.0 tags unchanged.

Limitations: Hosted offscreen GUI tests and fake-audio regressions are not native physical macOS hardware, headphone listening or sustained hot-plug validation. LIVE may incur noticeable latency, underflows or safe-stop interruptions and remains experimental. Stronger native proof, broad device compatibility, sustained-session performance, independent sound-quality assessment and distinct remastered DSP are not claimed.

Research and DSP preservation

• No new Velvet/Legato/Melt mathematical tuning, normalization or spectral-quality claim is made by this release.
• The existing streaming audio engine, bounded capture-recovery behavior, fail-closed routing rules, legacy ALIVE workflow, and fixed Experiment 02 research corpus remain under their existing provenance and validation limitations.
• This version is a control-path and usability release, not a declaration of v4.0 Remastered or an independently validated acoustic improvement.

V3.2.0 Window!

ISLA v3.2.0 — Audio Control Center

Release scope and provenance · 24 October 2026

ISLA v3.2.0 introduces an optional, clickable desktop Audio Control Center for local-file ALIVE playback and guided macOS LIVE system-audio monitoring, while retaining the terminal and offline export workflows. The version also includes the accumulated research-preservation and CI work merged since v3.1.1.

This is the source release note. A GitHub release and tag are separate publication actions; their existence must be verified independently. The package version alone is not proof that a GitHub release has been published.

Desktop control window — merged PR #34

• Launch an optional Qt Widgets interface with isla window, python -m isla window, or window from the existing isla> prompt.
• Preserve bare isla, python -m isla and the existing --input/--output/--mode offline CLI.
• ALIVE: WAV picker, preparation progress, playback controls, Velvet/Legato/Melt and original Bypass, playback position, and guarded selections.
• LIVE: BlackHole 2ch capture, verified physical stereo output selection, 44.1/48 kHz rates, routing-readiness checks, Sound Settings guidance, start/stop, mode selection and health feedback.
• Share the existing terminal controller and audio-engine ownership; no second DSP path and no unsafe Mac default-device writes.
• Qt dependency remains optional: python -m pip install -e '.[gui]'. CLI-only installation continues to work without Qt.
• See Control Center acceptance and mapping and the README GUI walkthrough.

Experiment 02 — merged research PR #33

• Preserves the Experiment 02 phase-governed resonance corpus, methods, reports, freeze, checksums and reproducibility index.
• The three distinct research adapters (Velvet, Legato, Melt) remain under tools/resonance02; they are not wired into the production ALIVE or LIVE modes in v3.2.0. This is not ISLA v4.0 Remastered.
• A separately identified post-freeze slowdown successor repairs isolated-impulse alignment fixtures without modifying frozen results. Dense or overlapping onsets, perceptual sound quality on real acoustic-piano recordings, and native audio hardware acceptance remain unverified.
• The frozen Experiment 02 verdict is unchanged: the original controlled tests do not establish a governance benefit or preferred sound; creator listening remains pending.
• Existing research evidence packages and large audio/data artifacts remain outside ordinary Git; see Experiment 02 research documentation.

Other changes since v3.1.1

• README release, platform and CI badge improvements from PR #31.
• Explicit production-only and complete test-suite steps, preserving visibility into research tests.
• Package, API and JUCE/CMake identity metadata aligned to 3.2.0, with compatibility assertions updated.

Installation and launch

From a Python 3.10+ checkout:

python -m pip install -e '.[gui]'
python -m isla window
# Or from an active environment:
isla window

ALIVE can process local WAV without BlackHole. LIVE requires macOS 13+, installed standalone BlackHole 2ch, a distinct physical stereo output and manual macOS Sound > Output selection of BlackHole 2ch. ISLA does not modify the macOS default output. Restore it manually after LIVE.

Validation evidence and limitations

PR #34 reports a complete local GUI-enabled test run (477 passed, 1 existing expected failure) on a Linux host, plus offscreen Qt interaction and CLI compatibility tests. A clean no-Qt installation was separately exercised. These are reported implementation observations, not proof of physical native Mac success. CI evidence for the exact proposed release commit must be verified before publication.

Physical macOS GUI/audio acceptance remains pending, including real BlackHole routing, device hot-plug, live listening, recovery over longer sessions and Bluetooth hardware. Hosted offscreen macOS tests are not equivalent to audio-device or user-interaction acceptance. LIVE remains experimental; no professional real-time guarantee, Windows LIVE implementation, standalone macOS installer, or production v4.0 remastered DSP is claimed.

Historical v3.1.1, v3.1.0 and v3.0.0 tags and published release artifacts must remain unchanged.

Release review gate: Verify main-branch CI and version consistency, test installed GUI on a native Mac where possible, record any pending limitations in the public release description, then create a v3.2.0 tag and GitHub release from the approved exact commit. Do not describe GUI testing on mock devices as physical macOS acceptance.

V3.1.1 Stability

🎹 ISLA v3.1.1 — LIVE Continuity & Recovery

Let the Music Flow. Even When the Audio Stream Stumbles.

ISLA v3.1.1 is a focused reliability patch for the experimental LIVE audio-monitoring engine.

Following v3.1.0's DSP stability improvements, this release addresses an issue discovered during real-world macOS listening: a single BlackHole capture-overflow event could cause ISLA to terminate monitoring, interrupting playback and requiring a manual restart.

Version 3.1.1 introduces bounded recovery from isolated capture interruptions, stronger audio-state protection, and expanded diagnostics—without changing the musical character of ISLA.

---

🔧 What's Fixed

1. Recovery From Isolated Capture Interruptions

Previously, a reported input overflow or underflow after startup caused LIVE to shut down immediately.

ISLA now attempts controlled recovery when a capture discontinuity is isolated and the audio system remains in a valid state.

The recovery sequence:

• Rejects the affected audio block and outputs silence during recovery.
• Prevents queued or in-flight audio from the previous stream state from being replayed.
• Clears application audio queues.
• Rebuilds the DSP's history, overlap-add state, and delay buffers.
• Revalidates device identity and routing.
• Resumes processing from fresh incoming audio when recovery succeeds.

Lost audio is not reconstructed or concealed as though no interruption occurred. A brief audible gap may remain.

2. Bounded, Safety-Conscious Recovery

To prevent repeated failures from becoming an unstable monitoring loop, recovery follows strict limits:

• A minimum 10-second healthy interval between accepted capture interruptions.
• No more than three recovery attempts per explicitly started monitoring session.
• Recovery must complete within eight additional audio callbacks.
• Persistent or repeated discontinuities trigger a safe shutdown.

At 48 kHz, the eight-callback recovery deadline is approximately 85 milliseconds.

Unsafe routing, disconnection, invalid timestamps, buffer failures, and persistent starvation remain fatal conditions.

The existing BlackHole single-path safeguards and physical-output restrictions are unchanged.

3. Improved Audio Diagnostics

The status command now provides more detailed visibility into LIVE monitoring.

New diagnostic information distinguishes:

• Driver-reported input and output overflows/underflows.
• Application ring-buffer failures.
• Capture-discontinuity events.
• Recovery attempts, completions, and failures.
• Deliberately silenced and discarded audio blocks.
• Fatal device errors.
• The last interruption and recovery state.
• Maximum observed worker, routing, and queue timing.

A successfully recovered session is explicitly identified as Recovered (audio lost), rather than being misrepresented as uninterrupted playback.

4. Native Reliability Testing Support

A new scalar-only reliability probe, tools/live_reliability_probe.py, supports longer monitoring observations.

It records diagnostic counters, timing, operating environment, and recovery state without recording or saving the user's music.

A companion native macOS checklist covers Bypass, Velvet, Legato, Melt, physical device changes, extended playback, and recovery behavior.

These tools provide a more systematic way to evaluate the reliability of ISLA on actual Mac hardware.

---

🧪 Testing and Engineering Evidence

The implementation includes additional deterministic tests covering isolated and repeated capture loss, recovery deadlines, stale-audio prevention, DSP state reset, buffer contention, device faults, mode switching, and normal uninterrupted processing.

Recorded validation includes:

• 305 passing Python tests
• 1 unchanged expected failure
• 36 passing focused recovery tests
• Successful four-platform hosted ALIVE Python acceptance on the merged implementation
• Worker-only performance benchmarking at 44.1 and 48 kHz

The normal LIVE processing window, hop size, algorithmic delay, musical profiles, and DSP equations remain unchanged.

Bypass continues to preserve its original exact waveform behavior when capture is uninterrupted.

---

⚠️ Known Limitations

ISLA LIVE remains experimental.

Native Core Audio and PortAudio scheduling, device-driver buffering, physical latency, and actual hardware reliability can differ from simulated testing.

This patch does not guarantee uninterrupted playback, recover missing samples, eliminate audible gaps, or make Python audio processing hard real-time.

Sustained native Mac listening acceptance is still required to establish whether the original intermittent shutdown is reliably mitigated on the creator's hardware.

---

❤️ Continuing the ISLA Journey

The foundation of ISLA is not only musical experimentation. It is dependable engineering.

Version 3.1.0 strengthened the mathematical behavior of ISLA's audio processing.

Version 3.1.1 strengthens how LIVE responds when the surrounding audio system encounters a temporary interruption.

That reliability work supports the next chapter: exploring more distinctive musical identities for Velvet, Legato, and Melt.

Dedicated to Isla, whose name continues to inspire the project.

---

🎹 ISLA v3.1.1 — LIVE Continuity & Recovery

More resilient monitoring. Clearer diagnostics. The same musical vision.

ISLA — Let the Music Flow.

V3.1 DSP Stability

🎹 ISLA v3.1.0 — DSP Stability & Hear the Difference

Let the Music Flow. With Greater Precision.

ISLA is ALIVE. ISLA is LIVE. Now, ISLA is getting stronger.

Following the release of ISLA v3.0.0 — ALIVE + LIVE, version 3.1.0 introduces an important engineering milestone: improved numerical robustness, reproducible musical listening comparisons, and a more intuitive macOS audio setup experience.

This release focuses on making ISLA's existing audio-processing foundation more reliable while preserving its creative identity.

The mission remains unchanged: explore how mathematics can transform musical continuity, harmonic texture, and expression.

---

✨ What's New in v3.1.0

🔬 1. More Reliable Real-Time Audio Processing

Two reproduced numerical issues have been addressed in ISLA's experimental LIVE engine.

Improved startup reconstruction

The weighted overlap-add (WOLA) processor now correctly accounts for the normalization weights of virtual silent frames preceding the first incoming audio block.

Previously, incomplete startup normalization could create significant unintended amplitude amplification.

In controlled Melt-mode testing with a half-scale constant input:

Sample rate | Before | After -- | -- | -- 44.1 kHz | 1.238757 | 0.458413 48 kHz | 1.135527 | 0.458413

This correction removes the reproduced startup amplification in the tested conditions without changing ISLA's established processing delay, harmonic strengths, or mode architecture.

Improved spectral reconstruction

LIVE now applies a continuous, precision-aware spectral correction when FFT coefficients approach numerical nulls.

The updated reconstruction reduces sensitivity to small representation-scale perturbations without introducing arbitrary phase-dependent energy at exact spectral nulls.

Controlled numerical sensitivity decreased from approximately 10–12% to 0.063–0.087% relative RMS difference in the selected tests.

These measurements represent improved numerical stability, not an independently established improvement in perceived sound quality.

🎧 2. Hear the Difference — ISLA's First Listening Showcase

For the first time, ISLA provides a reproducible listening showcase directly from its GitHub repository.

Eight FLAC audio samples are included:

• Four original LIVE baseline recordings: Bypass, Velvet, Legato, and Melt.

• Four corresponding recordings using the numerical robustness corrections.

The showcase uses Emotional Piano by triangelx, distributed under the CC0 public-domain dedication.

The comparisons use the same six-second musical excerpt, consistent presentation edits, documented loudness matching, and lossless FLAC storage of 24-bit PCM audio.

Processing provenance, source attribution, gain adjustments, rendering details, and SHA-256 checksums are preserved.

Explore the listening comparisons:

Hear the Difference — Baseline Samples

Before & After — Numerical Robustness

Listeners can now explore ISLA's harmonic transformations without configuring a live system-audio capture session.

🎛️ 3. Original Musical Profiles Preserved

ISLA continues to offer three LIVE harmonic-processing profiles:

Velvet (0.35) — A gentle, softened harmonic texture.

Legato (0.55) — An exploration of greater musical connectedness.

Melt (0.75) — A more pronounced, flowing harmonic transformation.

Bypass — The original signal, preserved for comparison.

The controlled DSP investigation confirmed that the current LIVE profiles represent different strengths of one underlying harmonic transformation.

Version 3.1.0 intentionally preserves those settings.

Independent musical-mode redesign and additional artistic tuning remain subjects of future research.

🖥️ 4. Improved macOS and BlackHole Setup

The guided setup audio workflow now explains the distinction between selecting an output inside ISLA and configuring macOS's system-wide audio destination.

The required LIVE route is:

Mac system audio → BlackHole 2ch → ISLA LIVE → Physical speakers/headphones

Setup now more clearly instructs users to:

• Select BlackHole 2ch as ISLA's capture input.

• Select physical speakers or headphones as ISLA's playback destination.

• Manually choose BlackHole 2ch under macOS System Settings → Sound → Output.

• Start LIVE monitoring only after the system-audio route is configured.

If the system output is incorrect, the routing error now identifies the selected macOS output when available and explains the required correction.

ISLA continues to reject unsupported, ambiguous, or potentially unsafe audio routes. No automatic macOS routing changes have been introduced.

🧪 5. Expanded Regression and Research Evidence

This release incorporates a comprehensive controlled DSP investigation and additional regression tests.

Research and testing include:

• Reproduction of numerical defects before applying corrections.

• Independent overlap-add reconstruction verification.

• Spectral-null and phase-stability experiments.

• 44.1 and 48 kHz testing in mono and stereo.

• Exact delayed Bypass checks.

• Warm processing-bank and bounded-state validation.

• Same-source waveform comparisons and numerical analysis.

• Synthetic CPU-load and worker-budget measurements.

• Audio provenance, source checksums, and reproducibility instructions.

• Cross-platform Python acceptance and package-build verification.

The completed DSP study examined 224 matched processing cases and 25,200 measured worker hops in its performance investigation.

No modeled worker-budget overruns were observed in that benchmark. Native audio callbacks, actual hardware latency, and physical-device reliability were outside its scope.

LIVE worker-test scheduling behavior has also received additional regression coverage, although intermittent simulated-worker timing failures observed in some CI runs warrant continued investigation.

📦 6. Version Alignment and Compatibility

ISLA's version metadata is now aligned at 3.1.0 across:

• Python distribution metadata.

• Public and legacy Python version identifiers.

• HTTP service metadata.

• JUCE/CMake project configuration.

• Automated version regression tests.

Existing public interfaces, historical compatibility aliases, plugin identifiers, and DSP operating modes are preserved.

The original v3.0.0 release remains separately available as the historical ALIVE + LIVE milestone.

---

⚠️ Experimental Status and Known Limitations

ISLA remains an experimental audio research project.

LIVE currently targets macOS with BlackHole 2ch and an independently selected physical stereo output.

The Python-based monitoring engine is not guaranteed to meet hard-real-time audio requirements. Longer-duration monitoring, physical disconnection recovery, acoustic latency, broader hardware compatibility, and new controlled listening acceptance remain open engineering activities.

The existing offline nonzero-strength chunk-streaming expected failure also remains documented.

The JUCE plugin pathway requires separate native acceptance and should not be considered a certified commercial audio plugin.

No claim is made that v3.1.0 introduces three independently designed musical effects or guarantees perceptually superior sound.

---

🚀 Getting Started

Repository: https://github.com/dfeen87/ISLA

For the terminal listening experience:

python -m pip install -e '.[playback]'
python -m isla

At the isla> prompt, enter:

setup audio

For LIVE system...

V3 Alive

🎹 ISLA v3.0.0 — ALIVE + LIVE

Let the Music Flow.

ISLA is ALIVE. And now, ISLA can listen LIVE.

Version 3.0.0 marks a major milestone in ISLA's evolution from an experimental musical signal-processing engine into an interactive audio application.

With ALIVE, listeners can open local recordings and explore ISLA's musical transformations. With LIVE, macOS users can process system audio continuously—including music playing through a browser—and listen through connected speakers or headphones.

At the heart of ISLA remains one question:

What if musical transitions could be softened, connected, and transformed through mathematics while preserving the character of the original performance?

This release brings that exploration closer to everyday listening.

---

✨ What's New in v3.0.0

🎧 ALIVE — Interactive Audio Playback

ISLA now offers an interactive terminal experience for exploring local music recordings.

• Load mono or stereo WAV recordings.
• Prepare and switch between Velvet, Legato, and Melt.
• Compare transformations against the original audio using Bypass.
• Play, pause, resume, stop, and replay recordings.
• Inspect playback status and processing progress.
• Perform file preparation separately from the audio playback callback.

ALIVE builds on ISLA's existing phase-aware DSP research and offline-processing foundation.

🌊 LIVE — Real-Time System Audio Exploration

For the first time, ISLA supports continuous processing of macOS system audio through a dedicated LIVE monitoring engine.

Using BlackHole 2ch as a virtual audio input, listeners can route supported system audio through ISLA and hear the result on their physical speakers or headphones.

LIVE introduces:

• Continuous audio capture and processing.
• Stereo support at 44.1 and 48 kHz.
• Stateful frequency-domain harmonic processing.
• Weighted overlap-add reconstruction across successive audio blocks.
• Smooth switching between musical profiles.
• Time-aligned Bypass comparison.
• Device-routing checks and safeguards against unsupported output configurations.
• Monitoring status, signal diagnostics, and explicit stream cleanup.

LIVE remains experimental and is not presented as a professional, hard-real-time audio solution.

🎛️ Three Musical Profiles

Velvet — Gentle harmonic softness

Explores a softer, less sharp harmonic texture.

Legato — Musical continuity

Explores a more connected musical impression.

Melt — Flowing harmonic transformation

Applies a more pronounced harmonic-smoothing effect.

Bypass — Hear the original

Returns to the original captured signal with the same LIVE processing delay, allowing comparisons without changing playback routes.

The current LIVE profiles use related spectral-processing mathematics at different strengths. Their musical descriptions express creative intentions rather than guaranteed perceptual effects.

🖥️ Guided Audio Setup

ISLA introduces an interactive device-selection workflow intended to make LIVE easier to configure.

After launching ISLA, enter:

setup audio

The guided setup helps select the required BlackHole capture device and an eligible physical stereo output.

Existing manual commands remain available for experienced users.

ISLA does not automatically change macOS system-audio routing or start monitoring during guided setup.

---

🧠 Engineering Foundations

Version 3.0.0 extends ISLA's deterministic DSP foundation with a practical playback and monitoring architecture.

Major engineering capabilities include:

• Stateful, continuous harmonic processing.
• 2048-sample analysis windows with 512-sample processing hops.
• Preserved analysis and reconstruction state between blocks.
• Separated audio callbacks and DSP worker processing.
• Bounded buffering and explicit failure handling.
• Regression tests for signal continuity, playback behavior, device management, and DSP invariants.
• Continued support for legacy Python interfaces and existing offline-processing functionality.

ISLA's ALIVE and LIVE engines intentionally use different processing pathways. Their outputs are not guaranteed to be identical.

---

🧪 Testing and Validation

ISLA v3.0.0 builds on automated tests and native macOS acceptance work.

Recorded development validation includes:

• Python regression testing and cross-platform CI.
• Actual macOS audio capture through BlackHole 2ch.
• Playback through physical audio outputs.
• Mode switching during continuous playback.
• Numerical waveform and continuity verification.
• Native testing at 44.1 and 48 kHz.
• Device-routing validation and audio-stream cleanup.

Owner listening confirmed that all three LIVE modes produced subtle differences, with Melt preferred during the documented listening session.

These results support the functionality of the experimental system, but do not constitute comprehensive musical-quality or hardware-compatibility acceptance.

Additional stress, long-session, physical-disconnection, and broader device testing remain necessary.

---

🚀 Getting Started

ISLA is open-source and available on GitHub:

https://github.com/dfeen87/ISLA

For the macOS terminal experience, install the Python playback dependencies and launch:

python -m pip install -e '.[playback]'
python -m isla

For LIVE audio, install BlackHole 2ch separately and follow the repository's macOS routing instructions.

The local-file ALIVE experience does not require BlackHole.

Python 3.10 or newer is required.

Please review the README's setup, audio-routing, safety and recovery instructions before using LIVE.

---

⚠️ Current Limitations

ISLA is an experimental audio research project.

• LIVE currently targets macOS 13+ with BlackHole 2ch and a separate physical stereo output.
• The Python-based LIVE system is not guaranteed to meet hard-real-time audio requirements.
• Monitoring introduces latency and may affect video synchronization.
• Processing is not automatically loudness-normalized.
• Wider device and operating-system compatibility remains unverified.
• The JUCE VST/AU plugin pathway remains experimental and is not certified for professional real-time operation.
• A documented legacy nonzero-strength chunk-streaming limitation remains under investigation.

ISLA should not be considered a finished commercial audio processor.

---

🔬 What's Next?

Future research will investigate whether Velvet, Legato, and Melt can develop more independent musical characteristics through carefully validated DSP improvements.

Additional priorities include:

• Mathematical robustness and numerical precision.
• Reproducible musical listening comparisons.
• Greater perceptual differentiation between processing modes.
• Long-duration LIVE stability.
• Audio-performance optimization.
• Broader native macOS acceptance.
• Future graphical interfaces and other listening experiences.

Experimental changes will be evaluated independently before inclusion in a release.

---

❤️ A Special Dedication

ISLA is dedicated to Isla, the creator's niece.

Her name represents the inspiration behind a project that brings together musical expression, mathematical curiosity, and thoughtful engineering.

What began as an exploration of phase-aligned note smoothing has grown into a working experimental audio system.

ISLA continues to evolve through human creativity, open-source collaboration, and AI-assisted engineering.

Special recognition goes to the foundational contributions of Microsoft Copilot and Google Jules, as well as OpenAI's ChatGPT and Codex in advancing ISLA's interactive ALIVE and LIVE experiences.

And to everyone who listens, experiments, tests, and contributes:

Thank you for helping the music flow.

---

🎹 ISLA v3.0.0 — ALIVE + LIVE

Not simply changing the sound.

Exploring how music can flow.

❤️ ISLA — Let the Music Flow.