Architectural and Cognitive Frameworks for Coherent Multisensory Performances: Integrating Sound, Voice, Text, and 3D Environments in Cicero, Illinois

Full report

On this page

The orchestration of a coherent multisensory performance requires a profound synthesis of advanced web technologies, cognitive neuroscience, psychometrics, and rigorous accessibility frameworks. When situating such an interactive installation in a localized context like Cicero, Illinois—a municipality characterized by dense urban infrastructure and a rich industrial history—the performance must meticulously bridge digital constructs and physical realities. Developing a three-dimensional spatialized environment that seamlessly integrates procedurally generated sound, synthesized and recognized voice, and localized text demands a robust computational architecture. The following analysis exhaustively deconstructs the computational mechanics, psychological phenomena, and physiological parameters necessary to execute a highly immersive, cognitively optimized, and universally accessible multisensory experience.

Acoustic Infrastructure: Web-Based Audio and Procedural Synthesis

The foundation of any web-based multisensory audio performance relies on the Web Audio API. This specification provides a versatile, high-level routing graph where individual AudioNode objects are instantiated and connected to define the overall audio rendering topology1. For an installation requiring real-time, highly responsive audio mapped to exact 3D spatial coordinates, standard high-level implementations are often insufficient due to inherent performance bottlenecks within the browser's execution environment.

The AudioWorklet and Main-Thread Optimization

Historically, JavaScript-based audio processing suffered from severe latency and playback degradation because audio processing shared the main control thread with user interface rendering, DOM manipulation, and garbage collection events4. The introduction of the AudioWorklet interface structurally mitigates this by allowing custom, low-level audio processing to run on a dedicated rendering thread2. In a demanding 3D performance setting, this enables sample-accurate processing for complex synthesizers and dynamic spatial effects without being interrupted by the main thread's graphical rendering tasks, such as WebGL frame updates or DOM repaints4.

However, memory management within the Web Audio API presents highly specific edge cases that must be addressed to maintain acoustic coherence. When an AudioWorkletNode is disconnected from the audio graph via the .disconnect() method, the logical expectation is that its process() function will cease execution to conserve resources8. Empirical evidence and architectural analysis indicate that the topological flood-fill process of the audio graph continues to trigger the disconnected node's process() function as a "phantom" node until it is fully collected by the browser's garbage collector8.

This behavior introduces severe instability if the node relies on a SharedArrayBuffer to communicate with the main thread. The main thread may prematurely execute a memory reallocation or free() command on shared data structures under the assumption that the disconnected node is dormant. Consequently, the active "phantom" worklet reads freed or reallocated memory, introducing severe audio glitches, processing jank, or total application crashes8.

To sustain a continuous, uninterrupted installation in Cicero, the software architecture must utilize a rigid memory pooling strategy. Pre-allocating a persistent pool of AudioBuffer arrays and AudioWorklet instances at initialization and reusing them—rather than dynamically creating and destroying nodes—prevents memory fragmentation and eliminates the unpredictable garbage collection sweeps that cause audible latency9. Additionally, computational frameworks like Web Audio Modules (WAMs) can be deployed to encapsulate these nodes into a DAW-like plugin architecture4. Utilizing WebAssembly (WASM), developers can port complex C++ or Rust digital signal processing (DSP) libraries, or utilize Domain Specific Languages like FAUST, compiling them directly into AudioWorklet instances for near-native execution speeds11.

Offline Rendering and Data Sonification via FM Synthesis

While real-time interaction dictates the use of the standard AudioContext, archiving or exporting the performance necessitates the OfflineAudioContext3. This variant processes the audio routing graph non-real-time, rendering the output as fast as the CPU allows directly into an AudioBuffer14. This allows the installation to capture live, interactive jam sessions or procedural generative outputs and instantly export them as polished WAV or MP3 files without routing through external recording software, preserving the exact spatial and acoustic metadata of the session17.

At the core of the installation's generative audio is parameter mapping sonification, an approach where individual data dimensions drive the control parameters of an audio synthesizer19. An installation in Cicero can leverage local environmental, demographic, or historical data streams by translating this data into procedural Frequency Modulation (FM) synthesis21.

FM synthesis generates highly complex timbres by modulating the frequency of a carrier oscillator with a modulator oscillator22. By mapping live data to these oscillators, the installation generates a continuously evolving soundscape. High-frequency data spikes (e.g., local transit activity or peak wind speeds) can be mathematically mapped to a high modulation index and a wide spectral spread, while slow-moving historical demographic data can map to smooth, attenuated low-frequency drones22. Generating audio procedurally via the Web Audio API eliminates the necessity of downloading massive pre-recorded audio assets, thereby drastically optimizing the initial load time and dynamic responsiveness of the 3D application10.

Spatial Dynamics and WebXR Integration

To ground the performance physically and psychologically, sound must be accurately spatialized. In human perception, the ability to locate sound sources in a 360-degree sphere relies on Head-Related Transfer Functions (HRTFs) and sophisticated distance attenuation processing25. The Web Audio API provides the PannerNode and AudioListener interfaces to simulate these complex acoustic physics within a WebXR context25.

Distance Models, Spatial Topology, and Acoustic Environments

The AudioListener represents the user's precise position and orientation within the virtual or augmented 3D scene. As the user navigates the digital representation of Cicero, their 6-Degrees-of-Freedom (6DoF) head tracking data—captured via the WebXR Device API—continuously updates the listener's spatial coordinates25. Simultaneously, virtual sound sources are affixed to PannerNode objects. The PannerNode utilizes various distance models—such as linear, inverse, or exponential—to dynamically calculate and reduce the volume, altering the frequency response of a sound as the virtual Cartesian distance between the source and the listener increases25. (For applications only requiring simple left-to-right panning without complex 3D math, a StereoPannerNode is available, though inadequate for true spatial reality27).

High frequencies are naturally absorbed by air over long distances; this attenuation can be procedurally simulated by routing the PannerNode output through a BiquadFilterNode acting as a low-pass filter, with a cutoff frequency mathematically inversely proportional to the calculated distance25.

Furthermore, environmental realism is achieved via the ConvolverNode, which applies an impulse response (a recorded acoustic fingerprint) of a specific physical space25. By capturing a high-fidelity impulse response from a local Cicero landmark—such as a historic industrial warehouse or a specific municipal transit station—and loading it into the ConvolverNode, the digital audio will reverberate and decay exactly as it would in that localized physical space. This directly bridges the gap between the virtual construct and local physical reality, drastically enhancing immersion.

Visual Integration and Stereoscopic Alignment

The audio spatialization must operate in tandem with visual rendering, typically achieved via WebGL libraries such as Three.js, where a Scene holds Object3D meshes, a Camera defines the view, and a Renderer executes draw calls via a requestAnimationFrame loop28. Because modern browsers strictly enforce autoplay policies, the entire AudioContext and its associated spatial nodes must be instantiated or explicitly resumed within the very first user interaction (e.g., a screen tap or click) to prevent the audio engine from being permanently suspended by the browser2.

When integrating 3D visuals with spatial audio, engineers must account for visually induced motion sickness (VIMS) and stereoscopic visual fatigue, which can destroy the coherence of the performance. The ISO 9241-392 standard dictates that visual fatigue in XR is heavily influenced by the interpupillary distance (IPD) of the user relative to the distance between the centers of the virtual lenses31. Vertical or rotational misalignments between the virtual images presented to the left and right eye, as well as uncalibrated magnification differences, will introduce severe visual discomfort, nausea, and cognitive dissonance, counteracting the immersiveness generated by the spatial audio31. Strict calibration protocols adhering to IEC 63145-20-10 procedures must be implemented to measure and correct luminance, contrast, and color crosstalk between the stereoscopic outputs31.

Voice Interaction: Speech Synthesis and Recognition

A critical component of a coherent multisensory experience is the bidirectional integration of voice. The Web Speech API provides two distinct interfaces: SpeechSynthesis (Text-to-Speech) and SpeechRecognition (Speech-to-Text), allowing the installation to read dynamic text aloud and respond to user vocal inputs in real-time33.

Speech Recognition and Privacy Considerations

The SpeechRecognition interface captures audio from the microphone and processes it to return text strings. By default, most modern operating systems and browsers route this audio to a cloud-based server engine for processing33. This introduces severe privacy concerns, as users may not consent to their speech being transmitted to third-party servers, alongside the obvious performance degradation caused by network latency during a live performance34.

To mitigate these issues, developers can mandate on-device processing. This is governed by the on-device-speech-recognition Permissions-Policy directive33. When enforced, the browser downloads a local language pack and performs all recognition algorithms locally, ensuring zero data transmission and enabling offline functionality34. This local execution is paramount for maintaining the low-latency responsiveness required for a coherent 3D performance.

Speech Synthesis and Browser Fragmentation

The SpeechSynthesis interface allows the application to vocalize procedural text without requiring gigabytes of pre-recorded voice lines. Implementing this across different browser engines, however, reveals significant technical fragmentation. The API operates asynchronously; browsers like Chrome, Edge, and Firefox require an onvoiceschanged event listener to populate the available voice list from the operating system, whereas Safari populates it synchronously upon execution30. Furthermore, mobile operating systems impose strict autoplay restrictions. On iOS Safari, the speechSynthesis.speak() method must be invoked directly within the call stack of a user gesture handler, otherwise the utterance is silently discarded without triggering an error30.

Abstracting these inconsistencies requires polyfill architectures or libraries like EasySpeech, which normalize the asynchronous voice loading, provide standardized Promise-based API wrappers, and automatically implement fallback logic based on the user's specific browser and OS combination37.

The Chromium 15-Second Cutoff Anomaly

A widely documented, long-standing anomaly in the Chromium browser engine causes long SpeechSynthesisUtterance objects to abruptly cancel after approximately 15 seconds of continuous playback39. This truncation occurs regardless of the word count or character limit; it is a strictly enforced chronological timeout that silently blocks the API, preventing subsequent onend events from firing39.

To ensure continuous, uninterrupted vocal performances during the Cicero installation, the architecture must implement specific, conditional workarounds based on the detected client environment:

 

For a robust, locally hosted installation, the text chunking method remains the most stable cross-platform solution38.

Cognitive Synchronization and Audiovisual Coherence

The convergence of procedurally generated 3D visuals, spatialized data-driven sound, and synthesized voice inherently strains human cognitive processing capacities. Cognitive load theory dictates that working memory has a strictly limited capacity; when multiple sensory streams demand simultaneous processing, competition between modalities occurs43.

Dual 2-back and Go/NoGo paradigm experiments demonstrate that when visual and auditory information are encoded differently or present incongruent information, clear neural competition emerges, severely disrupting working memory performance43. Incongruence mediated by cognitive load dynamics disrupts sensory processing more than fixed sensory hierarchies43. Conversely, when audiovisual stimuli are congruent and temporally aligned, cognitive outcomes are greatly facilitated43.

The Temporal Binding Window (TBW) and Causal Inference

Sensory signals (light and sound) propagate through physical space at vastly different speeds and are processed by the nervous system at different neurological latencies. Despite this fundamental mismatch, the human brain constructs a seamlessly unified perception of synchronous events if the stimuli fall within a specific chronological threshold known as the Temporal Binding Window (TBW)45.

The TBW spans several hundred milliseconds45. If the visual rendering of a 3D object in the performance occurs within this window relative to its corresponding audio cue, the brain executes causal inference, binding the two distinct sensory inputs into a single, unified crossmodal object45. The brain essentially integrates information from past, present, and subsequent events across modalities to construct a unified percept45. This is dramatically illustrated by the "Illusory AV Rabbit Illusion," where precise temporal staggering of flashes and beeps causes the brain to hallucinate a visual flash that did not exist, simply to maintain causal synchrony with an auditory beep45.

The precision of the TBW is highly malleable and heavily influenced by prior sensorimotor experience. Individuals with extensive musical training, for example, exhibit a significantly narrower TBW for music, demonstrating increased neural activation and effective connectivity in a superior temporal sulcus-premotor-cerebellar circuitry when asynchronous stimuli violate their precise temporal predictions48.

When engineering the Cicero performance, the Synchronization Accessibility User Requirements (SAUR) and electrophysiological data dictate strict latency thresholds. The audiovisual synchronization must fall well within a ±60 ms Stimulus Onset Asynchrony (SOA) to ensure implicit audiovisual temporal integration46. Industry standards (like EN 301 549\) dictate that audio arriving up to 45–100 ms after video is generally acceptable, but audio arriving more than 15 ms before video is highly objectionable and breaks immersion46. If the delay extends beyond this TBW, the participant will perceive the visual and auditory streams as completely separate events, drastically increasing cognitive load, inducing the McGurk effect (where incongruent audio and visual speech signals fuse into a corrupted percept), and shattering the performance's coherence43.

Spatial Audio and NASA-TLX Reductions

Precise spatial audio directly reduces the listener's cognitive load. Research utilizing the NASA Task Load Index (NASA-TLX)—a validated multidimensional assessment tool measuring mental, physical, and temporal demand, alongside performance, effort, and frustration—demonstrates that spatially rendered speech significantly outperforms non-spatial (monaural or diotic) audio52.

By aligning the perceived acoustic origin of a synthesized voice with its corresponding visual avatar or text block in 3D space, spatial audio leverages the "cocktail party effect." This allows participants to effortlessly isolate target information amidst a complex auditory background52. Spatial cueing practically doubles comprehension accuracy while subtly but consistently reducing perceived mental demand and listening effort52.

Interestingly, temporal synchrony alone can drive crossmodal integration even without exact spatial alignment—a phenomenon known as the "pip and pop" effect47. When a visual target changes in temporal coherence with an auditory pip, visual search times drop significantly, as the auditory cue instantly binds to the visual event, creating a highly salient multisensory object47. The installation can utilize this mechanism by aligning abrupt textural, topological, or lighting changes in the 3D environment with precise, synthesized auditory transients.

To objectively measure the success of this coherence during development, researchers utilize continuous physiological tracking. Specifically, the synchronization of 64-channel Electroencephalography (EEG) utilizing correlated component analysis, alongside high-resolution pupillometry (tracking pupil dilation) and Heart Rate Variability (HRV), provides empirical, real-time markers of deep cognitive immersion and physiological synchrony54.

Evaluating Neuromodulation Claims: The Binaural Beat Controversy

A frequent, albeit highly controversial, technique utilized in immersive audio installations to induce altered states is the implementation of binaural beats. This auditory phenomenon occurs when two pure sine waves of slightly different frequencies (typically below 1,000 Hz, with a difference of less than 35 Hz) are presented separately to each ear58. The superior olivary complex (SOC) in the brainstem, which is the first locus to process data from both ears, processes these distinct inputs, resulting in the subjective illusion of a third, oscillating "beat" at the exact difference frequency58.

Proponents of the "brainwave entrainment hypothesis" posit that listening to specific binaural beat frequencies forces the brain's electrocortical activity to synchronize with the beat via a frequency-following response (FFR). It is claimed that targeting specific EEG bands—Delta (1-4 Hz) for deep sleep, Theta (4-8 Hz) for meditation, Alpha (8-14 Hz) for relaxation, Beta (14-30 Hz) for concentration, and Gamma (\>30 Hz) for high-level cognition—can directly induce these respective cognitive or affective states58. Many commercial applications and performance artists aggressively market these beats as non-invasive neural modulators.

Empirical Failure of EEG Entrainment and Cognitive Risks

Rigorous neuroscientific scrutiny severely challenges the efficacy of binaural beats as a reliable entrainment mechanism. A comprehensive 2023 systematic review by Ingendoh, Posny, and Heine published in PLOS ONE synthesized 14 robust electroencephalographic (EEG) studies examining the brainwave entrainment hypothesis58. The analysis revealed severe inconsistency:

 

The extreme methodological heterogeneity across the literature precludes treating binaural beats as an established neurophysiological mechanism64. Furthermore, highly controlled experimental trials have demonstrated a complete failure of binaural beats to enhance EEG power across Theta, Alpha, Beta, or Gamma bands, while also failing to induce any measurable changes in objective physiological arousal markers such as heart rate or skin conductance60. While there are rare, highly specific clinical scenarios where binaural beats show a weak positive signal—such as reducing the required dose of remimazolam during general anesthesia induction—these are exceptions rather than the rule67.

More alarmingly, large-scale ecological studies indicate that under certain conditions, binaural beat stimulation can actively impair cognitive performance, resulting in worse test scores and disrupted working memory compared to silent or neutral-sound control groups63. Safety and adverse effect monitoring in this domain has historically been exceptionally poor, and the assertion that fixed cognitive functions can be confidently assigned to individual beat frequencies is fundamentally unsupported by current empirical data63.

Consequently, integrating binaural beats into the Cicero performance with the intent of forcing a neurological state is scientifically unjustifiable and potentially detrimental to the user experience. Any subjective benefits reported by users in previous studies are vastly more likely attributable to the placebo effect, demand characteristics, or the general relaxing nature of the carrier audio (such as ambient pink noise or musical soundscapes), rather than the physiological phenomenon of the binaural beat itself67.

Psychological Immersion and Trait Absorption

If artificial auditory entrainment is scientifically flawed, how does a multisensory installation achieve deep psychological immersion? The answer lies not in manipulating the stimulus to force a neural state, but in leveraging the innate psychological traits of the audience—specifically, trait absorption.

The Tellegen Absorption Scale (TAS) is a widely utilized psychometric instrument containing 34 items that measures an individual's disposition for experiencing episodes of "total" attention71. In these episodes, representational resources are fully engaged by an activity to the point of oblivious involvement with the surrounding environment72. Originally developed in the 1970s to assess hypnotizability (the innate ability to respond to imaginative suggestions), the TAS identifies individuals who readily blur the boundaries between their internal mindscape and the external world71.

Social cognitive theories of hypnosis reject the idea of a "hypnotic trance" as a special, altered state of consciousness. Instead, these theories argue that hypnotic responding is simply a trait ability, much like musical or athletic ability, driven by expectancies and imaginative involvement70. Hypnotizability is fundamentally linked to a person's baseline capacity for absorption70.

Dimensions of the Tellegen Absorption Scale

The TAS evaluates multiple facets of imaginative involvement, which can be categorized into distinct psychological domains:

 

It should be noted that the TAS has faced some criticism for circularity, as it includes items regarding the tendency to "sense a presence," which conflates standard aesthetic absorption with hallucination or mystical experiences74. Nevertheless, high scorers on the TAS demonstrate an amplified capacity to alter their own perceptual experience without requiring formal induction or external neurological tampering70.

In a 3D XR environment, highly absorbed individuals will organically fuse the spatial audio, procedural visuals, and synthesized text into a deeply unified and subjectively "real" experience71. To maximize the impact of the Cicero installation, the design should target these facets of absorption by providing rich, ambiguous, and aesthetically resonant crossmodal cues. Rather than presenting literal representations, the environment can use abstract FM-synthesized soundscapes synchronized with shifting 3D geometry, allowing the participant's own associative memory and imaginative involvement to fill in the semantic gaps.

Immersive Accessibility and Safety Protocols

Designing a multisensory performance necessitates a rigorous adherence to accessibility standards to ensure the environment does not exclude users with disabilities or induce physical harm. The World Wide Web Consortium (W3C) has established the XR Accessibility User Requirements (XAUR) to define necessary accommodations within virtual, augmented, and mixed reality paradigms78.

Motion-Agnostic Inputs and Semantic Customization

XR environments traditionally rely on 6DoF spatial tracking, requiring the user to physically turn, walk, or reach to interact with the environment81. For individuals with motor impairments, neurological conditions, or vestibular disorders, these mechanics are highly exclusionary. XAUR User Need 2 mandates "Motion agnostic interactions," requiring that all interactions within the 3D space can be executed via device-independent alternative inputs, such as keyboard navigation, voice commands, or specialized switch devices79.

Similarly, visually impaired users navigating a 3D representation of Cicero require rich spatial semantics. XAUR User Need 1 emphasizes the necessity for intuitive navigation mechanisms with robust affordances79. Assistive technologies must be able to parse the 3D scene graph, identifying objects, locations, and interactive elements. In spatialized environments, the combination of screen-reader support (via the Web Speech API) and sonic symbolism (earcons or spatial audio beacons) allows users to build a reliable mental map of the 3D layout without visual input79.

Spatial Text and WebVTT Captioning

For users who are deaf or hard of hearing, as well as those experiencing high cognitive load, text captioning is critical. In a standard 2D web environment, the Web Video Text Tracks (WebVTT) format is the standardized method for displaying timed text, operating via \<track\> elements and strictly aligning text to media timelines83. Furthermore, the RTC Accessibility User Requirements (RAUR) advocate for advanced UI controls, such as the ability to "pin" specific windows containing American Sign Language (ASL) interpreters alongside the primary content, ensuring continuous visual access to translation78.

In a 3D XR context, flat, screen-space captions placed at the bottom of a viewport are inadequate, as the user can physically look away from the viewport's edge. Spatial captioning involves anchoring the text near the origin of the sound in the 3D environment, behaving as a diegetic object87. The WebVTT format allows for precise control over cue placement using line and position settings within the VTT file, or via the lineAlign and positionAlign properties directly in JavaScript88.

The visual presentation of these captions is controlled via the CSS ::cue pseudo-element89. This allows developers to dynamically alter the typography, background contrast, and text color of the captions to ensure readability against complex, moving 3D backgrounds89. The inheritance model allows specific internal node objects (such as voice tags \<v\>) within the WebVTT cue to be styled independently, enabling distinct visual markers and colors for different synthesized voices operating simultaneously within the scene90.

Mitigation of Photosensitive Seizures

The absolute most critical safety requirement in any visual performance involves the prevention of photosensitive epilepsy (PSE) triggers. Rapidly flashing lights, high-contrast transitions, and bold geometric patterns can induce reflex seizures or severe visually induced motion sickness31. The Web Content Accessibility Guidelines (WCAG) 2.1 Success Criterion 2.3.1 (Three Flashes or Below Threshold) and the hardware-oriented ISO 9241-307 / 9241-391 standards strictly regulate visual flicker31.

The physical hazard threshold is defined mathematically. A sequence is deemed hazardous if it contains more than three flashes within any one-second period (a frequency of [Figure omitted from source export] Hz) and occupies a solid visual angle of at least 0.006 steradians (approximately 10% of the central visual field)91.

Additionally, the human retina is uniquely vulnerable to highly saturated red transitions, as deep red light stimulates long-wavelength sensitive cones without triggering the standard inhibitory mechanisms from medium and short-wavelength cones92. Under WCAG specifications, a saturated red state is defined mathematically as:

[Figure omitted from source export]

A transition involving this state is a strict violation if the colors differ by more than 0.2 units on the CIE 1976 UCS diagram92.

In the Cicero installation, any procedural 3D lighting, strobe effects, or rapid texture swapping must be clamped via a rendering algorithm that intercepts and throttles visual updates to stay strictly below the 3 Hz limit. Any strobing effects must automatically dim to ensure the relative luminance change between the dark and light states remains under 10%92. Modern operating systems offer prefers-reduced-motion settings (such as GNOME's WPE\_SETTING\_REDUCED\_MOTION); the 3D application must query this media feature and automatically disable all non-essential spatial animations if the user has flagged a sensitivity95. Implementing a safe-harbor toggle that mutes all non-critical environmental animations ensures compliance with XAUR personalization requirements and protects vulnerable participants from physical harm79.

Conclusion

The deployment of a coherent multisensory performance in Cicero, Illinois, represents a highly complex convergence of web audio engineering, cognitive psychology, and rigorous accessibility compliance. By prioritizing low-level AudioWorklet threads, memory pooling, and FM data sonification, the system can bypass the limitations of the JavaScript main thread to deliver latency-free, procedurally generated spatial soundscapes that reflect the local environment. Integrating the Web Speech API with advanced chunking algorithms ensures uninterrupted vocal delivery, while spatializing both audio and WebVTT text dramatically reduces the participant's cognitive load by leveraging the brain's natural temporal binding windows and the cocktail party effect.

Crucially, the installation must abandon unsupported neuromodulation theories, such as binaural beats, which risk impairing cognitive performance, and instead foster genuine psychological absorption through high-quality, congruent audiovisual design that caters to trait imaginative involvement. By rigidly enforcing WCAG and ISO seizure-prevention thresholds alongside W3C XAUR motion-agnostic navigation principles, the performance guarantees a safe, inclusive environment. This comprehensive architectural framework ensures that the intersection of sound, voice, text, and 3D space operates not merely as an aesthetic technical demonstration, but as a cognitively optimized and universally accessible digital reality.

Works cited

1. Web Audio Processing: Use Cases and Requirements \- W3C, https://www.w3.org/TR/webaudio-usecases/

2. Web Audio API \- W3C, https://www.w3.org/TR/2018/CR-webaudio-20180918/

3. Web Audio API 1.1 \- W3C, https://www.w3.org/TR/webaudio-1.1/

4. WAM-studio, a Digital Audio Workstation (DAW) for the Web, https://www.researchgate.net/publication/370414080\_WAM-studio\_a\_Digital\_Audio\_Workstation\_DAW\_for\_the\_Web

5. Building a Live Coding Audio Playground | jakelazaroff.com, https://jakelazaroff.com/words/building-a-live-coding-audio-playground/

6. Show HN: I'm building a browser-based DAW \- Hacker News, https://news.ycombinator.com/item?id=32165328

7. AudioWorklets create too much garbage \[40111364\] \- Chromium Issue, https://issues.chromium.org/40111364

8. Removing AudioWorkletNode from the AudioContext graph does not, https://github.com/WebAudio/web-audio-api/issues/2658

9. Audio playback slows down game \- javascript \- Stack Overflow, https://stackoverflow.com/questions/67293032/audio-playback-slows-down-game

10. Best architecture for scaling a browser-based Web Audio application, https://www.techrepublic.com/forums/discussions/best-architecture-for-scaling-a-browser-based-web-audio-application/

11. Ten Years of Web Audio Modules: Audio Plug-ins for the Web, https://www.researchgate.net/publication/398764786\_Ten\_Years\_of\_Web\_Audio\_Modules\_Audio\_Plug-ins\_for\_the\_Web

12. Powered By Faust \- Faust Programming Language \- Grame, https://faust.grame.fr/community/powered-by-faust/

13. Audio — list of Rust libraries/crates // Lib.rs, https://lib.rs/multimedia/audio

14. Web Audio API \- MDN Web Docs \- Mozilla, https://developer.mozilla.org/en-US/docs/Web/API/Web\_Audio\_API

15. Creating Audio on the Web Is Easy—Until It's Time to Export, https://danielbarta.com/export-audio-on-the-web/

16. TypeScript/index.d.ts at master \- GitHub, https://github.com/Tonejs/TypeScript/blob/master/index.d.ts

17. Real-Time Collaborative DAW: Your 2026 Production Guide, https://www.soundbridge.io/real-time-collaborative-daw-your-2026-production-guide

18. How can I export the OfflineAudioContext rendered audio buffer to a, https://stackoverflow.com/questions/42911622/how-can-i-export-the-offlineaudiocontext-rendered-audio-buffer-to-a-ogg-wav-or-m

19. Hearing the Invisible: Spatial Sonification for Scientific Discovery in VR, https://www.intechopen.com/online-first/1246021

20. A State-of-the-Art Report on the Integration of Sonification ... \- arXiv, https://arxiv.org/html/2402.16558v2

21. Creating Aesthetic Sonifications on the Web with SIREN \- arXiv, https://arxiv.org/html/2403.19763v1

22. GitHub \- strikeslip/SeisClaw, https://github.com/strikeslip/SeisClaw

23. Creating Aesthetic Sonifications on the Web with SIREN \- arXiv, https://arxiv.org/pdf/2403.19763

24. Spectastiq: a web component for exploring audio spectrograms, https://hardiesoft.com/posts/spectastiq-spectrogram-viewer

25. Spatial Audio for Websites Explained \- SUPADARK, https://supadark.com/notes/how-spatial-audio-works-for-websites

26. Unlocking HTML's Hidden Powers: 10 Surprising Things HTML Can, https://www.devtechpulse.me/blog/html-can-do-that

27. Web audio spatialization basics \- Web APIs \- MDN Web Docs \- Mozilla, https://developer.mozilla.org/en-US/docs/Web/API/Web\_Audio\_API/Web\_audio\_spatialization\_basics

28. Three.js Documentation \- GitHub Pages, https://expelledboy.github.io/threejs-manual-generator/

29. Three.js Visual & Interactive Encyclopedia \- A Complete Guide, https://neuralpixelgames.github.io/threejs-visual-guide/

30. Speech Synthesis API: Browser Support, Voices, Limitations, https://www.testmuai.com/learning-hub/speech-synthesis-api-browser-support/

31. Visual performance standards for virtual and augmented reality, https://www.frontiersin.org/journals/virtual-reality/articles/10.3389/frvir.2025.1575870/pdf

32. Visual performance standards for virtual and augmented reality, https://www.frontiersin.org/journals/virtual-reality/articles/10.3389/frvir.2025.1575870/full

33. Web Speech API \- MDN Web Docs \- Mozilla, https://developer.mozilla.org/en-US/docs/Web/API/Web\_Speech\_API

34. Using the Web Speech API \- MDN Web Docs, https://developer.mozilla.org/en-US/docs/Web/API/Web\_Speech\_API/Using\_the\_Web\_Speech\_API

35. Web Speech API \- GitHub Pages, https://webaudio.github.io/web-speech-api/

36. Web Speech API: Build Browser TTS, Best Practices, and ... \- DupDub, https://www.dupdub.com/blog/web-speech-api-tts

37. GitHub \- leaonline/easy-speech: Cross browser Speech Synthesis, https://github.com/leaonline/easy-speech

38. Cross browser speech synthesis \- the hard way and the easy way, https://dev.to/jankapunkt/cross-browser-speech-synthesis-the-hard-way-and-the-easy-way-353

39. SpeechSynthesis.speak (in Web Speech API) always stops after a, https://stackoverflow.com/questions/42875726/speechsynthesis-speak-in-web-speech-api-always-stops-after-a-few-seconds-in-go

40. Speech Synthesis stops abruptly after about 15 seconds \[41294170\], https://issues.chromium.org/41294170

41. SpeechSynthesisUtterance is stopping \[41084789\] \- Chromium Issue, https://issues.chromium.org/41084789

42. Speech gets cut off in firefox when page is auto-refreshed but not in, https://stackoverflow.com/questions/54328556/speech-gets-cut-off-in-firefox-when-page-is-auto-refreshed-but-not-in-google-chr

43. (PDF) Exploring the effects of audiovisual incongruence on working, https://www.researchgate.net/publication/392420290\_Exploring\_the\_effects\_of\_audiovisual\_incongruence\_on\_working\_memory\_performance\_in\_the\_combined\_2-back\_GoNoGo\_paradigm

44. Investigating the Effects of Plausibility Illusion and Congruency on, http://arno.uvt.nl/show.cgi?fid=162864

45. Causal inference shapes crossmodal postdiction in multisensory, https://pmc.ncbi.nlm.nih.gov/articles/PMC12929559/

46. The characteristics of audiovisual temporal integration in streaming, https://academic.oup.com/cercor/article/33/24/11541/7328987

47. Audio-visual spatial alignment improves integration in the presence, https://www.cmu.edu/dietrich/psychology/shinn/publications/pdfs/2020/2020neuropsych\_fleming.pdf

48. Long-term music training tunes how the brain temporally binds, https://www.pnas.org/doi/10.1073/pnas.1115267108

49. Long-term music training tunes how the brain temporally binds, https://www.pnas.org/doi/abs/10.1073/pnas.1115267108

50. Synchronization Accessibility User Requirements \- W3C on GitHub, https://w3c.github.io/saur/

51. A possible neurophysiological correlate of audiovisual binding and, https://pmc.ncbi.nlm.nih.gov/articles/PMC4244540/

52. Spatial Audio Rendering for Real-Time Speech Translation in Virtual, https://www.tandfonline.com/doi/full/10.1080/10447318.2026.2691238

53. Spatial Audio Rendering for Real-Time Speech Translation in Virtual, https://www.tandfonline.com/doi/pdf/10.1080/10447318.2026.2691238

54. Cognitive processing of a common stimulus synchronizes brains, https://pmc.ncbi.nlm.nih.gov/articles/PMC9802497/

55. Abstract \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC8842476/

56. This electronic thesis or dissertation has been downloaded from the, https://kclpure.kcl.ac.uk/portal/files/327243975/2025\_Le\_Cunff\_Anne-Laure\_1888236\_ethesis.pdf

57. Physiologically Adaptive Systems Across the Mixed Reality Continuum, https://edoc.ub.uni-muenchen.de/35223/1/Chiossi\_Francesco.pdf

58. (PDF) Binaural beats to entrain the brain? A systematic review of the, https://www.researchgate.net/publication/370900678\_Binaural\_beats\_to\_entrain\_the\_brain\_A\_systematic\_review\_of\_the\_effects\_of\_binaural\_beat\_stimulation\_on\_brain\_oscillatory\_activity\_and\_the\_implications\_for\_psychological\_research\_and\_intervention

59. Binaural beats to entrain the brain? A systematic review of the ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC10198548/

60. Effects of binaural and monaural beat stimulation on attention and, https://hersencentrum.nl/Research/Engelbregt2021\_Effects-Binaural-MonauralBeat-stimulation.pdf

61. Binaural Beat: A Failure to Enhance EEG Power and Emotional, https://www.researchgate.net/publication/321078055\_Binaural\_Beat\_A\_Failure\_to\_Enhance\_EEG\_Power\_and\_Emotional\_Arousal

62. A new perspective on binaural beats: Investigating the effects ... \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC11290623/

63. Binaural Beats: What Are They and What Are the Benefits? \- WebMD, https://www.webmd.com/balance/what-are-binaural-beats

64. Binaural beats to entrain the brain? A systematic review of the, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0286023

65. Binaural beats: evidence, effects, and risks \- yeshcube, https://yeshcube.com/en/research/binaural-beats-experimental-support-and-hypothesis/

66. Binaural beats to entrain the brain? A systematic review of ... \- PubMed, https://pubmed.ncbi.nlm.nih.gov/37205669/

67. Preoperative binaural beats reduce remimazolam dosage and, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0345960

68. Possible Effect of Binaural Beat Combined With Autonomous, https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2019.00425/pdf

69. Effects of self-administered binaural beats on meditative and, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0335580

70. Hypnotizability and the Natural Human Ability to Alter Experience, https://www.researchgate.net/publication/348896051\_Hypnotizability\_and\_the\_Natural\_Human\_Ability\_to\_Alter\_Experience

71. Consciousness and Cognition, https://thefpr.org/wp-content/uploads/Lifshitz-van-Elk-Luhrmann-Consciousness-and-Cognition.pdf

72. 209 Ioqv \- UNT Digital Library, https://digital.library.unt.edu/ark:/67531/metadc501110/m2/1/high\_res\_d/1002775157-Cawood.pdf

73. Relating the Tellegen Absorption Scale to Ideomotor Tasks, https://jewlscholar.mtsu.edu/bitstreams/0d29cbb9-3ae9-40ce-90b9-217fa95e4a9a/download

74. Hallucinations and the meaning and structure of absorption \- PMC, https://pmc.ncbi.nlm.nih.gov/articles/PMC8364205/

75. (PDF) Social cognitive theories of hypnosis \- ResearchGate, https://www.researchgate.net/publication/284338016\_Social\_cognitive\_theories\_of\_hypnosis

76. Music and hypnosis for well-being in retirement homes: A pilot study, https://www.tandfonline.com/doi/full/10.1080/00029157.2024.2435953

77. Tellegen Absorption Scale, https://www.ocf.berkeley.edu/\~jfkihlstrom/TAS.htm

78. Digital Accessibility User Requirements \- W3C, https://www.w3.org/WAI/research/user-requirements/

79. XR Accessibility User Requirements \- W3C, https://www.w3.org/TR/xaur/

80. Investigating VR Accessibility Reviews for Users with Disabilities, https://arxiv.org/html/2508.13051v2

81. What Augmented Reality Still Needs to Learn About Accessibility, https://wiprotechblogs.medium.com/beyond-the-hype-what-augmented-reality-still-needs-to-learn-about-accessibility-955d288f97e5

82. XR Accessibility User Requirements now a published W3C Note, https://www.accessibility.org.au/xr-accessibility-user-requirements-now-a-published-w3c-note/

83. Subtitle Resources & Standards — Formats, Netflix Rules, CPS Limits, https://videotext.io/subtitle-resources

84. Captions/Subtitles | Web Accessibility Initiative (WAI) \- W3C, https://www.w3.org/WAI/media/av/captions/

85. Browser compatibility \- Shiar's cheat sheets, https://sheet.shiar.nl/browser/css-supports-api

86. RTC Accessibility User Requirements \- W3C on GitHub, https://w3c.github.io/raur/

87. Accessible VR and WCAG: what maps, what breaks, and the, https://amplifiedcreations.com/journal/wcag-for-vr-accessibility

88. Web features explorer \- Limited availability, https://web-platform-dx.github.io/web-features-explorer/limited-availability/

89. Pseudo-elements \- CSS \- MDN Web Docs, https://developer.mozilla.org/en-US/docs/Web/CSS/Reference/Selectors/Pseudo-elements

90. cue CSS pseudo-element \- MDN Web Docs, https://developer.mozilla.org/en-US/docs/Web/CSS/Reference/Selectors/::cue

91. Web accessibility for seizures and physical reactions \- MDN Web Docs, https://developer.mozilla.org/en-US/docs/Web/Accessibility/Guides/Seizure\_disorders

92. Photosensitive Seizure Standards Compared: WCAG, Ofcom, ITU, https://video-audit.com/blog/seizure-safety-standards-compared

93. Final draft EN 301 549 V2.1.2 (2018-06) \- ETSI, https://www.etsi.org/deliver/etsi\_en/301500\_301599/301549/02.01.02\_30/en\_301549v020102v.pdf

94. Animations and Visual Effects \- University of Illinois, https://publish.illinois.edu/accessibility-training/2017/08/09/animations-and-visual-effects/

95. Planet WebKit, https://planet.webkit.org/