Open research questions in Music Technology and Sound Studies

What the literature leaves open

• This paper presents our methodology and results, and argues that without conceptual clarity on touch, important dimensions of musical expression and embodied interaction remain under-explored in NIME.

  What NIME talks about when it talks about touch · 2026 · DOI

• We would like to continue the competition and strengthen Ren- Con community building in the future. Collaborating with the MIREX initiative allows us to explore new research directions beyond standard rendering. For example, we could introduce tasks like predicting the technical difficulty of a piece from its score, which has applications in both music education and automated curation. Another plan for the future is moving toward more diverse data modalities. While most current systems focus on MIDI or audio, we want to include embodied performance data, such as the physical key and pedal movements captured by Yamaha Disklavier systems. The recent success of models like Contin-U shows that we can now work directly with score images to produce audio, effectively merging optical music recognition with expressive performance. This suggests that the competition should broaden its scope to include systems that process various inputs — from scanned sheet music to physical performance captures.

  RenCon 2025: Revival of Expressive Performance Rendering Competition · 2026 · DOI

• One limitation observed empirically, and also reported through an extended user study in related work, concerns the lack of support for saving or exporting modified, or “bent,” network states [15].

  Network Bending as Circuit-Bending Inspired Live Neural Synthesis Hacking · 2026 · DOI

• Future iterations of the MidiMbira could involve technical enhancements such as expanded microtonal capabilities, more intuitive MIDI mappings, and increased integration with virtual and augmented reality environments for immersive learning and performance experiences. Collaborative efforts could also be expanded internationally, creating networks of DIY researchers, educators, and musicians dedicated to adapting traditional instruments into digital hybrids. Building on the experience of the media-arts festival, one promising possible direction involves the deployment of interactive MidiMbira prototypes in museum contexts.

  The DIY MidiMbira: Bridging Traditional Playability and MIDI Technology forPedagogy, Innovation, and Performative Practice · 2026 · DOI

• Future work of Mezcal is already underway. Some of it is to address deficiencies or shortcomings in the current iteration. Some of it is to explore new ideas and endless features. The current design of Mezcal still needs some treatment. A lot of time and energy went into finding an ergonomic form-factor the allows various faders, search modals, controls and more to fit within tiny mobile devices. I’ve been stubborn about having full sized faders and the current layout is pretty cramped. A new layout is exploring the use of dial knobs in place of faders. I am also removing playlists. There are also minor features that could improve the performance aspect. Being able to group tracks would allow one to control volume for many sounding elements at once. Better handling of default settings and being able to save settings for various input devices could improve user experience. Currently, Mezcal sets all input to "echo-cancellation" on as a default. Besides layout and usability improvements, I am exploring some new infrastructural ideas. One current prototype has abandoned the use of Janus entirely and has a built-in MCU that acts as an audio server in the browser itself. This is an interesting development for many reasons. First, it can allow operators to do main downmix by hand, potentially on site. I am looking to explore pulling live inputs from mobile devices into theatrical ambisonic settings. Second, it allows Mezcal to be deployed as a slim service that only passes initial handshaking signals. The bandwidth can be offloaded to the participants in a way that allows Mezcal to more easily be handed out without cost. The current Mezcal is more of a service than a software. It needs multiple living components at once.

  Mezcal: A Collaborative Transmission Art Instrument · 2026 · DOI

• Future updates to the hardware and software components of the installation system will provide more variety in the work’s sonic collages, incorporating features that deepen the sense of a community’s impact on the sound world created in the Divination Phase. We plan to build an alternate version of the planchette that includes a capacitive touch pad on its top, enabling us to determine if the interface is being moved by a single user or a group. This will allow us to build a collection of sensor mapping presets and tie the selection of the configuration used in performance to the amount of participants working together at the Ouija board, resulting in some compositional states that can only be achieved through communal interaction. To enhance the spatial audio component of the piece, an ambisonic version of Host Your Ghosts for periphonic loudspeaker array is currently in the works. Each new analysis of Statement Phase audio resets the location of samples across the 2D plotters, so our method of virtually mapping the location of corpora-based sound spaces to corresponding locations on our physical Ouija board can result in large sonic dead spaces where participants’ movement of the planchette to certain symbols results in no sound being triggered sound from the CBCS engine. While we enjoy the indeterminate and chance-based nature of corpus distribution that the Ouija patch presents us with in each new Divination Phase, we have also begun to explore methods of adjusting the clustering scheme of the samples prior to their plotting through the use of an intermediate algorithm like those explored in prior projects by Diemo Schwarz and Ianis Lallemand [19, 29]. This will give us greater control over how our samples are virtually distributed to align with important spots on the physical board. To expand upon the concept of time-displaced contributions of audio collected in the Statement Phase, we plan to introduce a web-browser-based portal built using the p5.js5 and Tone.js6 as an alternative to the recording station and Max patch set-up.

  Host Your Ghosts: Recontextualizing the Ouija Board as a Communal NIME for Music Performance · 2026 · DOI

• CamJam demonstrates that camera-based sensing can support accessible, embodied musical interaction without requiring spe- cialized controllers or musical training. The system achieved comparable usability across musical backgrounds while main- taining a low barrier to entry. However, true collaborative music- making requires more than concurrent access: it demands design mechanisms that explicitly scaffold listening, turn-taking, and coordinated expression. Future work will focus on three areas: (1) enhancing real-time responsiveness through optimized computer vision pipelines and automatic sensor calibration, including formal end-to-end latency measurement, (2) implementing explicit collaboration cues such as role-rotation prompts, visual indicators of interdependence, and pleasant chimes reinforcing successful coordination, and (3) conducting longitudinal studies in ecologically valid contexts (fes- tivals, exhibitions) with diverse populations including children and individuals with disabilities. Technical improvements should address better-trained emo- tion recognition models capturing diverse expressions and eth- nicities, alternative emotion representations (avatars or abstract modes) for users uncomfortable with facial tracking, and en- hanced visual feedback (animated drum hits, string motion, wave- form evolution). The modular architecture enables rapid iteration, and the purely software-based approach ensures reproducibility across different hardware configurations. Evaluation methodologies should be expanded to include stan- dardized measures (System Usability Scale, enjoyment scales), task-based assessments (playing specific melodies), and compari- son with non-DMI baselines to better understand the system’s actual advantages for non-musicians. Testing should also exam- ine whether the string interaction’s success reflects resemblance to familiar instruments or simply clearer mappings, informing future design philosophy.

  CamJam: A Modular Collaborative and Accessible Digital Musical Interface · 2026 · DOI

• Developing SfM surfaced questions about prediction-based in- teraction that warrant further investigation: the tension between pre-reflective engagement and cognitive strategizing, the role of prediction horizon in phenomenological experience, and the interplay between visual and sonic attention in multimodal AR environments.

  Sounds from Mismatch: Sensorimotor Prediction Error as Sonic Material in Augmented Reality · 2026 · DOI

• Our study was conducted with a prototype in a controlled setup and not in a museum. When more people, especially children, use our exhibit, we expect a wider set of behavioral patterns and edge cases to become apparent. The sounds, note-space, and back- ing track also have a high potential to increase expressiveness. Our current choices in this regard also make assumptions about what is musically pleasant, coming from a Western cultural back- ground. Different sounds and modulation options might improve the overall experience. Based on our findings, it seems feasible to look deeper into the overall flow of the experience, including a short tutorial sequence, introducing additional complexity, or dynamically offering support or affirmation when needed.

  Feeling Connected: Designing Instruments and Haptic Feedback for Collaborative Music Exhibits · 2026 · DOI

• The augmented Harp-E demonstrates that mechanically grounded augmentation can extend instrumentality while preserving con- tinuity across acoustic, material, and computational domains. By coupling the harp’s string structure to architectural space and aligning synthesis with articulatory logic, the system main- tains coherent gesture–sound relationships while expanding into visual, spatial, and participatory dimensions. The work there- fore shifts augmentation away from abstract parameter mapping toward material propagation and perceptual legibility. This paper establishes a proof of concept rather than a compre- hensive technical, perceptual, or performative evaluation. Future work will study latency, responsiveness, and performer percep- tion of physically mediated delay. It will also examine how virtu- osic performance and participatory interaction coexist in larger settings, and how distributed agency, audience intervention, and shared material interfaces reshape authorship, control, and musi- cal structure over longer temporal forms.

  Extending Instrumentality Through Mechanical Augmentation and Sound Synthesis in a Microtonal Harp · 2026 · DOI

• This paper has presented Hypercuíca as an augmented percus- sion instrument conceived for live electroacoustic improvisation, combining the physical practice of cuíca playing with real-time digital processing as an open-ended framework for artistic and technical exploration. Rather than proposing a radically new sens- ing technology, the project investigates how the augmentation of a culturally specific percussion instrument can generate new relationships between gesture, timbre, and embodiment across different performance contexts. The instrument was presented in performances and workshops, suggesting its potential as a tool for artistic inquiry and its capacity to extend the expressive possibilities of the cuíca beyond traditional contexts. Future work includes further developing the system’s map- pings and expanding its sonic variability through continued ex- perimentation. This includes exploring alternative gesture-to- sound relationships, as well as machine learning approaches ca- pable of identifying and mapping longer, more complex gesture trajectories, potentially introducing more dynamic and interac- tive modes of engagement with the instrument. The ongoing performance and dissemination of Hypercuíca will remain essen- tial to the project, allowing the instrument to continue evolving through real-world artistic contexts and inventive practices.

  Hypercuica: Augmenting and Performing with an Afro-Brazilian Friction Drum · 2026 · DOI

• This paper has described artistic research interrogating the possibilities for dynamic tonality in shaping hybrid performance ecologies where varied musical resources intertwine.

  Grounded Yet Shifting: Shaping Hybrid Performance Ecologies with Dynamic Tonality · 2026 · DOI

• Drawing on how the bassnicophone and other Lunason instruments naturally produce infrasound, we are de- veloping an acoustic infrasonic source to induce low frequencies into CLSTR1’s frame—an underexplored direction in instrument design that the Lunason team and I are excited to pursue.

  Recipe for a ghost catcher: an infrasound-powered hybrid instrument-sonic installation · 2026 · DOI

• We present VESPer, an open-source library for integrating EVANS Hybrid Sensory Percussion and Unity, along with three etudes showcasing its potential as a practice aid, pedagogical game, and joyful catalyst. Building on our design reflections and practice Exploring Real-Time Interfaces With Sensory Percussion NIME 2026, June 23–26, 2026, London, UK [10] Thomas Fredericks. 2023. UnityOSC. https://github.com/thomasfredericks/ UnityOSC [11] Connie Hale and Susan Green. 2009. Six Key Principles for Music Assessment. Music Educators Journal 95 (05 2009), 27–31. https://doi.org/10.1177/00274321 09334772 [12] junferno. 2021. The Title of this Video was Typed with Donkey Kong Bongos. https://www.youtube.com/watch?v=ueZaTTSFU28 [13] Tod Machover and Joe Chung. 1989. Hyperinstruments: Musically Intelligent and Interactive Performance and Creativity Systems. International Computer Music Conference Proceedings (1989), 186–190. [14] Lisa Martin. 2020. Using Synchronous Formative Feedback to Facilitate Student Growth. Music Educators Journal 107, 2 (2020), 51–57. https://doi.org/10.117 7/0027432120976816 arXiv:https://doi.org/10.1177/0027432120976816 [15] Jonas Engelhardt Moritz Simon Geist. 2021. Popcorn Robot. https://www.mo ritzsimongeist.com/works/popcorn-robot [16] Adriana Sa. 2014. Repurposing Video Game Software for Musical Expression: A Perceptual Approach. In Proceedings of the International Conference on New Interfaces for Musical Expression. Goldsmiths, University of London, London, United Kingdom, 331–334. https://doi.org/10.5281/zenodo.1178925 [17] David Stanoch. 2009. Mastering the Tables of Time. Alfred Music, Los Angeles, CA. [18] Sunhouse Technologies, Inc. [n. d.]. Josh Green’s Epic Sensory Percussion- Controlled Light Rig. https://sunhou.se/blog/epic- drum- lights- setup https://sunhou.se/blog/epic-drum-lights-setup, Last accessed on 2025-04-28. [19] Sunhouse Technologies, Inc. [n. d.]. Philo Tsoungu Controls Worlds with Sensory Percussion. https://sunhou.se/blog/philomene-tsoungui-xr-event- wrap-up https://sunhou.se/blog/philomene-tsoungui-xr-event-wrap-up, Last accessed on 2025-06-12. [20] Sam Trolland, Alon Ilsar, and Jon McCormack. 2025. Visually-Led Design for Gestural Audiovisual Instruments. , Article 45 (June 2025), 9 pages. https: //doi.org/10.5281/zenodo.15699633 [21] Ge Wang. 2016. Game Design for Expressive Mobile Music. In Proceedings of the International Conference on New Interfaces for Musical Expression. Queensland Conservatorium Griffith University, Brisbane, Australia, 182–187. https: //doi.org/10.5281/zenodo.1176141 [22] Matthew Wright and Adrian Freed. 1997. Open SoundControl: A New Protocol for Communicating with Sound Synthesizers. In International Computer Music Conference (ICMC). 101–104. [23] xPaw. [n. d.]. SteamDB. https://steamdb.info/ https://steamdb.info/, Last [24] Louie Zong. 2023. Rhythm Hell. https://www.youtube.com/watch?v=a- accessed on 2025-04-29. EI30uucng A Video Examples Demonstrations of the etudes described above may be found here: https://www.youtube.com/watch?v=OXlaIJBN1iE B Code Repository The demos and starter code related to this paper can be down- loaded here: https://github.com/the-bard-in-the-lab/ehsp-interface-demos case studies, we discuss relevancy and legibility as important factors for synchronous formative feedback interfaces. We see opportunities to continue evaluating VESPer’s con- structive value across contexts. The first author’s case studies (section 5) are non-exhaustive evidence for VESPer in a longitudi- nal, real-life practice setting. Percussionists from different genres and experience levels will likely have different experiences with this toolkit. Future studies could engage with how new learners interact with VESPer, how educators might incorporate real-time interfaces into their lessons, or the different experiences between Classical and Jazz percussionists. Additionally, measuring the extent to which synchronized visuals affect a performer’s sense of rhythm quantitatively could provide insight for designers and educators. Regarding VESPer as an open-source library, we pave the way for developing more robust musical interfaces and games using EHSP. Future development work includes more specific parsing for OSC messages, such as MIDI CC messages, and making the interface compatible with other hybrid percussion. EHSP can also receive MIDI information from other devices, including virtual ones, which would allow Unity to influence sound selection in EHSP during a performance. We also hope to use EHSP’s unique features to create more etudes, such as games about smooth zone changes or fast transitions between techniques.

  Exploring Real-Time Interfaces With Sensory Percussion · 2026 · DOI

• From a signal processing point of view, the paths to follow are many. Modifications to the merged-cross structure can be added to tune the fundamental frequency and duration of the sound as needed, like adding all-pass filters and fractional delays like in the modifications introduced to the Karplus-Strong algorithm. This would allow us to have independent control of the pitch and timbre of the resulting signal, allowing a more customizable experience for the performer. In addition, further experiments will be done connecting more than one crossing sections in series, and real-time controls for the amount of crossings in the structure. This can give an option of utilizing the knot theory rules to distinct such topologies. Second, a meaningful connection with a GUI should be implemented to enrich the user experience. The development of the VST plugin with an interface resembling the one presented on Section 2.2 would help us transmit the physical and mathematical inspiration of the algorithm and let the user interact directly with changes in the structure, rather than more abstract parameters. Finally, the possibility of embedded hardware implementations allows the integration with physical objects and interfaces. NIME ’26, June 23–26, 2026, London, UK Kasich, Chacra Georg Essl. 2021. Iterative phase functions on the circle and their projections: Connecting circle maps, waveshaping, and phase modulation. In Perception, Representations, Image, Sound, Music: 14th International Symposium, CMMR 2019, Marseille, France, October 14–18, 2019, Revised Selected Papers 14. Springer, 681–698. Georg Essl. 2021. Topologizing Sound Synthesis via Sheaves. In 2021 24th International Conference on Digital Audio Effects (DAFx). IEEE, 260–267. Georg Essl. 2022. Deforming the oscillator: iterative phases over parametrizable closed paths. (2022). Adrian Freed. 2008. Application of new Fiber and Malleable Materials for Agile Development of Augmented Instruments and Controllers. In Proceedings of the International Conference on New Interfaces for Musical Expression. Genoa, Italy, 107–112. https://doi.org/10.5281/zenodo.1179539 W Austin Haney, Andrew J Clark, and Theodore A Uyeno. 2020. Characterization of body knotting behavior used for escape in a diversity of hagfishes. Journal of Zoology 310, 4 (2020), 261–272. Jack Hardwick, Stefano Fasciani, and Çağrı Erdem. 2024. CordChord: A String Instrument with Optical Sensing. In Proceedings of the International Conference on New Interfaces for Musical Expression. NIME, 104–109. Thomas Hélie and Martin Hasler. 2003. Representation of the weakly nonlinear propagation in air-filled pipes with Volterra series. In International IEEE Workshop on Nonlinear Dynamics of Electronic Systems. Scuol, Suisse, France, 105–108. https://hal.science/hal-01106126 cote interne IRCAM: Helie03e. Laddy Patricia Cadavid Hinojosa. 2020.

  Knotty Oscillator: Breaking knot topology for a new physically-inspired sound generator · 2026 · DOI

• This study was a small feasibility-oriented pilot. Its purpose was to examine whether the proposed workflow could operate in a real prototype, from calibration and profile assignment to U-space mapping, task interaction, and data logging. The two-participant example was therefore used to illustrate how the workflow operated in practice, rather than to provide statistical evidence of effectiveness. Several limitations remain. First, each participant completed only one session on a single day. It is not yet clear whether key variables used for U-space mapping, such as Pmax, Pcomf, and fatigue ratings, remain stable across days or repeated sessions. Future studies should include repeated sessions to distinguish day-to-day state changes from the usability of the mapping method itself. Second, the current task set is still prototype-oriented and assumes that participants can maintain a basic level of breath control. For users who cannot comfortably sustain a 10-second long tone, the sustained calibration steps and related tasks may need to be shortened or adapted. The crescendo–decrescendo task also raised a practical issue: one participant found it unintuitive to raise breath to a target level and then immediately ease off in a slow and controlled way. Future versions may therefore need simpler task formats and more gradual transitions. Third, the current grouping and boundary rules, including the B/F labels, coefficient choices, and fatigue-based boundary adjustments, should be understood as early pilot settings. As the study expands to larger samples, repeated sessions, and more playing-like tasks, these rules will need to be revisited, tested, and refined.

  Accessible Wind Instruments: Normalizing Breath Control Around Comfort · 2026 · DOI

• We aim to implement even more expressivity in our models, such as sustain pedal, pitch bend, and other expressive controls. Modeling these controls properly, however, requires precise tokenization. In particular, expressive controls can have very different densities than notes–pitch bend messages are much more dense, while sustain is much less dense. This proves challenging since a high density of controls can hinder the real-time performance of the model, while a sparse control can be difficult to model due to long dependencies extending outside of the context window. Additionally, while we have already experimented with diverse machine learning frameworks and optimization techniques, we believe that there are more methods, such as model distillation or more advanced quantization methods, that could continue to improve the model throughput. Finally, while using training data crafted by a single artist improves the output coherence and provides strong performance guarantees, it limits reuse by other performers and applications to different styles. Making the jam_bot more generalizable will require collecting more diverse data and developing fine-tuning strategies that broaden stylistic range without sacrificing the precise stylistic adherence that overfit models can provide.

  Enhancing Expressive Musical Conversation in the jam_bot · 2026 · DOI

• EyeLa can serve as an authentic musical instrument for individuals with limited cognitive and sensorimotor functioning. Through regular practice, users can develop instrumental skills, empowering them to make music both autonomously and collaboratively as musicians. It challenges established ableist practices in conventional music-making environments, creating opportunities for musicians with disabilities. Future directions for EyeLa include developing additional functionalities to support musical creativity and incorporating gaming elements to enable accessible gameplay in immersive environments [6].

  EyeLa · 2026 · DOI

• In the context of musical expression, the de- velopment of an interactive expressive sound instrument-system that uses artificial intelligence technology already available at- tempts to explore gestural interfaces to open questions about the causality of interactive systems from a critical perspective.

  Performing with the inclusive machine: An interdisciplinary roadmap for the design of AI collaborative musical instruments · 2026 · DOI

• Currently, Coypu uses a fixed metrical subdivision of 1/16th notes for all Sequencers; future work will introduce independent met- rical subdivisions per Sequencer. Coypu’s event scheduling has been refined to reduce jitter, and a wide range of methods for gen- erating rhythms and melodies has been implemented, including a Markov-chain-based jazz generator that follows a set of harmony constraints. Also, our current focus is on ensuring the stability of the system, refining the user interface, and implementing a few additional features, such as support for polyphony and cyclical structures in Coypu. The coverage of the Faust libraries currently stands at 70%, with plans to achieve full support by the end of the year. Once all the functions of Faust’s Physical Models library have been ported to Phausto, we plan to provide an extensive tutorial on building instruments based on this type of synthesis. We are also working on a more flexible and comprehensive MIDI implementation for controlling instruments created with Phausto; at present, multitimbrality and Control Change messages are not supported. In addition, we plan to implement offline audio analy- sis tools in Pharo, including FFT analysis and RMS detection, as well as the computation of spectral features such as the Centroid and Chroma from the magnitude spectrum. An internal oscillo- scope and spectrum analyzer for the real-time output of Phausto are also under development. Finally, we plan to provide a graphi- cal data-flow environment, inspired by Pure Data, for designing DSP by connecting Phausto’s Unit Generators. This will offer 26https://codeberg.org/musikinformatik/SuperDirt an alternative paradigm for users who prefer visualizing audio signal flow and will also support teaching and learning purposes.

  Coypu and Phausto: accessible live sound coding with Pharo · 2026 · DOI

• style, which may The dataset is relatively small and limited to a single dastgāh and metric restrict generalisability despite augmentation across tonal centres. In addition, the current system operates offline and does not yet support real-time interaction. Figure 4- Example of a generated melody In this example, 92.94% of the notes fall within the predefined allowable pitch set, while the remaining out-of-set notes, highlighted with red rectangles, represent deviations from the modal constraints. The excerpt also demonstrates a varied use of rhythmic values and pitch movement, indicating that the model produces musically diverse melodic material. Subjective evaluation: A perceptual evaluation was conducted using 10 randomly ordered 15-second audio excerpts, comprising 5 generated by PerFormer and 5 by a first-order Markov baseline. Two musicians with experience in Persian classical music participated in the study. Participants rated each sample on a 5-point Likert scale across four criteria: musical coherence (melodic structure), stylistic similarity to Persian classical music, pitch correctness (microtonality), and overall quality. Table 4 shows that PerFormer consistently outperforms the Markov baseline across all perceptual criteria, with the largest improvement observed in stylistic similarity and pitch correctness. Musical coherence improved from 2.2 to 3.9, indicating stronger structural increased from 1.4 to 4.4, organisation. Stylistic similarity demonstrating effective modelling of Persian musical characteristics. Pitch correctness also improved significantly (1.4 to 4.1), reflecting the model’s ability to capture microtonal relationships. Overall quality increased from 1.5 to 3.9, confirming that the generated outputs are perceptually more convincing.

  PerFormer: An AI-Driven Approach to Melody Generation in Microtonal Persian Music · 2026 · DOI

• This work has several limitations. Methodologically, it is a single- designer autoethnography supported by exhibition observation; it does not include a controlled comparison group, and it makes no claim that the findings generalise to all disabled designers or to all players. Reflexive thematic analysis is interpretive by design, and the designer’s positionality both enables the analysis (through embodied access to the design space) and constrains it (the same body cannot stand outside its own categories). The exhi- bition setting limits data collection: ambient noise, time-bounded curation slots, and the curator’s parallel programme reduced the number of structured conversations possible at most venues, and demographic statistics were unavailable from most festivals. Structured surveys were not deployed because they were judged intrusive in the exhibition flow, a defensible choice in context but one that trades depth for breadth. Technical limitations specific to individual versions have been reported alongside their iterations: at version 2 the Teensy USB failure at Feral Vector prevented full evaluation of touch and IMU together; version 4 is in active development and its haptic findings here are preliminary. Finally, the disability framing in this paper is grounded in my specific lived experience of EDS, chronic pain, and nerve damage. I have flagged where findings are tied to that specificity and where they invite extension; I do not treat one disabled designer’s perspective as representative of all crip practice.

  Crip Design for Collaborative Musical Interfaces: Iterative Development of Bot Party, a Touch-Based Sonic Game · 2026 · DOI

• 3 Limitations While our stress tests confirm server-side scalability to 2000+ connections and our largest live deployment reached 421 de- vices, the emergent acoustic properties of significantly larger arrays—aggregate sound pressure, spatial masking, and inter- device phase interactions—remain unexplored.

  Qianji: A Resilient Framework for Orchestrating "A Thousand Machines" in Distributed Performance · 2026 · DOI

• This paper presented a modular, cost-effective, and high-density loudspeaker array platform for embedded spatial audio applica- tions. By combining compact, daisy-chainable loudspeaker PCBs with an FPGA-based processing system, the proposed platform enables large channel counts and flexible array geometries within a portable and reproducible form factor. The modular design sup- ports rapid reconfiguration and adaptation to different use cases, making the platform well suited for spatial audio research, ex- perimentation, and artistic deployment. The current implementation is intentionally optimized for high spatial density and ease of integration, which imposes limi- tations on loudspeaker size and amplifier power. However, the use of a parametric PCB generation script makes it possible to accommodate larger loudspeakers. Applications requiring higher acoustic output would neces- sitate replacing the integrated amplifiers with external ampli- fication stages. Beyond PCB re-routing, this would introduce additional design constraints, including thermal management, power delivery, and mechanical integration, which constitute natural directions for future system extensions.

  Embedded, Modular, and Affordable High-Density Loudspeaker Arrays · 2026 · DOI

• The e-baton explores how conducting-style gestures can be recog- nised from a six-DoF IMU using a per-user retrainable classifier. Iterative prototyping suggested that feature sets extending be- yond raw acceleration, to include jerk and rotational data, were necessary to support the gesture vocabulary used in performance. Future work will include a controlled ablation to quantify the contribution of each feature group, and exploration of alternative communication protocols such as ESP-NOW or OSC over WiFi to address the BLE reliability limitations observed here. Following review and further improvements to the initial de- sign, the e-baton will be made open-source, enabling users to build their own e-baton and further experiment with the con- troller’s gestural capabilities.

  e-baton: Recognising Conducting Gestures with Machine Learning · 2026 · DOI

Most-cited papers in Music Technology and Sound Studies

• Learning metrics on spectrotemporal modulations reveals the perception of musical instrument timbre · Nature Human Behaviour · 2020 · 33 citations
• SoundScan and the consolidation of control in the popular music industry · Media Culture & Society · 1997 · 28 citations
• The Power of Consequential Product Sounds · Journal of Retailing · 2020 · 23 citations
• Cantus: Construction and evaluation of a software solution for real-time vocal music training and musical intonation assessment · Journal of Music Technology and Education · 2016 · 19 citations
• Making sound visible—a simple schlieren imaging setup for schools · Physics Education · 2021 · 12 citations
• Perspektiven einer Audiovisuellen Grounded-Theory-Methodologie und ihr Potenzial im Forschungsfeld »digitale Jugendkulturen« · psychosozial · 2020 · 9 citations
• Music to My Ears: A Material-semiotic Analysis of Fetal Heart Sounds in Midwifery Prenatal Care · Science Technology & Human Values · 2021 · 8 citations
• Appropriating Biosensors as Embodied Control Structures in Interactive Music Systems · Human Technology · 2021 · 6 citations
• The Soundlapse Project: Exploring Spatiotemporal Features of Wetland Soundscapes · Leonardo · 2021 · 6 citations
• Multiplicity, Audibility, and Musical Continuity · Dialogue · 2020 · 4 citations

Most recent work

• Exploring the harmonics of a heavy metal tone using ChatGPT and Phyphox: An AI-assisted classroom approach · The Physics Teacher · 2026
• ARGIRA B_v9: Spectro-Temporal Persistence in Sonified Visual Art — A High-Frequency Regime Transition Robust Across 72 Artworks · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Simulation-based deep learning for IoT-oriented music improvisation optimization · Discover Artificial Intelligence · 2026
• Multi-layer rhythmic modeling and interactive system for guzheng music style generation · Discover Artificial Intelligence · 2026
• The LADIM’s story or something over <i>One Thousand and One Nights</i> of an electroacoustic music laboratory in Venezuela in the 21st century. The electroacoustic music at <i>Universidad Simón Bolívar</i> · Organised Sound · 2026
• Shapes Are Frozen Wavelengths: Geometric Form, Acoustic Physics, and the Observer-Dependent Nature of Experience · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Input-Envelope-Output: Auditable Generative Music Rewards in Sensory-Sensitive Context · 2026
• A transformer-based intelligent online virtual conductor assistance system for choirs · Discover Artificial Intelligence · 2026
• Instability as a symphonic map for a family of examples · Nonlinear Analysis · 2026
• Benchmark study of pipe input impedance simulations and measurements for verification and validation in musical acoustics context · Acta Acustica · 2026

Related topics in Computer Science