Paper Archive

Recent advances in AI-assisted coding tools raise questions about how programming-intensive subjects such as audio digital signal processing should be taught and how exam projects should be evaluated. This paper presents a small-scale controlled exploratory study conducted in a graduate course on music DSP. As a final project at the end of the course, the students implemented a modular synthesizer plugin in C++. Half of the students had access to AI-assisted coding support, while the other group developed the plugin manually. All students had to follow a protocol and provide data at the end of the project, together with their code, which was discussed with them as part of the course exam. Although the scale of the study is small, the paper shares a qualitative analysis of the results and a few takeaway messages for future reference among lecturers in the field. Overall, AI-assisted coding does provide some advantage to students but only in certain regards. The used AI tools, trained on GitHub repositories, seem to have only partial awareness of the state of the art in digital audio processing (e.g. antialiasing oscillators, virtual analog filters, etc.). Finally, the use of AI seems to not interfere excessively with the ability of the students to learn from their practical experience.

Download

This paper proposes POLAR, a framework for the optimisation of loudspeaker signals using end-to-end differentiable perceptual loss functions. The framework optimises multiple perceptual attributes across multiple listeners, offering a versatile method for a range of problems. This versatility stems from the ability to customise the number of loudspeakers, listeners, and the weighting applied to different perceptual attributes. Here, we apply the method to four problems: (a) source panning for a single listener in stereo reproduction, (b) single-listener colouration matching in stereo reproduction, (c) extended sweet spot using stereo pairs beamforming, and (d) multi-attribute perceptually driven panning in stereo reproduction. The first three problems are evaluated against solutions traditionally used for these tasks: solutions of (a) are shown to be similar to those obtained with tangent panning law and vector-base amplitude panning (VBAP), solutions of (b) are shown to be similar to those obtained for cross-talk cancellation, and solutions of (c) are shown to be similar to those obtained in earlier work on directivity pattern optimisation for sweet spot widening. Each of these solutions was previously obtained using fundamentally different methodologies, demonstrating the flexibility and broad applicability of the proposed framework.

Download

This work extends our previously proposed Quadratic Programming (QP) approach for the emulation of nonlinear analog audio circuits by formalizing its main numerical ingredients and introducing a residual-driven adaptive multi-rate scheme. Starting from a state-space Differential Algebraic System of Equations (DAE) formulation, the nonlinear algebraic circuit device relations are replaced inside the QP by a first-order surrogate linear constraint, and the post-step nonlinear residual is shown to act as a valid defect indicator for adaptive step-size control. This yields a single-step simulation procedure that avoids the usual combination of nonlinear iterative solves and separate integration updates. The method is evaluated on a diode clipper, a BJT common-emitter amplifier, and a Colpitts oscillator, using SPICE as a baseline reference. The results show that adaptive step sizing considerably improves agreement with the reference solution, that the pseudo-inverse implementation is essentially equivalent to the full equality-constrained QP in the tested cases, and that the proposed formulation remains effective beyond the baseline clipper example, including for a self-oscillating circuit. These results position the proposed method as a promising bridge between SPICE-like interpretability and the efficiency demands of virtual analog (VA) audio applications.

Download

We introduce FoleySet, a human-annotated Foley sound dataset designed to support research on sound-event understanding and Foley sound generation. The dataset provides annotations at multiple levels of granularity, capturing both broad event categories and more detailed semantic or production-related attributes. This multi-level structure supports tasks such as classification, retrieval, captioning, and controllable generation. FoleySet is intended to address the limited availability of systematically annotated Foley material and to provide a common resource for evaluating models across different levels of semantic detail.

Download

Physical modeling sound synthesis has achieved remarkable success in terms of its fidelity to reality. In many cases, since modeling of the physical system is performed on the sounding objects that already exist in the real world, observation precedes the model itself. Departing from this convention, this paper aims to physically model the acoustic characteristics of objects that do not necessarily exist in reality. Specifically, we study wave propagation on compact two-dimensional (2D) manifolds that are non-orientable surfaces, such as the Klein bottle that cannot be embedded in three-dimensional Euclidean space without self-intersection. We derive closed-form expressions for the eigenfrequencies and mode shapes of non-orientable 2D topologies and study their acoustic characteristics. The modal structures are verified through comparison with finite-difference time-domain simulations. The results demonstrate how the topological character formed by the boundaries influences the acoustic resonances, and how the quotient-space framework provides a practical route to reverb synthesis on geometries with no physical counterpart.

Download