Download A Comparative Study of Kolmogorov-Arnold Networks and Multi-Layer Perceptrons for Virtual Analog Modeling in Wave Digital Filters
The design of Virtual Analog (VA) algorithms has traditionally been divided between white-box (physics-based) and black-box (data-driven) approaches. Recent work has shown that hybrid methods, combining physical modeling with neural networks, can effectively leverage the strengths of both paradigms. In particular, Wave Digital Filters (WDFs) can be coupled with Multi-Layer Perceptrons (MLPs) to model circuits with multiple nonlinearities in a fully explicit manner. In this paper, we present a comparative study investigating the use of Kolmogorov-Arnold Networks (KANs) for VA modeling within the WDF framework. Unlike MLPs, KANs shift the learning paradigm by parameterizing activation functions instead of relying exclusively on learned weight matrices, potentially enabling more compact representations. Results show that, for our case study, KANs achieve accuracy comparable to MLPs while requiring approximately 70% fewer parameters at the cost of increased computational complexity. These findings suggest that KANs may represent a promising alternative in scenarios where memory footprint is a primary constraint, such as embedded audio applications, or when target models feature numerous nonlinear elements.
Download Reservoir Computing: a powerful Framework for Nonlinear Audio Processing
This paper proposes reservoir computing as a general framework for nonlinear audio processing. Reservoir computing is a novel approach to recurrent neural network training with the advantage of a very simple and linear learning algorithm. It can in theory approximate arbitrary nonlinear dynamical systems with arbitrary precision, has an inherent temporal processing capability and is therefore well suited for many nonlinear audio processing problems. Always when nonlinear relationships are present in the data and time information is crucial, reservoir computing can be applied. Examples from three application areas are presented: nonlinear system identification of a tube amplifier emulator algorithm, nonlinear audio prediction, as necessary in a wireless transmission of audio where dropouts may occur, and automatic melody transcription out of a polyphonic audio stream, as one example from the big field of music information retrieval. Reservoir computing was able to outperform state-of-the-art alternative models in all studied tasks.
Download One Billion Audio Sounds From Gpu-Enabled Modular Synthesis
We release synth1B1, a multi-modal audio corpus consisting of 1 billion 4-second synthesized sounds, paired with the synthesis parameters used to generate them. The dataset is 100x larger than any audio dataset in the literature. We also introduce torchsynth, an open source modular synthesizer that generates the synth1B1 samples on-the-fly at 16200x faster than real-time (714MHz) on a single GPU. Finally, we release two new audio datasets: FM synth timbre and subtractive synth pitch. Using these datasets, we demonstrate new rank-based evaluation criteria for existing audio representations. Finally, we propose a novel approach to synthesizer hyperparameter optimization.
Download Real-Time Implementation of the Dynamic Stiff String Using Finite-Difference Time-Domain Methods and the Dynamic Grid
Digital musical instruments based on physical modelling have gained increased popularity over the past years. This is partly due to recent advances in computational power, which allow for their real-time implementation. One of the great potentials for digital musical instruments based on physical models, is that one can go beyond what is physically possible and change properties of the instruments which are static in real life. This paper presents a real-time implementation of the dynamic stiff string using finitedifference time-domain (FDTD) methods. The defining parameters of the string can be varied in real time and change the underlying grid that these methods rely on based on the recently developed dynamic grid method. For most settings, parameter changes are nearly instantaneous and do not cause noticeable artefacts due to changes in the grid. A reliable way to prevent artefacts for all settings is under development.
Download Upcylcing Android Phones into Embedded Audio Platforms
There are millions of sophisticated Android phones in the world that get disposed of at a very high rate due to consumerism. Their computational power and built-in features, instead of being wasted when discarded, could be repurposed for creative applications such as musical instruments and interactive audio installations. However, audio programming on Android is complicated and comes with restrictions that heavily impact performance. To address this issue, we present LDSP, an open-source environment that can be used to easily upcycle Android phones into embedded platforms optimized for audio synthesis and processing. We conducted a benchmark study to compare the number of oscillators that can be run in parallel on LDSP with an equivalent audio app designed according to modern Android standards. Our study tested six phones ranging from 2014 to 2018 and running different Android versions. The results consistently demonstrate that LDSP provides a significant boost in performance, with some cases showing an increase of more than double, making even very old phones suitable for fairly advanced audio applications.
Download Neural Modeling of Magnetic Tape Recorders
The sound of magnetic recording media, such as open-reel and cassette tape recorders, is still sought after by today’s sound practitioners due to the imperfections embedded in the physics of the magnetic recording process. This paper proposes a method for digitally emulating this character using neural networks. The signal chain of the proposed system consists of three main components: the hysteretic nonlinearity and filtering jointly produced by the magnetic recording process as well as the record and playback amplifiers, the fluctuating delay originating from the tape transport, and the combined additive noise component from various electromagnetic origins. In our approach, the hysteretic nonlinear block is modeled using a recurrent neural network, while the delay trajectories and the noise component are generated using separate diffusion models, which employ U-net deep convolutional neural networks. According to the conducted objective evaluation, the proposed architecture faithfully captures the character of the magnetic tape recorder. The results of this study can be used to construct virtual replicas of vintage sound recording devices with applications in music production and audio antiquing tasks.
Download Distributed Single-Reed Modeling Based on Energy Quadratization and Approximate Modal Expansion
Recently, energy quadratization and modal expansion have become popular methods for developing efficient physics-based sound synthesis algorithms. These methods have been primarily used to derive explicit schemes modeling the collision between a string and a fixed barrier. In this paper, these techniques are applied to a similar problem: modeling a distributed mouthpiece lay-reed-lip interaction in a woodwind instrument. The proposed model aims to provide a more accurate representation of how a musician’s embouchure affects the reed’s dynamics. The mouthpiece and lip are modeled as distributed static and dynamic viscoelastic barriers, respectively. The reed is modeled using an approximate modal expansion derived via the Rayleigh-Ritz method. The reed system is then acoustically coupled to a measured input impedance response of a saxophone. Numerical experiments are presented.
Download Physical Model of the Chinese Yehu for Sound Synthesis
The yehu is a Chinese bowed string instrument featuring a resonator carved from a coconut shell, a seashell-based bridge, and two silk strings. This paper proposes a physical model of the yehu and reports on simulations using a finite-difference scheme with measurement-based physical characterization. The proposed model consists of two stiff strings coupled at the bridge, a bow with elastic bow hairs, a stopping finger, and a modal model of the bridge. A non-iterative solver based on energy quadratization is used to model the finger–string contact force, while an iterative solver is used for elasto-plastic bow-string friction force. The bridge-body model is based on a modal characterization obtained from the measured bridge admittance. The measured radiation transfer function is represented as a bank of parallel second-order filters and is applied to the simulated bridge force to incorporate body radiation characteristics. Finally, computational performance tests are conducted, showing that the proposed model is capable of real-time computation.
Download A Digital Model of the Buchla Lowpass-Gate
In recent years there has been an increasing amount of interest in the style of synthesis implemented by Don Buchla in his instrument designs from the early 1960s until the present. A key part of the Buchla synthesizer and its characteristic quality is the ’lowpass gate’ filter and the acoustic-like plucked sounds that it provides. In this work we examine the circuit of the low-pass gate, both its audio and control portions. We propose a number of digital models of these circuits, as well as a model of the photoresistive optoisolator or ’vactrol’ used within them. In the case of the audio path of the device, we pay particular attention to maintaining desirable behavior under time-variation of its parameters. The resulting digital model retains much of the interesting character of the analog system, and is computationally cheap enough to use within a standard computer-music setup.
Download Design principles for lumped model discretisation using Möbius transforms
Computational modelling of audio systems commonly involves discretising lumped models. The properties of common discretisation schemes are typically derived through analysis of how the imaginary axis on the Laplace-transform s-plane maps onto the Ztransform z-plane and the implied stability regions. This analysis ignores some important considerations regarding the mapping of individual poles, in particular the case of highly-damped poles. In this paper, we analyse the properties of an extended class of discretisations based on Möbius transforms, both as mappings and discretisation schemes. We analyse and extend the concept of frequency warping, well-known in the context of the bilinear transform, and we characterise the relationship between the damping and frequencies of poles in the s- and z-planes. We present and analyse several design criteria (damping monotonicity, stability) corresponding to desirable properties of the discretised system. Satisfying these criteria involves selecting appropriate transforms based on the pole structure of the system on the s-plane. These theoretical developments are finally illustrated on a diode clipper nonlinear model.