Paper Archive

The increasing complexity and real-time processing demands of audio signals require optimized algorithms that utilize the computational power of Graphics Processing Units (GPUs). Existing Digital Signal Processing (DSP) libraries often do not provide the necessary efficiency and flexibility, particularly for integrating with Artificial Intelligence (AI) models. In response, we introduce TorchFX: a GPU-accelerated Python library for DSP, engineered to facilitate sophisticated audio signal processing. Built on the PyTorch framework, TorchFX offers an Object-Oriented interface similar to torchaudio but enhances functionality with a novel pipe operator for intuitive filter chaining. The library provides a comprehensive suite of Finite Impulse Response (FIR) and Infinite Impulse Response (IIR) filters, with a focus on multichannel audio, thereby facilitating the integration of DSP and AI-based approaches. Our benchmarking results demonstrate significant efficiency gains over traditional libraries like SciPy, particularly in multichannel contexts. While there are current limitations in GPU compatibility, ongoing developments promise broader support and real-time processing capabilities. TorchFX aims to become a useful tool for the community, contributing to innovation in GPU-accelerated DSP. TorchFX is publicly available on GitHub at https://github.com/matteospanio/torchfx.

Download

Antiderivative Antialiasing (ADAA), has become a pivotal method for reducing aliasing when dealing with nonlinear function at audio rate. However, its implementation requires analytical computation of the antiderivative of the nonlinear function, which in practical cases can be challenging without a symbolic solver. Moreover, when the nonlinear function is given by measurements it must be approximated to get a symbolic description. In this paper, we propose a simple approach to ADAA for practical applications that employs numerical integration of lookup tables (LUTs) to approximate the antiderivative. This method eliminates the need for closed-form solutions, streamlining the ADAA implementation process in industrial applications. We analyze the trade-offs of this approach, highlighting its computational efficiency and ease of implementation while discussing the potential impact of numerical integration errors on aliasing performance. Experiments are conducted with static nonlinearities (tanh, a simple wavefolder and the Buchla 259 wavefolding circuit) and a stateful nonlinear system (the diode clipper).

Download

Scattering delay networks (SDNs) provide a flexible and efficient framework for artificial reverberation and room acoustic modeling. In this work, we introduce a differentiable SDN, enabling gradient-based optimization of its parameters to better approximate the acoustics of real-world environments. By formulating key parameters such as scattering matrices and absorption filters as differentiable functions, we employ gradient descent to optimize an SDN based on a target room impulse response. Our approach minimizes discrepancies in perceptually relevant acoustic features, such as energy decay and frequency-dependent reverberation times. Experimental results demonstrate that the learned SDN configurations significantly improve the accuracy of synthetic reverberation, highlighting the potential of data-driven room acoustic modeling.

Download

Combining different room acoustic modelling methods could provide a better balance between perceptual plausibility and computational efficiency than using a single and potentially more computationally expensive model. In this work, a hybrid acoustic modelling system that integrates ray tracing (RT) with an advanced feedback delay network (FDN) is designed to generate perceptually plausible RIRs. A multiple stimuli with hidden reference and anchor (MUSHRA) test and a two-alternative-forced-choice (2AFC) discrimination task have been conducted to compare the proposed method against ground truth recordings and conventional RT-based approaches. The results show that the proposed system delivers robust performance in various scenarios, achieving highly plausible reverberation synthesis.

Download

Recent studies on classifying electric guitar effects have achieved high accuracy, particularly with deep learning techniques. However, these studies often rely on simplified datasets consisting mainly of single notes rather than realistic guitar recordings. Moreover, in the specific field of effect chain estimation, the literature tends to rely on large models, making them impractical for real-time or resource-constrained applications. In this work, we recorded realistic guitar performances using four different guitars and created three datasets by applying a chain of five effects with increasing complexity: (1) fixed order and parameters, (2) fixed order with randomly sampled parameters, and (3) random order and parameters. We also propose a novel Neural Architecture Search method aimed at discovering accurate yet compact convolutional neural network models to reduce power and memory consumption. We compared its performance to a basic random search strategy, showing that our custom Neural Architecture Search outperformed random search in identifying models that balance accuracy and complexity. We found that the number of convolutional and pooling layers becomes increasingly important as dataset complexity grows, while dense layers have less impact. Additionally, among the effects, tremolo was identified as the most challenging to classify.

Download

We present a Neural-Driven Multi-Band Processor (NDMP), a differentiable audio processing framework that augments a static sixband Parametric Equalizer (PEQ) with per-band dynamic range compression. We optimize this processor using neural inference for two tasks: Automatic Equalization (AutoEQ), which estimates tonal and dynamic corrections without a reference, and Production Style Transfer (NDMP-ST), which adapts the processing of an input signal to match the tonal and dynamic characteristics of a reference. We train NDMP using a self-supervised strategy, where the model learns to recover a clean signal from inputs degraded with randomly sampled NDMP parameters and gain adjustments. This setup eliminates the need for paired input–target data and enables end-to-end training with audio-domain loss functions. In the inference, AutoEQ enhances previously unseen inputs in a blind setting, while NDMP-ST performs style transfer by predicting taskspecific processing parameters. We evaluate our approach on the MUSDB18 dataset using both objective metrics (e.g., SI-SDR, PESQ, STFT loss) and a listening test. Our results show that NDMP consistently outperforms traditional PEQ and a PEQ+DRC (single-band) baseline, offering a robust neural framework for audio enhancement that combines learned spectral and dynamic control.

Download

Neural networks have become invaluable for general audio processing tasks, such as virtual analog modeling of nonlinear audio equipment. For sequence modeling tasks in particular, recurrent neural networks (RNNs) have gained widespread adoption in recent years. Their general applicability and effectiveness stems partly from their inherent nonlinearity, which makes them prone to aliasing. Recent work has explored mitigating aliasing by oversampling the network—an approach whose effectiveness is directly linked with the incurred computational costs. This work explores an alternative route by extending the antiderivative antialiasing technique to explicit, computable RNNs. Detailed applications to the Gated Recurrent Unit and Long Short-Term Memory cell are shown as case studies. The proposed technique is evaluated on multiple pre-trained guitar amplifier models, assessing its impact on the amount of aliasing and model tonality. The method is shown to reduce the models’ tendency to alias considerably across all considered sample rates while only affecting their tonality moderately, without requiring high oversampling factors. The results of this study can be used to improve sound quality in neural audio processing tasks that employ a suitable class of RNNs. Additional materials are provided in the accompanying webpage.

Download

Accurately estimating nonlinear audio effects without access to paired input-output signals remains a challenging problem. This work studies unsupervised probabilistic approaches for solving this task. We introduce a method, novel for this application, based on diffusion generative models for blind system identification, enabling the estimation of unknown nonlinear effects using blackand gray-box models. This study compares this method with a previously proposed adversarial approach, analyzing the performance of both methods under different parameterizations of the effect operator and varying lengths of available effected recordings. Through experiments on guitar distortion effects, we show that the diffusion-based approach provides more stable results and is less sensitive to data availability, while the adversarial approach is superior at estimating more pronounced distortion effects. Our findings contribute to the robust unsupervised blind estimation of audio effects, demonstrating the potential of diffusion models for system identification in music technology.

Download

This paper investigates the optimization of Truncated Backpropagation Through Time (TBPTT) for training neural networks in digital audio effect modeling, with a focus on dynamic range compression. The study evaluates key TBPTT hyperparameters – sequence number, batch size, and sequence length – and their influence on model performance. Using a convolutional-recurrent architecture, we conduct extensive experiments across datasets with and without conditioning by user controls. Results demonstrate that carefully tuning these parameters enhances model accuracy and training stability, while also reducing computational demands. Objective evaluations confirm improved performance with optimized settings, while subjective listening tests indicate that the revised TBPTT configuration maintains high perceptual quality.

Download

This study evaluates the performance of five objective audio quality metrics—PEAQ Basic, PEAQ Advanced, PEMO-Q, ViSQOL, and HAAQI —in the context of digital music production. Unlike previous comparisons, we focus on their suitability for production environments, an area currently underexplored in existing research. Twelve audio examples were tested using two evaluation types: an effectiveness test under progressively increasing degradations (hum, hiss, clipping, glitches) and a robustness test under fixed-level, randomly fluctuating degradations. In the effectiveness test, HAAQI, PEMO-Q, and PEAQ Basic effectively tracked degradation changes, while PEAQ Advanced failed consistently and ViSQOL showed low sensitivity to hum and glitches. In the robustness test, ViSQOL and HAAQI demonstrated the highest consistency, with average standard deviations of 0.004 and 0.007, respectively, followed by PEMO-Q (0.021), PEAQ Basic (0.057), and PEAQ Advanced (0.065). However, ViSQOL also showed low variability across audio examples, suggesting limited genre sensitivity. These findings highlight the strengths and limitations of each metric for music production, specifically quality measurement with compressed audio. The source code and dataset will be made publicly available upon publication.

Download