Download Dark Velvet Noise This paper proposes dark velvet noise (DVN) as an extension of the original velvet noise with a lowpass spectrum. The lowpass spectrum is achieved by allowing each pulse in the sparse sequence to have a randomized pulse width. The cutoff frequency is controlled by the density of the sequence. The modulated pulse-width can be implemented efficiently utilizing a discrete set of recursive running-sum filters, one for each unique pulse width. DVN may be used in reverberation algorithms. Typical room reverberation has a frequency-dependent decay, where the high frequencies decay faster than the low ones. A similar effect is achieved by lowering the density and increasing the pulse-width of DVN in time, thereby making the DVN suitable for artificial reverberation.
Download Higher-Order Scattering Delay Networksfor Artificial Reverberation Computer simulations of room acoustics suffer from an efficiency vs accuracy trade-off, with highly accurate wave-based models being highly computationally expensive, and delay-network-based models lacking in physical accuracy. The Scattering Delay Network (SDN) is a highly efficient recursive structure that renders first order reflections exactly while approximating higher order ones. With the purpose of improving the accuracy of SDNs, in this paper, several variations on SDNs are investigated, including appropriate node placement for exact modeling of higher order reflections, redesigned scattering matrices for physically-motivated scattering, and pruned network connections for reduced computational complexity. The results of these variations are compared to state-of-the-art geometric acoustic models for different shoebox room simulations. Objective measures (Normalized Echo Densities (NEDs) and Energy Decay Curves (EDCs)) showed a close match between the proposed methods and the references. A formal listening test was carried out to evaluate differences in perceived naturalness of the synthesized Room Impulse Responses. Results show that increasing SDNs’ order and adding directional scattering in a fully-connected network improves perceived naturalness, and higher-order pruned networks give similar performance at a much lower computational cost.
Download Multichannel Interleaved Velvet Noise The cross-correlation of multichannel reverberation generated using interleaved velvet noise is studied. The interleaved velvetnoise reverberator was proposed recently for synthesizing the late reverb of an acoustic space. In addition to providing a computationally efficient structure and a perceptually smooth response, the interleaving method allows combining its independent branch outputs in different permutations, which are all equally smooth and flutter-free. For instance, a four-branch output can be combined in 4! or 24 ways. Additionally, each branch output set is mixed orthogonally, which increases the number of permutations from M ! to M 2 !, since sign inversions are taken along. Using specific matrices for this operation, which change the sign of velvet-noise sequences, decreases the correlation of some of the combinations. This paper shows that many selections of permutations offer a set of well decorrelated output channels, which produce a diffuse and colorless sound field, which is validated with spatial variation. The results of this work can be applied in the design of computationally efficient multichannel reverberators.
Download Pyroadacoustics: A Road Acoustics Simulator Based on Variable Length Delay Lines In the development of algorithms for sound source detection, identification and localization, having the possibility to generate datasets in a flexible and fast way is of utmost importance. However, most of the available acoustic simulators used for this purpose target indoor applications, and their usefulness is limited when it comes to outdoor environments such as that of a road, involving fast moving sources and long distances travelled by the sound waves. In this paper we present an acoustic propagation simulator specifically designed for road scenarios. In particular, the proposed Python software package enables to simulate the observed sound resulting from a source moving on an arbitrary trajectory relative to the observer, exploiting variable length delay lines to implement sound propagation and Doppler effect. An acoustic model of the road reflection and air absorption properties has been designed and implemented using digital FIR filters. The architecture of the proposed software is flexible and open to extensions, allowing the package to kick-start the implementation of further outdoor acoustic simulation scenarios.
Download A Study of Control Methods for Percussive Sound Synthesis Based on Gans The process of creating drum sounds has seen significant evolution in the past decades. The development of analogue drum synthesizers, such as the TR-808, and modern sound design tools in Digital Audio Workstations led to a variety of drum timbres that defined entire musical genres. Recently, drum synthesis research has been revived with a new focus on training generative neural networks to create drum sounds. Different interfaces have previously been proposed to control the generative process, from low-level latent space navigation to high-level semantic feature parameterisation, but no comprehensive analysis has been presented to evaluate how each approach relates to the creative process. We aim to evaluate how different interfaces support creative control over drum generation by conducting a user study based on the Creative Support Index. We experiment with both a supervised method that decodes semantic latent space directions and an unsupervised Closed-Form Factorization approach from computer vision literature to parameterise the generation process and demonstrate that the latter is the preferred means to control a drum synthesizer based on the StyleGAN2 network architecture.
Download Differentiable Piano Model for Midi-to-Audio Performance Synthesis Recent neural-based synthesis models have achieved impressive results for musical instrument sound generation. In particular, the Differentiable Digital Signal Processing (DDSP) framework enables the usage of spectral modeling analysis and synthesis techniques in fully differentiable architectures. Yet currently, it has only been used for modeling monophonic instruments. Leveraging the interpretability and modularity of this framework, the present work introduces a polyphonic differentiable model for piano sound synthesis, conditioned on Musical Instrument Digital Interface (MIDI) inputs. The model architecture is motivated by high-level acoustic modeling knowledge of the instrument which, in tandem with the sound structure priors inherent to the DDSP components, makes for a lightweight, interpretable and realistic sounding piano model. The proposed model has been evaluated in a listening test, demonstrating improved sound quality compared to a benchmark neural-based piano model, with significantly less parameters and even with reduced training data. The same listening test indicates that physical-modeling-based models still achieve better quality, but the differentiability of our lightened approach encourages its usage in other musical tasks dealing with polyphonic audio and symbolic data.
Download Realistic Gramophone Noise Synthesis Using a Diffusion Model This paper introduces a novel data-driven strategy for synthesizing gramophone noise audio textures. A diffusion probabilistic model is applied to generate highly realistic quasiperiodic noises. The proposed model is designed to generate samples of length equal to one disk revolution, but a method to generate plausible periodic variations between revolutions is also proposed. A guided approach is also applied as a conditioning method, where an audio signal generated with manually-tuned signal processing is refined via reverse diffusion to improve realism. The method has been evaluated in a subjective listening test, in which the participants were often unable to recognize the synthesized signals from the real ones. The synthetic noises produced with the best proposed unconditional method are statistically indistinguishable from real noise recordings. This work shows the potential of diffusion models for highly realistic audio synthesis tasks.
Download A Direct Microdynamics Adjusting Processor with Matching Paradigm and Differentiable Implementation In this paper, we propose a new processor capable of directly changing the microdynamics of an audio signal primarily via a single dedicated user-facing parameter. The novelty of our processor is that it has built into it a measure of relative level, a short-term signal strength measurement which is robust to changes in signal macrodynamics. Consequent dynamic range processing is signal level-independent in its nature, and attempts to directly alter its observed relative level measurements. The inclusion of such a meter within our proposed processor also gives rise to a natural solution to the dynamics matching problem, where we attempt to transfer the microdynamic characteristics of one audio recording to another by means of estimating appropriate settings for the processor. We suggest a means of providing a reasonable initial guess for processor settings, followed by an efficient iterative algorithm to refine upon our estimates. Additionally, we implement the processor as a differentiable recurrent layer and show its effectiveness when wrapped around a gradient descent optimizer within a deep learning framework. Moreover, we illustrate that the proposed processor has more favorable gradient characteristics relative to a conventional dynamic range compressor. Throughout, we consider extensions of the processor, matching algorithm, and differentiable implementation for the multiband case.
Download A Comparison of Deep Learning Inference Engines for Embedded Real-Time Audio Classification Recent advancements in deep learning have shown great potential for audio applications, improving the accuracy of previous solutions for tasks such as music transcription, beat detection, and real-time audio processing. In addition, the availability of increasingly powerful embedded computers has led many deep learning framework developers to devise software optimized to run pretrained models in resource-constrained contexts. As a result, the use of deep learning on embedded devices and audio plugins has become more widespread. However, confusion has been rising around deep learning inference engines, regarding which of these can run in real-time and which are less resource-hungry. In this paper, we present a comparison of four available deep learning inference engines for real-time audio classification on the CPU of an embedded single-board computer: TensorFlow Lite, TorchScript, ONNX Runtime, and RTNeural. Results show that all inference engines can execute neural network models in real-time with appropriate code practices, but execution time varies between engines and models. Most importantly, we found that most of the less-specialized engines offer great flexibility and can be used effectively for real-time audio classification, with slightly better results than a real-time-specific approach. In contrast, more specialized solutions can offer a lightweight and minimalist alternative where less flexibility is needed.
Download Improved Automatic Instrumentation Role Classification and Loop Activation Transcription Many electronic music (EM) genres are composed through the activation of short audio recordings of instruments designed for seamless repetition—or loops. In this work, loops of key structural groups such as bass, percussive or melodic elements are labelled by the role they occupy in a piece of music through the task of automatic instrumentation role classification (AIRC). Such labels assist EM producers in the identification of compatible loops in large unstructured audio databases. While human annotation is often laborious, automatic classification allows for fast and scalable generation of these labels. We experiment with several deeplearning architectures and propose a data augmentation method for improving multi-label representation to balance classes within the Freesound Loop Dataset. To improve the classification accuracy of the architectures, we also evaluate different pooling operations. Results indicate that in combination with the data augmentation and pooling strategies, the proposed system achieves state-of-theart performance for AIRC. Additionally, we demonstrate how our proposed AIRC method is useful for analysing the structure of EM compositions through loop activation transcription.