Download Playing cylinders of mechanical organs with an optical reader This study presents an experimental setup designed to read, by means of optical techniques, the music inscribed on automatic organ cylinders. We describe the acquisition principle based on images taken by a CCD linear camera, and the various digital signal processing techniques employed to retrieve the music from the images. Preliminary results show that this original method is a relevant choice, since on our test cylinder about 90 % of the notes are correctly identified, with only 14 % of false alarms. However, for realistic estimates of the actual music, some improvements are still necessary, both in the experimental setup and in the way individual note positions are converted into music.
Download State-Space Representation for Digital Waveguide Networks of Lossy Flared Acoustic Pipes This paper deals with digital waveguide modeling of wind instruments. It presents the application of state-space representations to the acoustic model of Webster-Lokshin. This acoustic model describes the propagation of longitudinal waves in axisymmetric acoustic pipes with a varying cross-section, visco-thermal losses at the walls, and without assuming planar or spherical waves. Moreover, three types of discontinuities of the shape can be taken into account (radius, slope and curvature), which can lead to a good fit of the original shape of pipe. The purpose of this work is to build low-cost digital simulations in the time domain, based on the Webster-Lokshin model. First, decomposing a resonator into independent elementary parts and isolating delay operators lead to a network of input/output systems and delays, of KellyLochbaum network type. Second, for a systematic assembling of elements, their state-space representations are derived in discrete time. Then, standard tools of automatic control are used to reduce the complexity of digital simulations in time domain. In order to validate the method, simulations are presented and compared with measurements.
Download Acoustic Modelling of a Convex Pipe Adapted for Digital Waveguide Simulation This work deals with the physical modelling of acoustic pipes for real-time simulation, using the “Digital Waveguide Network” approach and the horn equation. With this approach, a piece of pipe is represented by a two-port system with a loop which involves two delays for wave propagation, and some subsystems without internal delay. A well-known form of this system is the “Kelly-Lochbaum” framework, which allows the reduction of the computation complexity. We focus this work on the simulation of pipes with a convex profile. But, using the “Kelly-Lochbaum” framework with the horn equation, two problems occur: first, even if the outputs are bound, some substates have their values which diverge; second, there is an infinite number of such substates. The system is then unstable and cannot be simulated as such. The solution of this problem is obtained with two steps. First, we show that there is a simple standard form compatible with the “Waveguide” approach, for which there is an infinite number of solutions which preserve the input/output relations. Second, we look for one solution which guarantees the stability of the system and which makes easier the approximation in order to get a low-cost simulation.
Download Perceptual Linear Filters: Low-Order ARMA Approximation for Sound Synthesis This paper deals with the approximation of a given frequency response by a low-order linear ARMA filter (Auto-Regressive Moving Average). The aim of this work is the audio synthesis, then to improve the perceptual quality, a criterion based on human listening is defined and minimized. Two complementary approaches are proposed here for solving this non-linear and non-convex problem: first, a weighted version of the Iterative Prefiltering, second, an adaptation of the Gauss-Newton method. This algorithm is adapted to guarantee the causality/stability of the obtained filter, and eventually its minimum phase property. The benefit of the new method is illustrated and evaluated.
Download Granular analysis/synthesis of percussive drilling sounds This paper deals with the automatic and robust analysis, and the realistic and low-cost synthesis of percussive drilling like sounds. The two contributions are: a non-supervised removal of quasistationary background noise based on the Non-negative Matrix Factorization, and a granular method for analysis/synthesis of this drilling sounds. These two points are appropriate to the acoustical properties of percussive drilling sounds, and can be extended to other sounds with similar characteristics. The context of this work is the training of operators of working machines using simulators. Additionally, an implementation is explained.
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 Expressive Piano Performance Rendering from Unpaired Data Recent advances in data-driven expressive performance rendering have enabled automatic models to reproduce the characteristics and the variability of human performances of musical compositions. However, these models need to be trained with aligned pairs of scores and performances and they rely notably on score-specific markings, which limits their scope of application. This work tackles the piano performance rendering task in a low-informed setting by only considering the score note information and without aligned data. The proposed model relies on an adversarial training where the basic score notes properties are modified in order to reproduce the expressive qualities contained in a dataset of real performances. First results for unaligned score-to-performance rendering are presented through a conducted listening test. While the interpretation quality is not on par with highly-supervised methods and human renditions, our method shows promising results for transferring realistic expressivity into scores.