ABSTRACT. The Balinese metallophone gender dasa consists of bronze plates of trapezoid shape, suspended over a bamboo resonator. This trapezoid shape has been shown to enhance the brightness of the plate considerably (R. Bader: Additional modes in a Balinese gender plate due to its trapezoid shape. 2009). Using laser interferometry, the transient traveling waves after a strike on the instrument as well as the mode shapes over time are measured. Next to expected modes, many additional complex modes shapes, degenerated modes, as well as asymmetric energy distribution on the plate are examined during the initial transient of about 100 ms. Modes shapes up to 20 kHz could be identified. This plate asymmetry seems to be the reason for the enhanced brightness of the instrument and therefore intendedly crafted by instrument builders.

ABSTRACT. This work seeks to enable researchers and makers to evaluate new idiophone designs based solely on geometry and material properties. Designs are evaluated by listening to virtual instruments, before investing in a physical prototype, and potentially even before purchasing materials.

Finite element analysis is a well-established and widely popular numerical method for analyzing physical phenomena and other mathematical problems. This method has a great capacity to simulate dynamic behaviour relevant to idiophones – including impact, contact mechanics and vibration. It can even capture nonlinear behaviour, where appropriate. Finite element models are defined based on the geometry and material properties of the objects under study.

The drawback of such versatility is computational cost. Finite element simulations of dynamic interaction and response in the time domain (so-called dynamic time-history models) can require vast amounts of computer memory, and may take days to run even on modern computers.

Modal models, by contrast, are computationally very efficient. Such models represent an instrument’s response to excitation as the sum of a limited number of modes, corresponding to natural frequencies of the instrument. Once a number of parameters are tuned for each mode, the modal model’s response over time is easily computed. It is common for recordings of existing instruments or experimental measurements to be used to fit these modal parameters. An instrument’s response must remain linear to be accurately represented as a combination of modes.

This work leverages the advantages of finite element time domain models and modal models to develop a hybrid model for the sound radiation of idiophones. Instrument excitation is modelled by finite element analysis, using only geometrical and material inputs, and accounting for any nonlinear interactions. Subsequent sound radiation is represented as a modal model, once the system has settled into a linear response. Specific knowledge of the mathematical problem is used to minimize the extent of the finite element model, reducing run times from days to less than an hour. Modal parameters are tuned based on finite element results, avoiding the need for recordings or experimental measurements.

Examples based on marimba and vibraphone bars will be provided. The work reported here partners with a previously reported approach for tuning idiophone bars via finite element analysis. Combining the two enables design and evaluation of bars within a single computational framework.

ABSTRACT. The Cristal Baschet is a musical instrument created during the 50’s by Bernard and Francois Baschet. It is composed of a large number of glass rods arranged in chromatic scale. The sound produced results of vibration induced by friction between wet fingers and glass rods. Each glass rod is connected to an assembly of threated shaft and mass. Mechanical properties of this assembly determine the pitch of the note. Then vibrations are transmitted to large metal panels or cones that act as radiating elements.

The manufacturing and tuning of this instrument is based on empirical knowledge and involves many parameters whose effects are not clearly understood. One of the encountered problems is the difficulty to produce sound in the high register of the instrument. In attempt to understand the influences of this parameters on sound and playability, a minimal physical model of the Cristal Baschet is developed. Thus, it will be possible to propose design rules to improve the playability of the instrument.

The minimal physical model focuses on the interaction between finger and isolated resonator. The resonator is an assembly of threaded shafts, mass and glass rod. Its dynamics behavior is described by a set of modes. The modals parameters can be obtained from finite element model or from experimental modal analysis of the instrument. The musician’s gesture is described by two control parameters: the velocity of the finger along the glass rod and the normal force applied by the finger on the rod. To describe the interaction between the finger and the resonator, a friction law is implemented. The influences of different parameters is studied by means of linear stability analysis and time-domain simulations. Specific criteria based on this two analyses are developed to highlight the role of design parameters on playability. The possibility to extend the physical model by taking into account one of the radiating elements, the large thin panel, is investigated and discussed.

ABSTRACT. Background SARS-CoV-2 is primarily spread by airborne transmission via aerosols and respiratory droplets [1][2]. As with singing, wind instruments are under heavy scrutiny during the ongoing COVID-19-pandemic. Several studies investigated droplet and aerosol emissions by wind instruments [3][4]. But since wind instruments don’t actually produce any “wind”, they also don’t blow out droplets. However, wind instruments still produce aerosols [3][5]. As counter measure, filters to be placed over the bell of brass instruments were proposed [5].

Questions This study investigated whether such filters have a noteworthy impact on the timbre of the filtered instrument, and whether the filters affect the radiation patterns of the same instruments.

Methods Four brass and one woodwind instrument (flute, French horn, trombone, trumpet, tuba) were tested in three registers (low/mid/high), each at two dynamic levels (pp/ff), and in three filter conditions (0/1/2 layers). The filters were sewn from molton fabric and provided with an elastic hem for easy use. The stimuli were played by musicians and recorded on stage to approximate concert conditions as closely as possible. The radiation patterns of the instruments were captured with an acoustic camera. A group of naive subjects rated the sound (timbre), intonation, and articulation of each stimulus. The ratings were compared by means of a paired t-test. Low-level audio features were extracted and analyzed to identify the acoustic properties affected by the filter.

Results One layer of filter had virtually no influence on the subjective ratings which is also confirmed by the spectral low-level features. With two layers the difference to the zero-filter-condition was significant. Overall, the perceived impact on intonation was considerably greater than on timbre. Differences in the radiation patterns were not found.

References [1] Meselson, M. (2020). Droplets and Aerosols in the Transmission of SARS-CoV-2. In: NEJM 382(21), 2063.

[2] Wiersinga, J. et al. (2020). Pathophysiology, Transmission, Diagnosis, and Treatment of Coronavirus Disease 2019 (COVID-19): A Review. In: JAMA 324(8), 782–793.

[3] Kähler, C. & Hain, R. (2020). Musizieren während der Pandemie – was rät die Wissenschaft? Über Infektionsrisiken beim Chorsingen und Musizieren mit Blasinstrumenten. Research report, Universität der Bundeswehr München.

[4] Gantner, S. et al. (2022). Impulse dispersion of aerosols during playing wind instruments. In: PLOS ONE 17(3), e0262994.

[5] Becher, L. et al. (2020). Einsatz von Filtern zur Reduktion der Ausbreitung der Atemluft beim Spielen von Blasinstrumenten und beim Singen während der COVID-19 Pandemie. Research report, Bauhaus University Weimar.

ABSTRACT. Modelling wind instrument, and studying these models using analysis methods for dynamical systems, allows extraction of quantitative information related to the behaviour of the instrument. These outputs are particularly relevant for the comparison of instruments with relatively close characteristics (bore shape), on the basis of descriptors related to the performance (“operational objective criteria”). Nevertheless, the question of the validity of these numerical results arise, in perspective with the musicians’ blowing feeling. This question is not trivial because it requires evaluation protocols to be developed, allowing the perceptual information to be collected, whilst this information is associated to physical quantities that musicians are usually not familiar with. In this work, we present the first results of playing tests conducted with professional trumpet players, in order to collect perceptual evaluations connected to numerical quantities calculated on different trumpets by numerical continuation (oscillation thresholds, hysteresis phenomena, variation of excitation parameters). We will discuss the strategies and protocols elaborated, the bias and difficulties identified, and we will present the results of this first series of tests, through comparisons with the numerical predictions.

ABSTRACT. In this paper we present recent work on the development of a physical model of the hand harmonium, a bellow driven free reed instrument popular in South Asian music. Western free reed instruments like the accordion, harmonica, reed organ, etc. share many similarities with the hand harmonium in their physical structure and sound timbre. Previous models and experimental works on western free reed instruments are studied and revisions are proposed to suitably match the physical setup and sound timbre of the hand harmonium.

A simplified physical model of the hand harmonium was created based on the minimal model of free reeds described by Millot and Baumann (2007). The Millot and Baumann model describes a harmonica system that is driven by a jet of air blown by the performer, which indirectly determines the upstream pressure at the reed, while the downstream is assumed to be exposed to atmospheric pressure. In contrast, the harmonium model is driven by the pressure generated in the upstream enclosure using a bellows and the air flow through the reed is controlled by the extent of key press which controls the downstream area for the air to escape. The upstream and downstream chamber resonances play a significant role in determining the sound timbre even though the fundamental frequency is practically equal to the principal eigenfrequency of the reed.

Experimental studies are under way to determine the model parameters such as reed chamber pressure and effective area for airflow around the reed. In particular, we are measuring the vibrations of the free reed when mounted on a mechanical blowing system using a laser doppler vibrometer to understand the relation between blowing pressure and the effective area for airflow. Work is in progress to measure the driving bellows pressure using pressure sensors placed inside an actual instrument, both to better understand the pressure range for the model and also to operate the mechanical blower system in a similar pressure range.

The ultimate goal of this research is to understand the critical parameters influencing the acoustic behaviour of free reed instruments like the hand harmonium and to develop a natural sounding real time physics-based synthesis system that can add additional affordances like pitch glides and microtonality which cannot be achieved in a real harmonium.

ABSTRACT. Physics-based spring reverb emulation has seen steady advances in the literature over approximately the past decade, with a recent focus on improved numerical simulation of helical spring vibrations. The input and output mechanisms of tanks have not yet seen extensive research, and as such a simplified driving and pickup model is commonly employed in simulation algorithms.

To help advance the physics-based models, this study investigates the underlying physics behind the driving mechanism, the design of which is based on the principles of electromagnetic induction. At the input of a tank, an alternating current is passed through a coil of wire wrapped around a small section of U-shaped stacked metal laminations. The alternating magnetic field induced then influences a small cylindrical magnet suspended in the air gap of the U-shape, which is in turn attached to the helical spring, thereby inducing mechanical waves along the spring corresponding to the electrical input signal.

We study how the electromagnetic field drives the magnetic bead here through experimental means. It is generally thought that the bead drives the spring through torsional motion (i.e., twisting), but specific aspects still require attention, including if the magnetic field can excite the bead in other polarisations and how the torsion transfer function amplitude varies across the audio frequency range.

Using sine-sweeps and high-speed video recordings, two polarisations of motion evident in the bead are tracked and analysed up to 5000 Hz. These are torsional and transverse (in the vertical plane) and are due to the input signal and the superposition of reflected waves along the spring, where the latter can be mitigated by adding damping along the length of the spring. An accompanying video compares the damped and undamped case for a lower frequency range and indicates that the transverse motion of the bead is due to the reflected wave propagations and the alternating magnetic field thus likely induces only torsion in the bead.

ABSTRACT. Transient energy distribution on soundboards is related to the acoustic salience of individual soundboard parts in instrument building, as well as to sound perception parameters like spaciousness, sound localization, or attack. As shown before with the piano (Plath, N.: From workshop to concert hall: Acoustic observations on a grand piano under construction, 2019), parts like bridges or ribs can act as waveguides or energy resources during the transient as well as the quasi-steady-state of the instrument sound. To estimate these parameters, a harpsichord is excited by an automated miniature impact hammer at five different positions at the bridge. The time-dependent velocity responses of the soundboard are recorded using a laser Doppler vibrometer. The detected bending waves show a propagation mainly along the bridge. Furthermore, during the first 50ms, which are crucial for human sound localization, most of the vibrational energy is located around the excitation point. While playing the harpsichord, different strings continuously excite the soundboard at different positions at the bridge, which may suppress the perception of a fusion of musical chords and allow for a more separated perception of polyphonic lines, present and important in Baroque or Renaissance music, e.g., in fugues.

ABSTRACT. The current work is presented in the frame of a series of efforts to study the directivity of pipe organs. The radiation patterns of musical instruments are potentially necessary for simulating musical instruments. Although many instruments can be modeled as point sources, organs are large complex instruments. While there is great detail to be found in the last several decades towards describing elements of the pipe operation, little is said about the scattered field inside the organ’s enclosing cavity (the buffet) or the radiated field through the organ façade into the room. The focus in the present work is toward characterizing the radiation pattern of the pipes within the buffet.

In previous work, the organ buffet was simulated with electroacoustic sources. The benefit of this approach is the precision and repeatability of the measurement. In the current work, the organ buffet was excited by natural means: namely with individual, isolated pipes within the organ buffet as well as clusters of simultaneously excited pipes to create a wideband, noise-like stimulus. Using the organ pipe sources excite the organ façade in a more natural way than with electroacoustic sources.

The organ chosen for this study was the positive section of Sainte Élisabeth d’Hongrie, which embodies the simplest buffet geometry from up-to-date church tribune organs. The relatively simple arrangement of the organ made it feasible to compare to laboratory measurements. Additionally, in this church, the positive radiates in free-field into the nave, making measurements feasible.

The measurement method, common with other measurements carried out by the same team during the period 2021-2022, consisted in positioning microphones along a line of points parallel to the façade. In the current case, this was done by hanging microphones up in the air and sliding them for each set of acquisitions. Additional reference microphones were placed within the inside of the cavity and inside the resonator of several of the excited pipes.

It was shown elsewhere in previous work on organ radiation of this same team that the cavity and the façade will convey differences in the radiated response for different windowed instances. The same principle is thus followed here with the aid of the microphones inserted in the pipe’s resonators: they help estimating the formation and propagation times of the scattered and transmitted fields inside the buffet and in the church respectively.

ABSTRACT. Most flue organ pipes are either cylindrical and made of metal or rectangular and made of wood. It is often assumed that the shape is not an important parameter provided the cross sectional dimensions are small compared to the wavelength of the sound, however, recent work has indicated that the pipe shape may be important in determining the resulting sound. The flow through transparent organ pipes has been imaged using high speed transmission electronic speckle pattern interferometry (TESPI) and preliminary results indicate that in some instances the character of the flow through flue organ pipes is dependent upon the pipe geometry. Depending on the details of the pipe construction, the exit flow may be slow and laminar or fast and turbulent. High speed TESPI images combined with pressure and flow measurements are providing insight into the importance of pipe geometry on the final sound.

ABSTRACT. The contribution explores the changes in labial air jet movements and sound caused by two different types of voicing adjustments in labial organ pipe. The movement of the air jet in the labium was observed on a transparent rectangular organ flue pipe (a copy of a wooden pipe) with adjustable upper lip possition using laser Particle Image Velocymetry (PIV). The changes were studied by modifying the mouth cut-up height and the windchest air pressure. Visualizations of the time development of the air particle velocity vectors of the jet are shown and linked to sound spectra and to descriptions of the perceived sound quality obtained from a listening test.

ABSTRACT. Pipe organ builders have made numerous inventions and spent considerable effort where they included deep bass into pipe organs despite these being located in spaced-restricted rooms. This study reviews these technical approaches discovered over centuries and illustrates them with drawings and pictures. Whereas mitre bends and stopped pipes are well-known methods to cope with limited building height there are less common approaches like reed ranks in which half-length resonators are combined with free reeds. Special attention is directed to Haskell construction of pipes. These are tonally based on the principle of adding a so-called helper rank (like the Copula), because their sound spectra reveal both, the effect of a shortened pipe resonator as well as that of an open pipe at octave pitch. Haskell pipes also relate to bi-phonic pipes, which can produce two notes with the aid of an auxiliary valve. The latter leads to the polyphonic organ pipe with a compass of up to nine semitones. The shape of the resonator may change from an elongated pipe to a cuboid, then acting as a Helmholtz resonator. The largest of them have been built as walk-in pipes for demonstration. Finally, the principle of Resultant stops, in which at least two pipes a quint interval apart sound simultaneously, is treated. These compound stops generate their designated pitch only indirectly, i.e. as virtual pitch. Although there are numerous examples of Resultant stops (in larger pipe organs), their auditory effect is often not convincing, however occasionally they do perform outstandingly well in providing their designated pitch. Which parameters are responsible for the latter seems to lie beyond current psycho-acoustic models. A final overview comprises technically viable approaches to generate tones of the lowest octave, if space restrictions apply.

ABSTRACT. The sound of pipe organs is complex, in part due to their large size and vast number of pipes. Large pipe organs are typically custom built for the spaces that house them, with the associated room acoustics being intrinsically linked to their sound. The present work investigates the perceptual effect of radiation patterns on the sound of an organ in a geometric acoustics room simulation.

Musical excerpts were recorded with two microphones positioned inside and two microphones in front of the facade of the positive section of an organ buffet in an attempt to capture the "dry" sound of the organ. Simultaneously, the music was also recorded with a binaural head from the perspective of a listener in the nave of the church (the reference). A geometric acoustics model of the church (Église Sainte-Élisabeth-de-Hongrie de Paris) was constructed and calibrated to room impulse response measurements. The dry organ recordings were then auralized with the acoustic model and with a variety of different directivity patterns. These included the directivity pattern of a proxy organ measured under anechoic conditions, the church organ used for recording the music measured in situ, and idealized directivity patterns (omnidirectional and cardioid). A listening test was performed to evaluate the extent that the directivity patterns effected the realism and plausibility of the simulations.