Rolf Bader
Institute of Musicology, University of Hamburg

Andreas Bamberger
Physical Institute of the Albert-Ludwigs-University, Freiburg, Germany

The excitation mechanism and the acoustic power emission for flue instruments is an ample discussed issue in musical acoustics. Over the years of investigations two seemingly contradictory mechanisms have been put forward: The momentum transfer of the jet due to dissipation in the resonator and the resulting radiation damping on one hand, and the vortex sound theory on the other hand. Here, the experimental investigations of jet - labium interaction in flutes with the Particle Image Velocimetry are presented. The acoustic power of the instrument is quantitatively compared with estimates based on the vortex sound theory. The result is understood as a delicate superposition of vortex sheets of the jet with opposite sign between the lips and the labium. This leads to a net vorticity which interacts with the local acoustic field across the embouchure. The power according to the vortex sound theory tends to be positive, hence being interpreted as a contributor to the acoustic far field. The validity of the above models will be addressed.

P.J. Berry and L. Jones
London Metropolitan University

The bore of the hautboy, like other baroque wind instruments, shows a number of localized expansions or chambers. Examination of the bore plots of four oboes by Thomas Stanesby Sr. showed that these expansions were made in specific regions. However, the exact location of each expansion varied between different instruments suggesting that they were individual tuning or voicing adjustments.

Copies of an instrument by Thomas Stanesby Sr. were made without these expansions. A set of specially made reamers was then used to add the chambers found in the original instrument, and the effect on tuning and playing characteristics was recorded.

This secondary reaming resulted in rather subtle changes in intonation, but tended to correct errors of tuning inherent in instruments with the 'basic' bore. The most striking effect was on their sound, which was greatly strengthened.

Increasing the diameter of the choke or narrowest section of the bore by reaming flattened left hand notes in the lower register, and sharpened the equivalent notes in the second register.

The effect of localised alterations to the diameter of the bore was investigated further by inserting short lengths of plastic tubing to narrow sections of the bore. The pitch of each note of the 2-octave scale of C major was recorded in hertz for each of 15 positions of the tube. Results were analysed and tabulated to show the effect of each tube position on the 2-octave scale, the effect of the 15 tube positions on each individual note, and the effect of each tube position on each octave. Results suggest that the maker may have reamed the bore at positions most useful for selectively correcting octaves.

Sandra Carral
Institut für Wiener Klangstil, Vienna, Austria

Wind instrument players tend to agree that the material of their instrument influences the its behaviour dramatically, from the subtle "feel of it" (in terms of i.e. playability) to the actual sound produced by it. On the one hand, from a scientific standpoint, it is still not clear whether or how the material could have such a dramatic effect, especially on the produced sound of the instrument. On the other hand, young musicians are paying considerable amounts of money to buy "better" instruments made with more expensive materials. The question is: Do instruments that differ "only" in the material really feel and sound different, or do they feel and sound different (if they do) because the musician expects them to do so, and plays accordingly? A typical example of this is the case of the flute. The study presented here focuses in comparing two flutes of the same maker and model, one made of gold, and the other one made of silver. There are three levels where differences between one and the other could be found: 1. what the player feels, 2. what the player hears, 3. what the listener in the far field hears. Furthermore, it is possible that differences in these three levels are more pronounced if the player sees what instrument he/she has in his/her hands. Two experiments were performed: In the first experiment, a last year flute student of the University of Music in Vienna was asked to play the same fragment of a piece of music 30 times with the silver flute, and subsequently 30 times with the gold flute. A microphone was placed close to her left ear, and another one 3 metres away from her. She was asked to describe the differences she felt and heard. The recorded sounds were subsequently analysed using the program SNDAN, and the spectral centroid was calculated for each of the 60 samples. The second experiment consisted on a blind test, whereby the (blindfolded) player was asked to play the same piece of music, this time with the flutes presented in random order (30 times each flute), and asked which flute she had just played. The sound was also recorded close to her ear and 3 metres away from her, and the sound samples subsequently analysed as done in the first test. These two experiments were conducted in an anechoic chamber. Later on, for each test all the sound samples for each flute were separated, resulting in two groups (silver and gold) of 30 samples each, and the differences between each group were found by means of statistical analysis for each of the two tests. The results obtained will be presented, along with some thoughts on the possible sources for them.

René Caussé, Nicholas Ellis and Joël Bensoam
IRCAM, 1 place Igor Stravinsky, F-75004 Paris

The starting point of this study is the carrying out of a virtual bass clarinet for a musical project at the IRCAM. Not only is the modeling of normal notes looked for, but the sounds coming obtained by fork fingerings, multiphonics and the possibility to drive the model with the real fingerings and instrument player's control parameters as well.

The bore (internal geometry) of the virtual duplicate is the copy of the real instrument. The 3D mesh is constructed with Matlab scripts and the sound is computed with Modalys, a physical modeling synthesis software using modal theory. The blowing pressure and the flow inside the mouthpiece are related through a nonlinear reed-valve characteristic.

The first step consisted in measuring the bore of the real instrument, which called for a complete disassembly of all keys. For the various holes, we refined the measures by making moldings of the chimneys in silicone resin.

After that, we validated not only the mesh chosen for the resonator (barrel and mouthpiece included), but the hypotheses made in the 3D finite model as well. The validation has been carried out by comparing resonances amplitudes and frequencies, for various fingerings of the bass-clarinet, multiphonics included. These values are obtained from acoustical input impedance measurements.

At the same time, we carried out a comparative study between two models, the first, old fashioned, is a one-dimensional model including visco-thermal losses by a method of discretization, using truncated cones and the second, a 3D finite element model using the software Modalys. The results of that study will be presented and the influence of various errors and approximations (geometry, mesh and measures) will be discussed for the two models.

Finally, sound examples will illustrate this presentation on the making of a bass clarinet, showing possibilities for instrument craftsmanship and musical projects.

Delphine Chadefaux (1) and Wilfried Kausel(2),
(1) Université Pierre et Marie Curie, Paris, France (currently intern at Institut für Wiener Klangstil, Vienna)
(2) Institut für Wiener Klangstil, University of Music and Performing Arts, Vienna, Austria

The technical committee musical acoustics (TCMA) of the EAA has established a working group (WG2) for Musical Synthesis and Modelling, initiated by M. Kob and the supervising author. The first contribution by A. Braden was object oriented C++ code as part of his thesis based on the work of J. Kemp, both graduates of Edinburgh. The original code was for Linux and consisted of bore discontinuities, straight and circularly bent cylindrical tubes and cones and Bessel horn sections for single and multimodal treatment in the frequency domain. As a first step this code has been ported to Windows and restructured to speed it up. As a second step new interfaces have been added allowing quiet shell execution using command line switches and it has also been compiled as a dynamic or static library to be used by external applications providing a graphical user interface. Currently tone hole models are being implemented and preparations are being made for caching intermediate results of instrument parts which are not modified during optimisation or interactive bore modifications. Rectangular ducts are ready for implementation and feasibility of cones or Bessel horns without discretization is currently being studied. In future time domain simulation should be included.

John Chick, D. Murray Campbell and Arnold Myers
University of Edinburgh

As brass musical instruments age, the inner surfaces of the bore of the instrument are likely to change. These changes can be as a result of a number of different corrosion regimes, and/or as a result of lack of cleaning and the accumulation of material on the tube wall. Although the bore profile may appear to be superficially unchanged from that of the instrument in "as new" condition, changes in surface condition can have significant effects on the acoustic response, and hence playing characteristics, of the instrument.

This paper explores the use of non invasive measuring and modelling techniques which may be applied to historic brass instruments to provide insight into the playing characteristics of the instrument. The presence and location of leaks can be identified using pulse reflectometry techniques. For instruments with no significant leaks, the acoustic input impedance of the instrument is measured and the results compared with those from computational simulations based on physical measurements of the bore profile. By adjusting the attenuation coefficient in the model to provide a good match with the measured data we can make some observations about the internal condition of the instrument, and suggest ways in which the playing characteristics of the instrument in its current state may differ from those of the instrument immediately following its manufacture.

Jean-Pierre Dalmont
Laboratoire d'Acoustique de l'Université du Maine, UMR CNRS 6613, Av. O. Messiaen, 72085 Le Mans cedex 9, France

It is now well known that the bore profile of a wind instrument can be deduced from its reflection function r(t) [1]. This reflection function is usually measured by using the acoustic pulse reflectometry technique [2]. One of the problems of this technique is the difficulty to reach a high signal to noise ratio. Another solution is to perform an input impedance measurement, from which the reflection function R(w) can be deduced. An inverse Fourier transform leads to the reflection function r(t) and the same algorithm [1] can be used to obtain the bore profile. The advantage is that input impedance measurement is a mature technique with a high accuracy especially by using the new input impedance sensor which has been recently built and which allows low frequency measurements [3]. Results show that the bore profile can be efficiently deduced from impedance measurements when strong discontinuities are avoided and when measurements are performed beyond the cut-off frequency. The question of the accuracy is discussed.

[1] N. Amir, G. Rosenhouse, U. Shimony: "A discrete model for tubular acoustic systems with varying cross sections - the direct and inverse problems. Part 1: theory" Acustica, 81, No.5, 450-462 (1995)
[2] D. B. Sharp and D. M. Campbell,"Leak detection in pipes using acoustic pulse reflectometry", Acustica, 83, 560-566 (1997).
[3] J.P. Dalmont, J.C. Le Roux, B. Gazengel, "A new impedance tube for large frequency band measurement of absorbing materials",.Acoustic's 08 Paris, 1-6.

Mathew Dart
London Metropolitan University

Woodwind instruments from the early eighteenth century give the appearance of simple technology with just six to eight open fingerholes and a few keys. However they often contain complex bore geometries not found in their modern counterparts which are only revealed by careful measurement.

The musical demands made on these instruments were considerably greater than those of their predecessors. They were required to play in a wide range of keys while making the enharmonic distinctions of non-equal temperaments, to play with increased expressive capability and to fill solo and ensemble roles in the latest concertos, orchestral and chamber music.

Subtleties of bore design were part of the means by which the playing characteristics were developed in answer to these pressures, while keeping the number of toneholes to a minimum.

Makers reached a peak of expertise in the manipulation of tuning and response through bore modification in the eighteenth century, but this decreased in significance as the process of adding greater numbers of toneholes with associated keywork got underway. It is now a field of rediscovery for reconstructors of baroque woodwind instruments.

This paper looks closely at the bore design of bassoons by Johann Poerschman (Leipzig c 1680-1757) and attempts to understand the reasoning behind specific features of the bore. Use is made of mathematical modelling of bore resonances alongside playing tests on an instrument reconstructed using multiple small reamers in order to show how some of the problems of early bassoons are met.

Joël Gilbert
CNRS, Laboratoire d'Acoustique de L'Université du Maine (UMR CNRS 6613), Le Mans, France

When a brass instrument is played making a crescendo, the energy level of the higher harmonics increases, and the sound becomes brighter. If the brass is played loudly, then the increase is dramatic, and the sound is becoming dramatically bright, then called "brassy" or cuivré. The brightness and brassiness in brass instruments is controlled by several phenomena:

• the embouchure of the player who controls the modulated flow entering into the instrument,
• the bore of the instrument defining the transfer function (linear acoustic) between the input acoustic pressure and the radiated sound,
• the non-linear steepening of the wavefront and, in extreme cases, the generation of shock waves within the instrument bore.

Timo Grothe (1), Johannes Baumgart (2) and Roger Grundmann (1)
(1) Institute for Aerospace Engineering, Technische Unviversität Dresden, Germany
(2) Insitute of Scientific Computing, Technische Unviversität Dresden, Germany

The intonation of woodwind instruments is a crucial topic from the manufacturers point of view. Most instruments are manufactured in serial production. Even if the parts are made using high precision woodworking machinery, every single instrument needs individual reworking of the intonation. Making highest quality instruments, the maker has to perform delicate corrections at the tone holes of the ready-built instrument, in order to meet the customer needs. For musicians, the intonation is one of the most important quality aspects of woodwinds. The ease of playing, and therefore the possibilites of artistic articulation on an instrument highly depend on the tuning corrections that are necessary during performance.

In this study we present curves of the dynamic range at a constant pitch for all playable notes of the bassoon. The curves were obtained from sound samples of several german manufactured bassoons, played by a professional bassoonist. Several reeds and bocals were used in the investigation. The rising of overtones, as the dynamic level is increased from pianissimo piano to forte fortissimo, is mapped using principal components analysis.

This experimental study was supported by instrument makers and focuses on their needs of objective measures in the process of fine tuning. Comparing these curves and perceptional judgements of the bassoonist, we discuss quality aspects of the bassoons from the musicians point of view.

Michael Hutley

We seek to apply an understanding of the physics of the operation of reeds at a level which will be helpful to the maker and musician. To do this we need data which are sufficiently precise and accurate to reflect subtle differences that are often of great musical significance. This paper will review progress in developing techniques of measurement which are both accurate and sufficiently convenient that they may realistically be used by reedmakers. The current work concentrates on optical and mechanical methods for measuring the stiffness profile of double reeds.

Wilfried Kausel
Institut für Wiener Klangstil, University of Music and Performing Arts, Vienna, Austria

Previous experiments demonstrated a significant audible influence of wall vibrations on the radiated sound of french horns. Timbre differences observed in the near as well as in the far field could with high statistical significance be attributed to the vibration amplitudes measured near some antinodes of certain deflection shapes of the bell. Sound levels originating from comparable multi pole radiators have been estimated and more accurate boundary element simulations have been performed both strengthening the hypothesis of an audible effect of vibrating brass mainly of the bell. Finally near sound field recordings have been made with Rolf Bader's 128 microphone array[1] at the Institute of Musicology of the University of Hamburg. After post processing 5GB of wave files 250GB of video files have been generated illustrating the time resolved sound pressure distribution in a plane directly in front of the bell and in a plane perpendicular to that plane. Many of those animations do show complicated modal patterns with several nodal diameters, which are not compatible with the assumption of an axial symmetric air column as the only source of sound. In some human lip driven cases, recorded at higher sound levels, the formation of shock waves becomes apparent.
[1] Thank you Rolf for giving me that unique opportunity!

Jonathan Kemp, Maarten van Walstijn and Richard Smith

Acoustic methods of measuring tubular systems (such as brass musical instruments) can be characterised either as impedance head methods which analyse signals entirely in terms of fixed amplitude infinite duration sinusoids and reflectometry methods that also exploit causality. Traditionally, acoustic pulse reflectometry has largely been based on the assumption that the source tube must be long enough to completely isolate all the reflections from the object under test. The current paper discusses how this condition can be relaxed: only the primary reflections from the object under test need to be isolated.

James Kopp
Hoboken, New Jersey, U.S.A.

For more than a century (1780-1911) bassoon methods called for the player to use an oblique embouchure. That is, the reed was to be rotated slightly on the crook so that formed "an angle with the lips". This widespread practice, largely forgotten during the 20th century, had significant consequences for early reed design, fingerings for early bassoons, and octave venting. A simple A-B comparison of results obtained via the oblique embouchure vs. a conventional embouchure offers valuable insights into some of the effects of altered lip-damping conditions. After a summary of the written evidence, I will demonstrate some of the acoustical effects of the oblique embouchure.

Jean-François Petiot (1), Marie Françoise Lucas (1), Joël Gilbert (2)
(1) Institut de Recherche en Communications et Cybernétique de Nantes (UMR CNRS 6597) - 1 rue de la Noë, BP 92101, 44321 Nantes Cedex 3 France, {petiot,lucas}@irccyn.ec-nantes.fr
(2) Laboratoire d'Acoustique de l'Université du Maine (UMR CNRS 6613) Av. O. Messiaen, 72085 Le Mans Cedex 9, France, Joel.Gilbert@univ-lemans.fr

This paper addresses the use of sound simulations for instrument characterisation. We focus on the ability of simulations by physical modelling to create sounds characteristic of a given instrument: is this technique accurate enough to produce dissimilar sounds for two different instruments? Are there similarities with sounds played by a musician? We used the harmonic balance technique to generate trumpet sounds in permanent regime. The input parameters of the simulations are the input impedance of the trumpet (resonator), the control parameters are the characteristics of the virtual musician (excitator), and the outputs of the simulations are the playing frequency and the magnitude of the 6 first harmonics of the notes. Three different trumpets, obtained by geometrical variations of the leadpipes, were first simulated using several virtual musicians, and second played by a 'real' musician. The two populations of sounds produced were characterized by their spectrum in permanent regime. Firstly, in order to see in which extent the simulated sounds are different, Principal Component Analysis (PCA) and Factorial Discriminant Analysis (FDA) were used to analyse the population of sounds. For very dissimilar trumpets according to their input impedance, the simulated sounds are clearly differentiated. For more similar trumpets, the variability due to the virtual musician is in the same order than those of the instruments. More information will be necessary for separating the sounds, as the evolution of the sounds characteristics during an increasing of the pressure supply (crescendo sounds).

Secondly, the similarities between the simulated sounds and those played by a musician, were studied. For a same trumpet, the simulated sound and the sound played by a musician were obviously very different. But similarities in the differences between the sounds were noticed for the set of 3 trumpets.

As a conclusion, either with a simulated or with a 'musician' sound, an acoustical signature characteristics of the instrument can be found in the signal.

Adrien Mamou-Mani and David Sharp
The Open University, Milton Keynes

For large-scale musical wind instrument manufacturers, the ability to produce instruments in a repeatable fashion is essential. However, despite the tight manufacturing tolerances used, professional musicians are often able to discern small, but perceptible, differences between the playing characteristics of instruments manufactured in an identical way. These differences are most likely a result of tiny disparities in bore profile or in the positioning/sealing of any side holes.

This talk outlines a programme of work designed to investigate the consistency with which manufacturers are able to make wind instruments and to explore the causes of any musical differences between the instruments. State-of-the-art techniques will be used to measure the internal geometries and resonance characteristics of batches of nominally identical instruments. Meanwhile, the musical qualities of the instruments will be established through a series of psychoacoustical tests. The potential effectiveness of the proposed approach will be demonstrated through a set of measurements made on two low-cost, mass-manufactured trumpets.

Jer Ming Chen, John Smith and Joe Wolfe
School of Physics, University of New South Wales, Sydney, Australia

In a simple model of reed or lip reed operation, the impedance of the bore of a instrument (looking in from the mouthpiece) and that of the player's vocal tract (looking in from the lips) are acoustically in series [1]. The acoustical impedance spectra of instruments have a series of impedance peaks, typically several tens of megohms, and these usually determine the playing regime. That this is not the whole story is demonstrated by the didjeridu, in which much of the musical interest comes from the variations in spectral envelope produced by the effects of resonances in the vocal tract [2], as well as interference effects produced by simultaneous vibration of lips and vocal folds.   We report the results of measurements of the impedance spectrum of the vocal tract of performers, while they are playing. Here we concentrate  on clarinets and saxophones. On the saxophone, the magnitude of impedance peaks decrease strongly with frequency in the high registers, a result of the conical bore. This limits the performance range - unless the performer can tune a tract resonance so that its series combination with a weak, high resonance of the instrument can control the reed. This technique is not needed for the altissimo range of the clarinet. Nevertheless, clarinettists can learn to produce resonances in the tract with sufficiently high impedance that they can largely determine the playing pitch [3]. The famous glissando of Gershwin's Rhapsody in Blue is a well-known example. 

[1] Benade, A.H. (1985). "Air column, reed , and player's windway interaction in musical instruments", in Vocal Fold Physiology, Biomechanics, Acoustics, and Phonatory Control, Chap 35, pp.425-452.
[2] Tarnopolsky, A, Fletcher, N. Hollenberg, L., Lange, B., Smith, J. and Wolfe, J. (2005). "The vocal tract and the sound of a didgeridoo", Nature, 436, 39.
[3] Chen, J.M., Smith, J. and Wolfe, J. (2008). "Experienced saxophonists learn to tune their vocal tracts", Science, 319, 726.

Mitsuru Ikushima* and Shigeru Yoshikawa
Graduate School of Design, Kyushu University, 4-9-1 Shiobaru, Minami-ku, Fukuoka, 815-8540 Japan
*Currently with Disco Corporation, 2-13-11 Omori-Kita, Ota-ku, Tokyo 143-8580 Japan

Hiromi Tashiro* and Shigeru Yoshikawa
Graduate School of Design, Kyushu University, 4-9-1 Shiobaru, Minami-ku, Fukuoka, 815-8540 Japan
*Currently with Roland Corporation, 2036-1 Hosoemachi-Nakagawa, Kita-ku, Hamamatsu, 431-1301 Japan

Shigeru Yoshikawa
Graduate School of Design, Kyushu University, 4-9-1 Shiobaru, Minami-ku, Fukuoka, 815-8540 Japan

The so-called "cross fingering", which crosses open tone hole(s) and closed tone hole(s), is usually used to lower the tone by a semitone in woodwind instruments. However, the cross fingering in Japanese shakuhachi is applied to raise the tone by a semitone, too. This reverse phenomenon is called "intonation anomaly" in the present paper. Such a phenomenon is also seen in simple flutelike or recorderlike instruments like Chinese flute and Irish tin whistle. However, most players are almost unconscious that the cross fingering is used to bending up as well as to bending down, and they do not think this intonation anomaly as the anomaly in the scene of actual playing. A few data on the intonation anomaly are obtained from the shakuhachi that has four holes in the front and one hole in the back. The fingering Wu3 (only the 3rd hole open) yields the bending-up when the shakuhachi is blown in the second register: The tone frequency of Wu3 is about 904 Hz; that of Chi (the 1st, 2nd, and 3rd holes open) is about 880 Hz (A₄). However, the Wu3 never yields the bending-up when the shakuhachi is played in the first register. Numerical calculation is carried out by applying the branching-system theory to these two fingerings (Chi and Wu3), and the result is compared with the measured data. Also, the respective standing-wave patterns in the bore are measured, and then the possible cause of the intonation anomaly is discussed.

For further information, please contact Murray Campbell, School of Physics University of Edinburgh, James Clerk Maxwell Building, King's Buildings, Mayfield Road, EDINBURGH EH9 3JZ, E-mail: D.M.Campbell@ed.ac.uk or Arnold Myers, Edinburgh University Collection of Historic Musical Instruments, Reid Concert Hall, Bristo Square, Edinburgh EH8 9AG, E-mail: A.Myers@ed.ac.uk