NAG/DAGA 2009 - Rotterdam Reconstructing the Lituus: A Reassessment of Impedance, Harmonicity, and Playability A. G. Apostoli1,S.M.Logie2,A.Myers3,J.A.Kemp4,J.P.Chick5,andA.C.P.Braden6 1 School of Physics and Astronomy, University of Edinburgh, United Kingdom, Email: [email protected] 2 School of Physics and Astronomy, University of Edinburgh, United Kingdom, Email: [email protected] 3 School of Arts, Culture and Environment, University of Edinburgh, United Kingdom, Email: [email protected] 4 School of Physics and Astronomy, University of Edinburgh, United Kingdom, Email: [email protected] 5 School of Engineering, University of Edinburgh, United Kingdom, Email: [email protected] 6 School of Physics and Astronomy, University of Edinburgh, United Kingdom, Email: [email protected] Introduction studies included theoretically modelling the fiscorn from measurements of the bore profile, using BIES to calculate J. S. Bach’s evocative funeral motet ‘O Jesu Christ, the theoretical input impedance curves, measuring the meins Lebens Licht’ (BWV 118), appearing in two ver- input impedance curves using the Brass Instrument sions, was thought to have first been performed in 1736-7. Analysis System (BIAS) designed by scientists at the On inspection the first version of Bach’s score calls for Institut f¨ur Wiener Klangstil, Vienna[3], and finally four voices, cornett, three trombones, and two litui, and conducting a brief playing test using two experienced the second version for four voices, strings, bass continuo, brass instrument players. and two litui. But what exactly are these mentioned litui? Are they a specific type of instrument or just an Following these initial studies an optimisation recipe alternative name for an instrument already in existence consisting of an input impedance curve template, target at the time? To try and answer these questions a input impedance peak frequencies, and a template for team of researchers from the Schola Cantorum Basiliensis the instrument geometry with assigned upper and lower in Basel, Switzerland, in collaboration with scientists bounds were generated and input into BIES to find the and musicians from the University of Edinburgh, United optimal design for the lituus within specified parame- Kingdom, have designed and built two functioning litui ters. Input impedance measurements were taken for one which may be similar to the instruments Bach could have of the completed instruments and these measurements had at his disposal during a performance of BWV 118. were then compared with theoretically calculated input The project concluded with a public concert performance impedance curves for the optimised lituus. and formal presentation of the litui in Basel in January 2009. Measurement and Analysis Tech- These questions of exactly what the litui Bach refers niques to in his score looked and sounded like are not novel. Indeed, the musicologist, Ebenezer Prout, is known Input Impedance to have pondered these very same questions in an The acoustic input impedance Z(ω) of any brasswind article to the Royal Musical Association back in 1885[1]. instrument is defined as the ratio of acoustic pressure Work conducted by researchers at the Schola Cantorum p(ω)tovolumeflowU(ω), taken at the input plane of Basiliensis indicated that the lituus may have been a the instrument, where ω is the angular frequency for a long wooden natural instrument, sounding somewhere sinusoidal input signal at the input plane. Acoustic input between a trumpet and an alphorn, and playing a impedance is measured in units of acoustic ohms (Ω). series of natural harmonics. Scientists at the Musical Impedance is an important measurement to have in the Acoustics Laboratory at the University of Edinburgh context of brasswind instruments, as it is this which de- were contacted to collaborate on the design process. termines the playing characteristics and linear acoustical Provided initially with a bore profile formed from in- behaviour for the instrument under investigation[4, 5]. terpolated measurements made for a Nagel trumpet The positions of the peaks on an input impedance (17th Century Nuremberg natural trumpet in D flat) spectrum also determine whether or not a note should and straight b¨uchel (wooden alphorn-like instrument), be easy to play. On a trumpet, for instance, if a theoretical models were then constructed and analysed particular input impedance peak has a high magnitude using the Brass Instrument Evolution Software (BIES), at a frequency at which the player wishes to play, then recently developed at the University of Edinburgh[2, 5]. this note will be easy to play. When the opposite To acquire a better understanding of instruments similar happens and the player wishes to play a note at which the in design to that proposed for the lituus, initial studies corresponding input impedance peak is of low magnitude, were undertaken on a 2.4m fiscorn (a type of Catalo- then the note is difficult to play. In this situation the nian flugelhorn) in C, from the Edinburgh University player is required to lip away from the input impedance Collection of Historic Musical Instruments. These initial ‘slots’ in order for the note to sound from the instrument. 854 NAG/DAGA 2009 - Rotterdam Equivalent Fundamental Pitch constant usually defined in terms of specified length and Another important measurement in producing an opti- input/output radii for the section expressed[2]. mal design for the lituus was the Equivalent Fundamental Pitch (EFP). Derived from the acoustic input impedance Initial Investigation using a Fiscorn curves and closely related to the Equivalent Fundamen- Considering it is accepted that discrepancies do exist tal Frequency, an EFP plotted graphically provides a between the input impedance of an instrument when straightforward method for assessing how close individual measured and calculated theoretically[5] and to assess input impedance peaks are to corresponding members of the input impedance results obtained theoretically with a single harmonic series. The EFP is calculated for each BIES, initial investigative work was undertaken on an input impedance peak using the following equation instrument with notionally similar dimensions to the proposed lituus design. A 2.4m long fiscorn in C attached to a suitable mouthpiece had input impedance measure- 1200 fi EFP(fi)= × log [cents] (1) ments taken using BIAS. Theoretical input impedance log(2) iF curves also were calculated using BIES from a model th created from detailed measurements taken by hand and where fi is the frequency of the i input impedance th finally a brief playing test was conducted by two experi- peak and iF is the frequency of the i harmonic based enced brass players. on an arbitrary fundamental frequency F [2]. The EFP, 5 expressed in the more musically relevant unit of cents, 10 Measured Fiscorn therefore indicates for our given value of i the deviation Theoretical Fiscorn of the impedance peak frequency from the frequency of ) its corresponding component from the harmonic series. Ω 4 EFP was extremely useful throughout the optimisation 10 routine, as it was possible to draw quick conclusions as to the suitability of theoretical instrument designs based on their harmonicity. Acoustic Impedance (k 3 10 Brass Instrument Evolution Opti- 0 200 400 600 800 1000 1200 1400 1600 1800 2000 misation Software Frequency (Hz) The Brass Instrument Evolution Software (BIES), devel- Figure 1: Input impedance curves measured for a 2.4m oped at the University of Edinburgh[2, 5], provides an long fiscorn at playing conditions (solid line) and theoretically dashed line integrated software package which allows for the design, calculated ( ). testing, and optimisation of brass instruments theoreti- An interesting comparison to be considered was between cally. Specifically, BIES allows the musical instrument the measured input impedance for the fiscorn, warm and researcher to obtain calculated input impedance curves in playing condition, and the input impedance calculated for theoretically described instrument geometries and an theoretically (see Figure 1). Although the measured optimisation routine where a target input impedance is and theoretical input impedance peaks are at almost specified and, should suitable optimisation parameters the same frequency, their relative peak magnitudes are have been chosen; an optimal input impedance curve and quite different. It is believed that this discrepancy is associated smooth bore profile is produced. Throughout most likely to be a combination of acoustic energy loss the optimisation process spherical-wave propagation was due to dirt and sediment lining the internal tubing, as assumed. mentioned in Watkinson et al[7], and leaks due to the age of the instrument. The playing test results confirmed One of the features of BIES is that it uses the Rosenbrock that the measured input impedance peak frequencies direct-search algorithm to find an optimised bore profile were close to the frequencies at which our two brass with the required input impedance curve across a more players played. Typically, the deviation was within ±5Hz constrained design space. This has the effect of improv- of the measured impedance peak. ing the convergence to a realistic instrument design and the time within which it does this. A more detailed consideration of BIES and the optimisation process is Designing the Lituus available in other sources[2, 5]. Another feature of BIES With the initial investigation work complete, the original is the ability to express the bell of an