Page  485 ï~~Analysis of "Touch Effect" on the Transient of Pipe Organs with Mechanical Transmission. Laura Bazzanella, Giovanni B. Debiasi C.S.C.- D.E.I., Universita' di Padova, via Gradenigo 6A - 35131 Padova (Italy) tel. +39+49-8287500, fax +39+49-8287699, Email Abstract In pipe organs with mechanical transmission, the organists notice a response from the instrument depending on the type of solicitation, i.e. the so called "touch". In order to explain this intuitive observation, we have considered the sounds of several stops of organs with mechanical transmission, using for all notes two different ways of "touch": slow and fast, employing for this aim a mechanical finger. With different analysis methods, we got the evolutions of the timbre of the considered sounds, obtaining very different behaviours. Our results have proved a real difference between the same note played with a different "touch". From a general point of view, we can notice a remarkable sensitivity of the second spectral component. The results of the present research can be useful for the synthesis of pipe organ sounds in high-level electonic organs. 1. Introduction Sounds of pipe organs are generated from pipes boosted with an air jet driven by the keys of one or more keyboards. Most organists judge that the pipe organs with mechanical transmission are preferable in order to act with the touch on the transient of each note. However there are discordant opinions about the effectivness of the touch influence and so we decided to give a new contribution for the solution of the question. 2. Methods of sound recording and analysis We considered the sounds of flue pipes of two high quality modem organs with mechanical transmission, namely the organ of Vincenzo Mascioni of the Conservatory "C. Pollini" in Padua and the organ of Franz Zanin of the S. Giorgio Church in Vicenza. For each organ we recorded all stops considering for each note two modalities of touch: slow and fast. The definition of these two categories has been done by expert organists and they adjusted a mechanical finger to obtain two very constant key velocities, slow and fast, according with their subjective judgement. Each note has been recorded from the starting transient up to five seconds with an AKG condenser microphone mod. C414B-ULS and with a D.A.T. Denon mod. DTR 2000, at the sampling frequency of 48 kHz. We also used a B&K phonometer with a condenser microphone for monitoring the intensity level (dB). We adopted a system of acoustic shields in order to reduce the incidence of the reverberating sounds on the microphone. The sampled sounds have been processed with different analysis methods: in particular with the STFT (see Limg & Oppenheim [1988]), with a proprietary algorithm for discrimination between harmonic and inharmonic components of the sound and with a method for automatic evaluation of timbre and fluctuations of pipe organ sounds [Dal Sasso et al., 1991]. 3. Experimental results For both organs, the starting transients of the sounds analized with STFT showed a remarkable difference between the time-evolution of the amplitude of the first six harmonics in correspondence with slow and fast touch. Particularly, we find that the steady-state of the amplitude of the second harmonic is reached with an overshoot in the case of fast touch, while in the case of slow touch we can observe no overshoot (Figure 1). The automatic assessment of the timbre on a polar chart demonstrates a real difference in the timbre evolution on the starting transient of sounds with a different touch (Figure 2). Finally, the extraction of the inharmonic component has shown that its starting transient also presents a different evolution, according to the type of touch (Figure 3). These two last methods of analysis, to our knowledge, have not be used before for this kind of analysis. The differences of the second harmonic evolution are in accordance with the theoretical prevision of Fletcher [1976], Finch & Nolle [1986], Caddy & Pollard [1957], and all other measurement we did confirm the effective dependence of the starting transient on the touch. Nevertheless, we also noticed that sometimes the notes of the stops with long-distance mechanical transmission (having very long torque bars) can present no differences between slow and fast touch. We try to explain this fact thinking that, with slow touch, there is torsional energy accumulating in the bars and, as this energy reaches the needed value, it causes the abrupt opening of the air valve. ICMC Proceedings 1994 485 Acoustics

Page  486 ï~~It is our intent to probe this hypothesis in a future work. In any case, the results we achieved consent to explain the doubts and the contradictions found about this matter (see, for instance, Girardi [19901). 4. Conclusions In order to get a deeper knowledge of the behaviour of pipe organs, in view of a more realistic electronic synthesis of their sounds, we undertook a detailed analysis of the characteristics of pipe organs. Particularly, we studied how the touch can modify the starting transient of pipe organ sounds. The results we obtained confirm that the notes of pipe organs with short-distance mechanical transmission have a starting transient strongly depending on the touch. Two of our analysis methods can be considered innovative and they also confirm the influence of the touch on the timbre during the starting transient. Nevertheless, this effect can disappear for notes of stops with long-distance mechanical transmission, and this can explain some of the doubts still existing on this matter. References [Caddy & Pollard, 1957] S. Caddy & H.F. Pollard. Transient Sounds in Organ Pipes. ACUSTICA, vol.8, pp. 277-280, 1957. [Dal Sasso, Debiasi & Spagiari, 1991] M. Dal Sasso, G.B. Debiasi & G. Spagiar. Method for Automatic Evaluation of Timbre and Fluctuations of Pipe Organ Sounds. International Computer Music Conference, 1991. [Finch & Nolle, 19861 T.L. Finch & A.W. Nolle. Pressure Wave Reflections in an Organ Note Channel. Journal of the Acoustical Society of America, vol. LXXIX-5, pp. 1584-1591, 1986. [Fletcher, 1976] N.H. Fletcher. Transients in the Speech of Organ Flue Pipes - A Theoretical Study. ACUSTICA. vol. 34, pp. 224-233, 1976. [Girardi, 1990] E. Girardi. Problematiche relative all'Esecuzione su Organi Liturgici in Situazioni e Modalita' Diverse di Alimentazione". III Convegno di Organologia sul Tema: La Riforma de1'Organo Italiano, Pisa, 1990. [Lima & Oppenheim, 1988] J.S. Lima & A.V. Oppenheim. Advanced Topics In Signal Processing, Prentice Hall 1988. tiune- time 1st I 2nd a time lOOms lse b 2nd 'I, _ _ lt looms time Figure 1: starting transients of 1st and 2nd harmonic of Zanin Principal 8!-C3; a: slow touch; b: fast touch. -4 SOOm. 500ms Figure 3: starting transients of inharmonic components of the same note; a: slow touch; b: fast touch. Figure 2: polar chart of timbre evolution (each dot every 20 ins) on the starting transient of the same note; a: slow touch; b:fast touch. Acoustics 486 ICMC Proceedings 1994