Bassoon spectrum transformation function
A. H. Benade, Jessie C. Carman, P. H. Barrett · The Journal of the Acoustical Society of America · 1981
Individual notes and also musical passages have been recorded with one microphone moved about the room, and the other probing the bassoon's reed cavity signal. Room-averaged (“external”) spectra p‖(ext) and mouthpiece (“internal”) spectra p‖(lnt) are measured for the harmonic components of the mf level notes C2, F2, C3, F3. The precision is ±2 dB from all causes (player variability and room fluctuations). The smoothed transformation function 〈T(f)〉 = 〈pext/plnt〉 is calculated as a continuous function of frequency. This has a broad peak (30-dB high, 500-Hz wide at 15 dB down) centered at 600 Hz. Above this it fluctuates smoothly before rolling off steeply above 2000 Hz (20 dB down by 2500 Hz). The T‖'s of individual components differ from the corresponding values of 〈T(f)〉 by about ±6 dB. Loudspeaker playback of music using the internal signal via an electronic equalizer set to 〈T(f)〉 sounds like a bassoon but lacks “life.” Boosting or cutting several randomly chosen ⅓ octave bands by 4–8 dB improves the realism considerably, in the judgement of bassoonists and others. [Work supported by NSF.]