Pressure measurements in an air supply system of a flue pipe organ by P.-M. Koenig (~1970)

Data relative to Reduced model of the air supply system of a flue pipe organ Structure of the archive <root>/AcquisitionsRaw: entire data dump of each pipe's 300-second acquisition <root>/Geometry: dimensioning of instrument and pipework <root>/MetricsExtracted: processed data from /AcquisitionsRaw folder Data in <root>/AcquisitionsRaw 22 .mat files, one per pipe, 700 MB each. The filename starts with a numbering from A01 to A56 corresponding to the note played within the span of the keyboard, following inventory notation. Next information is the pitch, diatonic or sharp. The last piece of information in the filename before the file extension is the time stamp when the acquisition started. Data structure inside the files PR_data: 6 channels with all data scaled to physical units ([Pa] for pressure, [m/s] for key velocity) presented in a 6x15360000 matrix form: Ch 1: Pressure in bellows, $P_b(t)$ Ch 2: Pressure in pallet box, $P_p(t)$ Ch 3: Pressure in groove, $P_g(t)$ Ch 4: Pressure in foot, $P_f(t)$ Ch 5: Sound pressure at 20 mm in front of center of the labium, $P_{rad}(t)$ Ch 6: Key velocity (ca. 7mm from the edge of each key) PR_metadata Temperature [º C] Relative humidity [%] mics: serial number of each probe (Keller, Endevco, Endevco, Endevco, Brüel & Kjaer) device: acquisition device details dataFilename PR_params Tacq: time duration of the acquisition [s] SR: sampling rate [Hz] chanNumber: number of acquired channels sens: conditioner's output sensitivity applied to each probe after calibration Wm: mouth cut-up distance [m] tvec time vector in [s] Data in <root>/Geometry Origin of the pipe work used in the experimental setup: the pipes are from a Principal purchased in Stoke-on-Trent (UK), from David Cooper, who reported that they were manufactured by Wadsworth of Manchester for the Stanley Street Methodist Church (Openshaw, Manchester). The files containing geometrical information are: Big_H_m.mat: $H$ distance in jet section [m] (equivalent to width of the mouth and width of the flue) Dp_m.mat: pipe's resonator diameter [m] FlueExit_Sj_m2.mat: jet section $[m^2]$ FootInlet_Sin_m2.mat: foot inlet section $[m^2]$ GrooveSlotWidth_m.mat: width of slot and groove [m] Lp_m.mat: length of resonator [m] PalletGeometry_m.mat: LHS_m: distance to neighbouring pallet on the Left Hand Side of current pallet [m] Pallet_Length_m: 0.1575 [m] Pallet_Travel_m: distance from closed position to stationary opening [m] Pallet_Width_m: width of pallet [m] RHS_m: distance to the pallet's neighouring next pallet on the Right Hand Side [m] Small_h_m.mat: little $h$ distance in jet section [m] (from bevel to lower labium) ToneHoleDiam_m.mat: diameter of cilindrical perforation [m] on the top board (immediately under the pipe foot) Vf_m3.mat: foot volume $[m^3]$ Wm_m.mat: $W_m$ cut-up distance [m] (from flue exit to upper labium) Data in <root>/MetricsExtracted data_proc.mat contains steady-state and transient values extracted from the measured attacks. The fields contain one value per acquisition (mean and standard deviation): Ppall_mean and Ppall_std: target pallet box pressure at pipe's steady-state [Pa] Pgrv_mean and Pgrv_std: target pressure in the groove at pipe's steady-state time [Pa] Pf_mean and Pf_std: foot target pressure in steady-state [Pa] PRTgrv_mean and PRTgrv_std: pressure rise time in the groove [s] PRTf_mean and PRTf_std: pressure rise time in the foot [s] PRTrad: mouth-radiated pressure rise time [s] Prad: total pressure target envelope of radiated sound (max amplitude) [Pa] F1: resonator's fundamental oscillating frequency [Hz] as per radiated sound, full $P_{rad}(t)$ signal, at 20 mm from the labium and extracted via Yin algorithm t10grv: $t^{10}$ of groove pressure since key-stroke $t_0$ [s] t10ft: $t^{10}$ of foot pressure since key-stroke $t_0$ [s] t10rad: $t^{10}$ of radiated pressure since key-stroke $t_0$ [s] MaxNegKeyVelocities.mat contains a $50\times 22$ matrix with all transients' maximal key-downwards velocity [m/s] for the 22 pipes and 40-50 transients per acquisition. Where no data exists, it has been populated with NaN.