What do we hope to measure?

Transcription

1 Impact of excitation and acoustic conditions on the accuracy of directivity measurements, Sebastià V. Amengual Garí Detmold University of Music Erich Thienhaus Institute What do we hope to measure? A: Radiation properties of an instrument or B: Perception of sound in the far field What do we measure? Sound pressure in a room at various locations 2 of 14 1

2 Variety of methods Human vs. artificial player Single vs. multiple microphone set-up Spherical vs. plane mic array Absolute vs. normalised SPL values Polar plots vs. animated balloons Selected frequency bands vs. high resolution Use: texbooks vs. sound synthesis: 3 of 14 Classic approach Place musician with instrument in anechoic room Record SPL in 2 m distance Repeat measurement at different angles Average SPL in frequency bands Normalise each band to 0 degree value Create a polar plot for each band Meyer/Marshall 1985; Kob of 14 2

3 Classic problems Repeatability of excitation mostly poor (cf. presentation of Dusty Eddy tomorrow!) Musician is part of the pattern Looong procedure Some partials hit the freq. bands, some don t Anechoic room imposes problems: huge room needed for large instruments not the favourite performance room for artists FFT over sustained sound lacks transient directivity effects 5 of 14 Modern approach Simultaneous recording in a plane or on a sphere Real-time imaging of directivity Representation in balloons Problem of acoustic centre 54-channel sphere in Graz ISMA Le Mans of 14 drawing: Franz Zotter 3

4 Representation of results Decomposition into spherical harmonics of a limited order Images by Franz Zotter Superposition of bubbles of three harmonics: + + = 7 of 14 Impact of reflections Example: Music stand in anechoic room Dodecaeder Six channels Placed on turntable White noise of 14 4

5 Conference, Location, Date Autor 9 of XX Impact of reflections Reflecting surfaces no music stand of 14 5

6 Impact of reflections Reflecting surfaces with music stand of 14 Comparsion w/wo music stand No difference at low frequencies < 2 khz Amplitude changes/inversion from 2 to 5 khz 12 of 14 6

7 Radiation from a bassoon cromatic Scale, Artist: Timo Grothe 15 of 14 Measurements: conditions Condition Stand position Oriented Distance Height Inclination C1 No stand on stage C2 Front Yes C3 Front No C4 Front-right No 81 (front), 75 (right) C5 Front-right Yes 81 (front), 75 (right) C6 Right Yes 62 (right) C 1 C 4 C5 C6 16 7

8 Results: Frequency Response 17 Results: Time Domain 18 8

9 BAC K BAC K LEF T Results: Spatial Visualization Broadband, entire IR, 3 projected views FRO NT FRO NT RIGH T BOTT OM BOTTO M [Tervo2016] Tervo, S.; Pätynen, J.: SDM Toolbox, Matlab File Exchange SDM Toolbox [Tervo2016] 19 Results: Spatial Visualization Filtered, 10 first ms, view from top 20 9

10 Results: Room parameters 21 Results Directivity on stage Up to 9 db of increase at high frequencies Worst case: High frequency contribution bigger fromstand than fromsource 6 db differences between conditions just by changing stand inclination 15º More directive sources: Possible shift in perceived direction (Precedence/Haas effect)? +/- 1.5 db at mid frequencies Increased early reflections up to 50 ms EDT affected (in some cases over JND) T30 not significantly affected C80 systematically increased over 500 Hz Between 3 and 6 db Perceivable by a player, but never payed attention to it Interview response during an experiment 22 10

11 Sound examples Convolved samples Stereo IR Anechoic recordings Mic at trumpet bell Schoeps CCM 4V Trumpet C1 (clean stage) C2 (front, directed) C5 (front-right, directed) Piccolo C1 (clean stage) C2 (front, directed) C5 (front-right, directed) 23 Summary The impact of reflections is very important Polar diagrams do not give a complete picture of radiation properties Frequency bands should be very fine or carefully selected for music instruments Acoustic cameras can be very helpful to investigate complex, dynamic sound radiation What to do if there are many acoustic centres? 25 of 14 11