Room Acoustic Measurements Using a High SPL Dodecahedron

Audio Engineering Society Convention Paper Presented at the 140th Convention 2016 June 4–7, Paris, France This paper was peer-reviewed as a complete manuscript for presentation at this convention. This paper is available in the AES E-Library (http://www.aes.org/e-lib) all rights reserved. Reproduction of this paper, or any portion thereof, is not permitted without direct permission from the Journal of the Audio Engineering Society. Room acoustic measurements using a high–SPL dodecahedron Dario D’Orazio1 , Simona De Cesaris1 , Paolo Guidorzi1 , Luca Barbaresi1 , Massimo Garai1 , and Roberto Magalotti2 1 DIN, 2 B&C University of Bologna, Bologna, Italy Speakers, Bagno a Ripoli, Italy Correspondence should be addressed to Dario D’Orazio (dario.dorazio@unibo.it) ABSTRACT In this paper a dodecahedron with high powered loudspeakers is presented. The source is designed to allow high SPL with very low distortion. By comparing the prototype with a reference sound source, the high SPL dodecahedron show a flat frequency response over the 80 ÷ 5000 Hz one third octave bands, enough to meet all the ISO 3382 criteria. Laboratory measurements have been performed to test the performances and the robustness of the dodecahedron using different techniques at different sound pressure levels and background noises. The prototype allows a good signal-to-noise ratio of the impulse response also when 75 dB of stationary noise is added during the measurements. 1 Introduction In the field of acoustic measurements it has been widely discussed if one technique is more performant than the other one, i.e. MLS versus Exponential Sine Sweep [1]. The performance of the compared methods is often analysed with regards to robustness in harmonic distortion of the loudspeaker or “ideal” noise spectra, rarely with regards to robustness against “real” background noise [2, 3, 4]. Recently Guski and Vorländer pointed out the disadvantage of the ESS measurements in presence of an impulsive noise [5]. They proposed an algorithm to detect an impulsive noise and, if necessary, repeat the measurement. In the authors’ experience the noise conditions are not predictable during the measurements, e.g. in historical opera houses, due to presence of people or HVAC noise [6]. The ESS techniques may be not immune from these kind of noise, due to complex statistic characteristics of the signals [7] and an hardware solution is needed. 2 Design The goal of the authors is to develop a dodecahedron for room acoustic measurements with a good robustness against the typical noises present in “real” environments. The dodecahedron has been designed using the following requirements, similar to other academic projects [8]: 1. good mechanical robustness, for in situ measurement procedures in the field of architectural and building acoustics; 2. total weight lower than 20 kg, for the needs of portability; D’Orazio et al. High SPL dodecahedron 3. enough sound power in the octave bands from 125 Hz to 4000 Hz, according to the requirements of ISO 3382 criteria [9]. A prototype of the dodecahedron has been built recycling the wooden shell of an old B&K sound source with 6.5 inch drivers. Custom loudspeaker drivers have been provided by Italian pro audio manufacturer B&C Speakers, based on the standard 6NDL38 model. The custom driver is a 170 mm diameter cone loudspeaker featuring a 38 mm diameter voice coil and a power handling of 150 W according to the AES2-2012 [10] standard. The magnet assembly is based on a neodymium ring magnet, resulting in a sensitivity of 92 dB 1W/1m. Low resonance frequency (72 Hz), rubber surround and large excursion capability (xmax = 6 mm) allow enhanced performance in the low frequency range, while the waterproof treatment makes it usable even in harsh environmental conditions. The prototype of the dodecahedron was assembled in September 2014 and has been used in the Opera House measurement campaigns done by the acoustic group of the Bologna University [11, 12]. More in detail, the prototype was set up in order to provide enough sound power during the measurements at Bayreuth Festspielhaus. When the sound source is placed in the orchestra pit of the Bayreuth Festspielhaus, the measured value of sound strength in the audience is very low (G < −5 dB): it may be regarded as the worst condition of source-receiver communication in the opera house acoustics. The prototype was able to measure impulse responses during these gravious conditions and so it has been called “Siegfried”, with reference to the Wagner’s Ring. Preliminary measurements tested the directivity the dodecahedron, according to ISO 10140 [13] and ISO 3382 [9]. Directivity has been tested placing the microphones at 1 m from the source, more strictly than the ISO 3382 recommendations, due to the configuration used in the Support criterion measurement. The dodecahedron show a sufficient omnidirectionality (see figure 1), compliant with ISO 3382, even if the microphone was kept at 1 m distance from the sound source while ISO 3382 requires a 1,5 m distance. It is worth noting that this aspect can be considered negligible when measuring standard room criteria if the measurements are done by stepwise rotating the source [14]. Fig. 1: The prototype “Siegfried” during its assembly (top) and placed in the Bayreuth Festspielhaus orchestra pit during the measurement sessions in September 2014 (bottom). 3 Methods In order to evaluate the performance, the prototype has been compared with a reference dodecahedron B&K Omnipower 4296 using two procedures: 1. Sound power level measurements in a reverberation room, according to ISO 3741[15]. Sound power levels have been extracted in onethird octave band from 50 to 5000 Hz. 2. Impulse responses have been measured in reverberation room, adding noise from another sound source during the exponential sweep sine. Sinusoidal signals and pink noise have been respectively added, spanning to 65 dB to 85 dB of SPL. Signal-to-noise ratios of impulse response and energy decays have been estimated. AES 140th Convention, Paris, France, 2016 June 4–7 Page 2 of 7 D’Orazio et al. High SPL dodecahedron 10 dB 5 0 −5 −10 250 500 1,000 2,000 4,000 Hz Fig. 2: Directivity of the prototypal dodecahedron (black line) with the upper and lower limits of directivity of ISO 10140[13] (dashed line) and ISO 3382[9] (dotted line) 250 Hz 500 Hz 1000 Hz 1 THD (%) 1.5 0.5 0 104 106 108 110 112 114 116 single tone SPL (dB) Fig. 3: Prototype “Siegfried” powered with Crown XLS 2500 in bridge mode. Measured values of THD versus sound pressure level measured at 1 m. Single tone excitation at 250 Hz (orange), 500 Hz (green), 1000 Hz (brown). Fig. 4: Setup for the measurements of impulse responses in presence of noise of the prototype “Siegfried” at the reverberation room of the University of Bologna, March 2016. 4 Experimental results Measurements have been done in the reverberation room of the Bologna University. Due to the volume of the room only measurements above 100 Hz could be allowed, according to ISO 3741. Nevertheless the measurements have been done down to third octave band of 50 Hz by averaging a large number of receiver positions. Both sound sources have been set to the maximum level without distortion (THD<1 %) or signal compressions on the signal path (A/D converter, power amplifier). A RME fireface 800 was used as D/A converter with enough headroom (>6 dB) during measurements. The prototype was powered by Crown XLS 2500 in bridge mode without additional equalization. The reference source B&K Omnipower 4296 was powered by power amplifier B&K 2716 in bridge mode. B&K 4190 microphones have been used with preamplifier B&K 2996 without further preamplification of the RME fireface 800. 4.1 Sound power level measurements The measured sound power levels of the reference sound source and the prototype “Siegfried” are shown in Fig. 5. The reference sound shows a noticeable rolloff at low frequencies below 100 Hz. This is probably due to the high resonance frequency of the mounted AES 140th Convention, Paris, France, 2016 June 4–7 Page 3 of 7 High SPL dodecahedron Sound power level (dB re 1 pW) D’Orazio et al. 120 100 80 60 reference 63 125 250 prototype 500 1,000 2,000 4,000 Hz Fig. 5: Comparison between the sound power level of the reference dodecahedron (blue) and the prototype “Siegfried” (red) measured in the reverberation room of the University of Bologna, according to ISO 3741. 5 inch loudspeakers, optimized for building and room acoustics measurements in the range 100÷ 5000 Hz. The prototype shows a frequency response flatter than the reference. Moreover in each third of octave the measured sound power levels are 10÷15 dB higher than the reference ones. It is important to note that the size and weight of the prototype are about twice as much as the ones of the reference sound source. 4.2 Robustness to stationary noises The measurement setup in the reverberation room consists of: 1. the sound source under test (prototype or reference), playing an exponential sweep sine without loudspeaker distortion (THD <1%); 2. an additional sound source, playing the additional noise in order to have a most diffuse noise field in the reverberation room; 3. monaural microphones in ten receiver positions; For every source-receiver combination seven impulse responses have been recorderd: the first one without adding noise, the other six with pink noise or 1 kHz sinusoidal signal. Figure 6 shows the time-frequency representations of the impulse response, highlighting the presence of the pink noise (figures 6(c), 6(d)) or the single tone (figures 6(e), 6(f)) during the measurements. The signal-to-noise ratio of each measurements has been evaluated for each configuration. The averaged values over all the source-receiver combinations are shown in fig. 7. The threshold of SNR = 40 dB may be considered the minimum requirement to extract room criteria values unaffected by the background noise using reference noise compensation algorithm (The compensated Schroeder method in the version presented by Guski and Vorländer [16]- method E: truncation,correction and subtraction - using the MATLAB ITA Toolbox [17]). 5 Discussion The experimental results suggest that there are some advantages when using an high–SPL dodecahedron. The comparison of sound power level measurements done in the reverberation room show that the use of professional loudspeakers in a dodecahedron may allow a flat emission over all the octave bands required by the ISO3382 criteria. Moreover, a high–SPL dodecahedron gives out more than 10 dB of SPL with respect to a commercial dodecahedron. In the authors’ AES 140th Convention, Paris, France, 2016 June 4–7 Page 4 of 7 D’Orazio et al. High SPL dodecahedron (a) reference sound source (b) prototype sound source (c) reference with 85 dB of pink noise (d) prototype with 85 dB of pink noise (e) reference with 85 dB of 1 kHz single tone (f) prototype with 85 dB of 1 kHz single tone (dB) 50 40 prototype reference 30 0 65 75 85 pink noise SPL (dB) impulse response SNR impulse response SNR (dB) Fig. 6: Time-frequency visualization of the impulse responses measured with prototype and reference sound sources with an incremental amount of noises during measurements (on the horizontal axis the time of the impulse response, on the vertical axis the frequency). 50 40 prototype reference 30 0 65 75 85 1 kHz single tone SPL (dB) Fig. 7: Analysis of the robustness against stationary noises. Averaged values of the signal-to-noise ratios of impulse response measured increasing the stationary noise SPL. The threshold of 40 dB may be considered a mininum requirement to extract room criteria values unaffected by the background noise, as explained in the text. AES 140th Convention, Paris, France, 2016 June 4–7 Page 5 of 7 D’Orazio et al. High SPL dodecahedron experience this possibility allow to get reliable room acoustic measurements also in presence of extraneous noise. Previous literature discussed the performance of the ESS technique in presence of impulsive noise during the measurements [5]. In the present work the ESS technique has been tested in presence of stationary noise (pink noise or single tone), simulating real conditions. The results (see figure 7) allow to evaluate the useful range of the impulse response for the extraction of the room criteria. Using reference methods of background noise compensation [17, 16] 40 dB of signal-to-noise is the minimum requirement to extract room criteria values unaffected by the background noise. Further studies can be done measuring other kind of stationary noise (e.g. modulated narrow band noise at low frequency) or using different techniques to extract the envelope from impulse responses [18, 19]. 6 Summary The meaning of this work is to overcome a disadvantage of sweep measurements, its sensitivity to environmental noise. A hardware solution is proposed: a high SPL dodecahedron able to cover the octave bands from 125 Hz to 4000 Hz. A prototype of the dodecahedron has been built and its directivity has been tested according to ISO 3382 requirements. The prototype has been compared with a reference sound source (B&K 4296) in laboratory measurements. Sound power levels of both sources have been measured in a reverberation room, according to ISO 3741: the frequency response of the prototype is more linear than the reference one and has an extended bass response down to 50 Hz. The robustness against background noises has been tested in a reverberation room adding noise during the sweep recording. Pink noise and sinusoidal signals have been added, at increasing SPL. To conclude, the prototype allows a good signalto-noise ratio of the impulse response also when 75 dB of noise is added during the measurements. In the same condition the reference sound source doesn’t allow an adequate decay of the impulse response. References [1] G. Stan, J. J. Embrechts, D. Archambeau, Comparison of Different Impulse Response Measurement Techniques., J. Audio Eng. Soc., 50(4) 2002, p. 249–262 [2] A. Farina, Advancements in Impulse Response Measurements by Sine Sweeps. 122nd Convention of the Audio Engineering Society, 2007, paper 7121. [3] A. Torras-Rosell, F. Jacobsen, A New Interpretation of Distortion Artifacts in Sweep Measurements, J. Audio Eng. Soc., 59(5) 2011 p. 283– 289. [4] P. Dietrich, M. Guski, M. Vorländer, Influence of Loudspeaker distortion on Room Acoustic Parameters, Proc. of 40th Italian (AIA) Annual Conference on Acoustics and the 39th German Annual Conference on Acoustics (DAGA), 2013. [5] M. Guski, M. Vorländer, Impulsive Noise Detection in Sweep Measurements, Acta Acustica united with Acustica, 101 (2015) 723–730. [6] N. Moriya and Y. Kaneda, Impulse response measurement that maximizes signal-to-noise ratio against ambient noise, Acoust. Sci. & Tech. 28, 1 (2007). [7] P. Guidorzi, L. Barbaresi, D. D’Orazio, M. Garai, Impulse responses measured with MLS or SweptSine signals applied to architectural acoustics: an in-depth analysis of the two methods and some case studies of measurements inside theaters, Energy Procedia, 78, 1611–1616 (2015). [8] M. Horvat, H. Domitrovic, S. Grubesa, Design of a new omni-directional sound source, 3rd Congress of the Alps Adria Acoustics Association 27–28 September 2007, Graz (Austria). [9] ISO 3382-1:2009 Acoustics – Measurement of room acoustic parameters – Part 1: Performance spaces. International Organization for Standardization (2009). [10] AES2-2012: AES standard for acoustics - Methods of measuring and specifying the performance of loudspeakers for professional applications Drive units. [11] M. Garai, F. Morandi, D. D’Orazio, S. De Cesaris, L. Loreti, Acoustic measurements in eleven Italian opera houses: Correlations between room criteria and considerations on the local evolution of a typology, Build. Environ., 94(2), 2015, 900– 912. AES 140th Convention, Paris, France, 2016 June 4–7 Page 6 of 7 D’Orazio et al. High SPL dodecahedron [12] M. Garai, K. Ito, D. D’Orazio, S. De Cesaris, F. Morandi The acoustics of Bayreuth Festspielhaus, 22th International Congress of Sound and Vibrations, Florence, July 2014. [13] ISO 10140-1:2010 Acoustics – Laboratory measurement of sound insulation of building elements – Part 1: Application rules for specific products. International Organization for Standardization (2010). [14] F. Martellotta, Optimizing stepwise rotation of dodecahedron sound source to improve the accuracy of room acoustic measures, J. Acoust. Soc. Am., 09/2013; 134(3), 2037–48. [15] ISO 3741:2010 Acoustics – Determination of sound power levels and sound energy levels of noise sources using sound pressure – Precision methods for reverberation test rooms. International Organization for Standardization (2010). [16] M. Guski, M. Vorländer: Comparison of noise compensation methods for room acoustic impulse response evaluations. Acta Acustica united with Acustica 100 (2) (2014) 320–327. [17] P. Dietrich, M. Guski, M. Pollow, B. Masiero, M. Müller-Trapet, R. Scharrer, and M. Vorländer, ITA-Toolbox – An Open Source MATLAB Toolbox for Acousticians, in 38th German Annual Conference on Acoustics (DAGA), Darmstadt, Germany, March 2012. [18] D. D’Orazio, S. De Cesaris, M. Garai, Measuring reverberation time using preprocessed energy detection, Proc. of Internoise 2012, New York City, 18.08.2012. [19] S. De Cesaris, D. D’Orazio, F. Morandi, M. Garai, Extraction of the envelope from impulse responses using pre-processed energy detection, J. Acoust. Soc. Am., 138, 2513 (2015). AES 140th Convention, Paris, France, 2016 June 4–7 Page 7 of 7