Sound recording with the application of microphone arrays

Transcription

1 Coputer Applications in Electrical Engineering Sound recording with the application of icrophone arrays Eugeniusz Kornatowski West Poeranian University of Technology 7-26 Szczecin, 26 Kwietnia, e-ail: In this article the issues concerning sound recording with the use of three-diensional systes of several icrophones were considered. The issues under study concern the so called beaforing, which is odeling three-diensional directivity patterns of icrophone arrays, as well as odern technologies of ultichannel recording production with the purpose of reproduction in surround sound systes.. Introduction Modern technology of sound recording is, in ost cases, based on ultiicrophone technology. For this purpose the icrophones with specific directivity patterns are used: onidirectional (donut shaped), figure 8 and cardioid (including the following subgroups: sub-, super-, hyper- cardioid). Application of a single icrophone - with a definite directivity pattern - to sound recording enables full control of the recording: in all probability we can predict the agnitude and geoetry of an environent in which the sound eitted by particular sources will be successfully recorded. However, while applying the ulti-icrophone technology the final result is often difficult to predict. In such cases, in sound engineering, the ost popular ethod consists in conducting several test recordings while changing the geoetry of the icrophone syste. It is still ore coplicated when a produced recording is supposed to be reproduced on a surround sound syste, e.g. 5.. In such a case, apart fro fulfilling the requireent of recording the sound fro particular sources with high quality, the possibility of apparent sound source planar localization during ultichannel reproduction should also be ensured. Bearing in ind all these considerations, it can be claied that an analytic device enabling siulative icrophone syste (icrophone array) configuration would be extreely useful. In the literature this process is often referred to as beaforing [, 2]. 2. Directivity patterns of icrophone The ost iportant properties of icrophones are deterined by two characteristics: the sensitivity and the directivity pattern. The sensitivity at a specific frequency is the ratio of the voltage at the terinals (output) of a 23

2 E. Kornatowski / Sound recording with the application of icrophone arrays icrophone loaded with noinal ipedance to the acoustic pressure in the place of location of the icrophone [3]: U V S u, p () Pa The directional properties of a icrophone are deterined by the ratio its of sensitivity at any direction of incidence of a sound wave on the icrophone to the sensitivity at a perpendicular direction of incidence on the eleent receiving the acoustic energy. The range of this ratio in the function of incidence angle of wave is called the directivity pattern [3]. In siulation testing the ost frequently odelled pattern is the directivity pattern of a single icrophone as a difference of the characteristics of an onidirectional (pressure) icrophone and a figure-8 pattern (pressure gradient) icrophone, located fro one another at a distance of l. If is the delay between signals of the two icrophones, f deterines the frequency and direction (aziuth angle) fro which the wave plane reaches both icrophones, then the directivity pattern of the icrophone odelled is expressed by the following equation [4]: l cosθ U(f,θ) A exp j 2 Π f τ (2) c where: A,, c speed at which sound travels in the air is equal to 34 /s. Depending on coefficient A and delay tie it is possible to siulate a directivity pattern fro an onidirectional to a figure-8 pattern. Figure shows an exaple pattern of a icrophone with a cardioidal pattern. In Figure (b) the axes are diensionless. Every point of the surface presented represents the sensitivity of the icrophone along direction of a surface point to the centre of the co-ordinate syste. Modelling of directivity patterns of arrays (atrices) of icrophones in general can be perfored for a near field or for a far field. In case of a near field it is necessary to take into account the distance of a sound source fro individual icrophones during calculations. For the far field it is assued that the distances between icrophones in the array are uch saller than the distance of geoetrical centre point of the icrophone array fro the sound source; the front of acoustic wave is flat. Furtherore, it is assued that the far field case is considered. For an array of icrophones located on a horizontal surface, the directivity pattern of such an array can be described using the following equation: 24

3 E. Kornatowski / Sound recording with the application of icrophone arrays U M (f, θ) U f, θ α l exp j 2 Π f c cos( θ θ) (3) a) b) Fig.. Directional characteristics of cardioidal icrophone for one frequency of Hz: (a) polar plot, (b) 3-D directivity pattern, as a function of elevation and aziuth angle 25

4 E. Kornatowski / Sound recording with the application of icrophone arrays Forula (3) enables the deterination of a directivity pattern U M for a planar array of icrophones, taking into account the directivity patterns U of the icrophones foring the pattern. The position of each icrophone is described by the distance l fro the centre of the XY co-ordinate grid and the angle of direction fro the icrophone to the sound source. Angles are the angles of rotation around the axis of each icrophone and are easured in relation to the positive part of axis X in a counter-clockwise direction. The forula (3) can be generalized for the case of a 3-D array including a nuber of icrophones, then: U (f, θ, φ) U f, θ α, φ β M l (4) exp j 2 Π f cos( θ θ) cos( φ φ) c where:, angles of aziuth and elevation of icrophone,, angles of aziuth and elevation of direction icrophone sound source. 3. Experiental and siulation tests Tests were perfored using the Atos icrophone syste coprising Brauner s five icrophones VM ounted on five-ared planar stand ASM5. The Atos syste is designed to ake recordings dedicated to surround sound systes 5. (5.). The icrophones, all having cardioidal pattern, record the following signals: C front center (center), LF and RF front left and front right, LR and RR rear left and rear right respectively. Each icrophone can be rotated around its axis within the range +/ 9 in relation to the axis of a atching stand ar. The construction of the array is syetrical to the front-rear axis and its geoetry is shown in Fig Fig. 2. Geoetry of stand ASM5 with icrophones VM of Atos syste

5 E. Kornatowski / Sound recording with the application of icrophone arrays The producer recoends that during the recording the axes of the icrophones axiu sensitivity should overlay the axes of respective ars of the stand (standard set-up). The Atos array was used several ties to do test recordings of concerts perfored by the syphony orchestra at the Philharonic in Szczecin. The signal was recorded by the Zaxco Deva digital recorder with the resolution of 24 bits and the sapling frequency of 96 khz. During the test the axes of the axiu sensitivity of each icrophone overlaid the axes of respective ars of the stand (standard set-up). During the playback of the perforance recorded in the way already described, it turned out that, despite loudspeaker systes set-up fulfilling the standard ITU-R- BS.775-, in the front sound stage an inaccurate apparent sound source localization could be noticed, with the clear doination of the background (rear channels) and center channel. It can be assued that the reason for this negative effect lies solely in the wrong set-up of angles of particular icrophones. The quality of the icrophones theselves is beyond question the VM icrophones are aong the best studio icrophones in the world. In order to possibly validate this hypothesis, siulation tests were conducted using the ethod described in Section 2 with the assuption that the sound source is located far fro the icrophone. Hence the siulation tests were conducted for the so called far field. a) b) Fig. 3. Atos syste: a) syste set-up in Studio S in Polish Radio in Szczecin, b) recording event at the Philharonic in Szczecin 27

6 E. Kornatowski / Sound recording with the application of icrophone arrays The directivity pattern of the Atos syste with standard set-up has a shape as shown in Fig. 4. a) 2 Back Left Right Front b) Fig. 4. Directivity pattern of Atos syste with icrophones with standard set-up angles for the frequency f =5 Hz: a) 3-D graph, b) directivity pattern in the XY plane In the 3-D graph the axes are diensionless. Each point of the surface presented represents the effectiveness of array along the direction: surface point the center of the coordinate syste XYZ. The obtained patterns substantiated the observations related to degradation of the front sound stage. The directivity pattern

7 E. Kornatowski / Sound recording with the application of icrophone arrays is very uneven (far fro onidirectional) especially in the area of the front sound stage. The following test phase consisted in searching for optial angles of rotation of particular icrophones. While changing these angles, with the use of nuerical odeling, such set-ups were sought which would provide axially onidirectional directivity pattern, especially in the front sound stage. The tests yielded the following conclusion: the LF icrophone should be turned through an angle of +7 in relation to the standard set-up (clockwise direction of rotation), the RF icrophone -7, the icrophones C, LR and RR should be left as in the standard set-up. The pattern obtained is presented in Fig. 5. a) 2 Back Left Right Front b) Fig. 5. Directivity pattern of Atos syste after the adjustent of angles for the frequency f =5 Hz: a) 3-D graph, b) directivity pattern in the XY plane 29

8 E. Kornatowski / Sound recording with the application of icrophone arrays Having altered the set-up of the icrophones, the recording session in the concert hall took place one ore tie. Subjective readability and the localization of apparent sound sources in the front sound stage was considerably enhanced in coparison with the standard set-up of the icrophones. The recordings were subsequently subitted for subjective verification to the Laboratory of Sound Engineering and Abiophonics at the Faculty of Electrical Engineering, West Poeranian University of Technology in Szczecin Fig. 6. Fig. 6. Laboratory of Sound Engineering and Abiophonics at the Faculty of Electrical Engineering, West Poeranian University of Technology in Szczecin The tests of that kind are coonly applied. Due to their vital iportance in quality assessent of ultichannel signals and systes of recording and surround sound playback, the procedure and conditions under which the tests are conducted are standardized [5, 6]. The test - concerning a subjective assessent of the Atos syste after its optiization was conducted with 28 participants, listeners-experts. The participants task was to copare the quality of recording saples obtained using the Atos syste with the icrophones set-up both in a standard and optiized way. The assessent concerned the following paraeters:. The sound quality of front channels understood as: stability of front sound iage, width of front sound stage, precision of apparent sound source localization, sense of appropriate localization of sound sources depending on the type and character of a recorded event (e.g. instruent groups localization during the playback of syphony orchestra concert being in line with expectations). 2. The sound quality of rear channels understood as: stability of rear sound iage (analogous to Point.), coherence of sound space (no feeling of void in certain spot, e.g. directly behind the listener, in the area of rear sound space), 22

9 E. Kornatowski / Sound recording with the application of icrophone arrays appropriate arrangeent and localization of sound sources in the space. 3. The sense of spatiality, which is: feeling of the size of the place, appropriate length of reverberation tie for a given event conditions (for a given space), realis of sound space, sense of presence : feeling of being in a place where a recording was done as a easure of sound realis, ratio of direct sounds to reflected sounds: if reflected sounds doinate, the sense of the so called artificial acoustic perspective arises. 4. Clarity (lucidity), which is: speech clarity, ability to identify and differentiate between voices and sounds to be heard siultaneously, separation of individual short sounds occurring in short tie intervals. 5. Balance: dynaics of a played back track (appropriate to the nature of an acoustic event), loudness ratio between front and rear channels. 6. Overall assessent: subjective assessent of a recording coprising forerly described paraeters as well as the quality of the recording as a whole and general ipression the recording ade on the listener. The listener-expert assessent consisted in copleting a questionnaire for and arking individual paraeters fro (bad) to 6 (excellent). The results of the tests proved (Fig. 7) that the optiization of the icrophones set-up was especially beneficial to the quality of the front sound stage. Furtherore, the feeling of spatiality of played back recordings was greatly enhanced. And the balance of the sound surrounding the listener was highly arked Pa r a e t e r n u b e r St andar d Af t er opt i i zat i on Fig. 7. Test results for the Atos syste before and after optiization of rotation angles of icrophones 22

10 E. Kornatowski / Sound recording with the application of icrophone arrays The tests in question concerned the case of recordings conducted in a large concert hall. It can be claied with all probability that for recording in spaces of sall cubic capacity and a sall nuber of sound eitting objects, an optial configuration of the Atos syste will be undoubtedly different fro the one presented above. In such a case, siulation tests of patterns should be conducted for the so called near field [7]. References [] Van Trees H.L.: Optiu Array Processing, John Wiley & Sons, New York, 22. [2] Raichel D.: The Science and Applications of Acoustics, Springer, New York, 26. [3] Leighton T.G. : The Acoustic Bubble, Acadeic Press, San Diego, 999. [4] Jaroz A.: The Design and Use of a Double Cardioid Stereophonic Microphone, Journal of the Audio Engineering Society, vol. 8 No. 2, 96, pp. 4. [5] European Broadcasting Union Tech E: Assessent ethods for subjective evaluation of the quality of surround prograe aterial Music, EBU Official Technical Texts, Geneva, 997. [6] European Broadcasting Union Suppleent to Tech. 3286: Assessent ethods for the subjective evaluation of the quality of sound prograe aterial Multichannel, EBU Official Technical Texts, Geneva, 2. [7] Chen H., Ser W., Yu Z.: Optial design of near field wideband beaforers robust against errors in icrophone array characteristics, IEEE Transactions on Circuits and Systes, No. 54(9), 27, pp