Interactive auralization of powertrain sounds using measured and simulated excitation

Transcription

1 7NVC-162 Interactive auralization of powertrain sounds using measured and simulated excitation Daniel Riemann and Roland Sottek HEAD acoustics GmbH, Herzogenrath, Germany Tadakazu Naritomi, Shin Kishita and Akira Yamada Nissan Motor Co.; LTD., Yokohama, Japan Copyright 27 SAE International ABSTRACT Interior vehicle sound is an important factor for customer satisfaction. To achieve an optimized product sound at an early stage of development, subjective evaluation methods as well as analysis and prediction tools must be combined to provide reliable information relevant to product quality and comfort judgments. Binaural Transfer Path Synthesis (BTPS) is a wellknown method to calculate interior noise and vibrations based on multi-channel input measurements. Recent enhancements of the BTPS method enable taking into account also simulated excitations, for example engine mount vibrations calculated using MBS and/or FEM simulations, allowing the prediction of interior noise even if the engine is not available in hardware. Interactive evaluation of the generated sounds in a vibro-acoustic driving simulator helps to increase understanding of customer responses and perception of target sounds. This paper describes the engineering tools and methods developed for vehicle product sound design and comfort judgments by means of a BTPS model of a Nissan vehicle. INTRODUCTION Based on the demands of the automobile industry, Binaural Transfer Path Analysis and Synthesis (BTPA/BTPS) techniques have been developed for predicting vehicle interior sound. These methods allow not only predicting order levels and spectra, but also binaurally auralizing various driving conditions, as well as predicting vibrations at the driver s seat and steering wheel for interactive evaluation in the acoustic driving simulator H3S. The synthesis procedure is based on input measurements in the engine compartment of a car or on an engine test rig. During the development process of new engines, prototypes are rarely available. In addition, it is important to have as much information as possible about NVH issues of the engine in the early design phase. Therefore, recent extensions of BTPA and BTPS concern the use of simulated excitation data for airborne and structure-borne contributions. The methodology also allows combining measured and simulated excitation data. Acoustic simulation software permits predicting level and phase values of sound pressure or vibration, generally for steady-state conditions at certain rpm values. These data cannot be used directly for the time-domain approach of BTPS, which is essential for auralization. Therefore, a suitable order generator has been developed, taking into account adequate phase synchronization of all input data to ensure a correct summation of the contributions from the different sources. BINAURAL TRANSFER PATH ANALYSIS AND SYNTHESIS Binaural Transfer Path Analysis (BTPA) was originally developed for assessing the binaural contributions of individual vehicle noise paths [1]. It is a powerful modeling tool enabling engineers to identify causative mechanisms for noise transfer by distinguishing between excitation source strengths and the transfer behavior of individual elements. The method is well-known for powertrain noise analysis, including both point-coupled structure-borne and identified airborne source paths to a receiver location. MEASUREMENT OF TRANSFER FUNCTIONS For structure-borne sound, transfer functions from engine mount to each ear are measured for each individual noise path (Fig. 1). For airborne noise, acoustic transfer functions are similarly determined. For the latter, all relevant sound sources must be known,

2 especially because no discrete reference points exist as they do in the case of structure-borne sound. Fig. 1: Structure-borne noise path: F Engine Mount Body Airborne noise path: Vibrating surface Q reciprocal measurement Transmission reciprocal measurement p Ear p Ear Transfer functions measured by BTPA. Reciprocal measurements The transmission of airborne and structure-borne sounds to the driver's ears can be measured reciprocally using a recently-developed binaural volume velocity transducer in the driver s seat (Fig. 2) [2] as well as microphones or accelerometers at the noise source positions. BTPS MODEL Binaural Transfer Path Synthesis (BTPS) is the process of creating listenable vehicle interior noise data based on a BTPA model or modifications of it (Fig. 3). For engine and transmission mounts, vibration signals in all three directions (x, y, z) are considered. Each individual path or combination of paths can be auditioned independently, to assess their respective impact on the overall sound quality. Paths can be modified to simulate countermeasures and their effects on the interior noise. Operating data measured on different sources, such as engine test-rig data, may also be put into the model to predict how interior noise is influenced by the different sources [5]. Structure-borne noise Left X Left Y Left Z Right X. Engine + mount. Airborne noise Body. + P Ear Engine Transmission Source 1 Source 2... Fig. 2: Binaural sound source with low-frequency cabinet (for frequencies below 2 Hz). Each ear contains an electrodynamic horn loudspeaker. The geometry of the head is quite similar to the artificial head used for interior noise measurements. The directional patterns of source and receiver are the same, the basic requirement to apply the reciprocity method [3]. The advantages of taking reciprocal measurements of acoustic transfer functions are evident: First, significant time is saved, since all paths can be measured simultaneously. Second, less space is required for sensors than for sources. Thus the measurement positions can be chosen almost without restriction, leading to higher accuracy. Fig. 3: Binaural synthesis based on airborne and structure-borne sound transfer paths Input data for the binaural interior noise synthesis are the measured airborne and structure-borne excitations. The airborne radiation of the engine is acquired with several microphones around the engine and at the intake and exhaust system. The structure-borne excitations are measured with triaxial sensors at the attachment points of the engine mounts. All input signals are measured time-synchronously in one measurement to ensure accurate phase relationship between the transfer paths. In order to combine several measurements or include simulation data, suitable time synchronization is essential. Extensions of methodology using four-pole technology Recent improvements to the BTPA/BTPS methodology involve new techniques for replacing intricate timeconsuming measurements with simulations using computer models [4].

3 METHODS FOR ORDER MODIFICATION There are different methods for order modification. Wellknown methods involve order filtering using tracking filters. Filter parameters such as the center frequency vary as a function of rpm. The accuracy depends on the rpm resolution and on the update-rate of the filter parameters. Very fast changes of filter parameters may cause audible noise due to transient effects. Of course, in any case abrupt changes of amplification (level changes of several decibels) should be avoided. But also, very fast changes of center frequency are difficult to handle. In addition, there is always a trade-off between the resolutions in the frequency and time domains. Narrow filter bandwidths mean longer transient responses of the filter. Thus, it is almost impossible to modify a single order without influencing other orders (or partial orders) in its vicinity. Despite the described difficulties, order filtering is applicable in many cases, to reduce or increase certain engine orders of a given sound recording. But in the case of BTPS, where it is necessary to generate time data with a certain phase relation to other measured or simulated data, the common tracking filter is not suitable. Normally, minimum-phase filters are used: their phase depends on their amplitude characteristic and cannot be adjusted independently. For a single path this is not of great importance, but for the sum of several paths the resulting processing errors due to erroneous phase relations may become very high. To correct the situation, so-called FIR (finite impulse response) filters could be used. They allow for independent control of magnitude and phase. But it is difficult to update the parameters of FIR filters continuously. Normally, averaged rpm values are used for order tracking (Fig. 4), 2nd Order 12-3 rpm 5m.1 t/s p/pa nd Order 12-3 rpm signal vs.rotation p/pa Cycle/cycle: cycle leading to unwanted order-level fluctuations due to the mismatch of actual signal frequency and center frequency according to the frequency-dependent characteristic of the order filter. Fig. 4: Instantaneous rpm curve and averaged curve Therefore a method based on order analysis and synthesis using a high rpm resolution and a non-linear resampling technique is proposed for order modification, to overcome the described drawbacks of order filtering. Fig. 5 illustrates the effect of the resampling technique. In this example, the sweep signal with constant level and varying frequency is compressed and stretched, controlled by the instantaneous (not averaged) rpm value in order to obtain the correct order level and phase values (phase is not shown in Fig. 5). 2nd Order 12-3 rpm.fft 5m 75m.1 t/s L/dB[SPL] 8 82 time signal vs rotation.fft O/Order cycle t/s L/dB[SPL] Fig. 5: Order analysis using a resampling technique. The left diagrams show the time curve of a sweep signal (upper) and the resampled signal on an angular axis (lower); the right diagrams show the corresponding spectrograms.

4 ORDER ANALYSIS AND SYNTHESIS The following diagrams illustrate the entire process of order analysis, modification and synthesis, starting with the recorded audio signal (for example from the intake) (Fig. 6, left). After applying a nonlinear resampling technique we obtain the pressure curve for each camshaft cycle on a linear angle scale (Fig. 6, right). By means of the fourier transform the order spectrum can be calculated. The resolution of the order spectrum shown in Fig. 7 (left) is very high, because of the improved resampling technique in combination with a high angular resolution based on the high pulse rate of the crankshaft sensor signal provided by the engine control unit. Fig. 7 (right) shows, as an example, the magnitude and (unwrapped) phase curves of the 2 nd order of the intake signal. These analyses form the basis for further modifications, such as changes of order levels and phase values, or exchange of orders with results from simulation. Normally, the results of simulation programs (like GT- Power) are magnitude and phase values as a function of rpm. Thus another mapping process from camshaft angle to rpm and vice versa is required. Fig. 8 (left) represents the result of such a mapping. In Fig. 8 (right) a simple modification is applied, here for demonstration purposes only. Of course other level and phase vs. rpm curves could be used. Abrupt level changes should be avoided because of audible transient effects. The next processing step realizes the transform from rpm to (Fig. 9, left). For each camshaft cycle the level and phase of the modified orders must be adapted. The modified values can be seen in the order spectrogram (Fig. 9, right). By means of the Inverse Fourier Transform the order spectrogram can be converted to a pressure signal vs. camshaft angle, Fig. 1 (left), containing all the modifications applied to the different orders. The final processing step results in a modified audible sound file using a nonlinear mapping from camshaft angle to time, as shown in Fig. 1 (right). The results show no audible noise due to errors caused by the signal-processing steps Pa Pa Fig. 6: s Audio signal (left) and result of nonlinear mapping to the camshaft angle (right) Magnitude of 2 nd order in db order Phase of 2 nd order in rad Fig. 7: High resolution order spectrogram (left) and magnitude and phase curves of the 2 nd order of the intake signal as a function of (right)

5 115 Magnitude of 2 nd order in db 115 Magnitude of 2 nd order in db rpm Phase of 2 nd order in rad rpm Phase of 2 nd order in rad rpm rpm Fig. 8: Magnitude and phase curves of the 2 nd order as a function of rpm (left) and modified 2 nd order, 4dB level reduction (right) Magnitude of 2 nd order in db Phase of 2 nd order in rad order Fig. 9: Magnitude and phase curves of the modified 2 nd order as a function of (left) and modified high resolution order spectrogram (right) Pa Pa s Fig. 1: Modified pressure signal as a function of the camshaft angle (left) and corresponding audio signal as the result of the nonlinear mapping from camshaft angle to time (right)

6 ACOUSTIC DRIVING SIMULATION SYSTEM The first subjective sound evaluations took place in laboratories using playback via loudspeakers or headphones. Jury tests on interior car noise have shown that to achieve realistic results, it is not sufficient to present only the original recorded airborne sound via headphones to the evaluator sitting in an office. It is necessary to present the sounds in an environment where the tester feels immersed in the actual situation of driving. Furthermore, realistic reproduction of vibrations at the seat and the steering wheel helps to improve the perceived user realism. Therefore the SoundCar playback system has been developed as a realistic playback environment for the reproduction of vehicle interior sounds. It utilizes a carefully-equalized loudspeaker system including a powerful subwoofer for the airborne sound reproduction. The seat as well as the steering wheel can be excited using electrodynamic shakers. The virtual vehicle cabin playback environment is shown in Fig. 11. It combines real and virtual sounds. The real sounds could be from existing vehicle cabin components such as the radio or dashboard warning chimes. Virtual sounds are those directly related to motion and the physical driving process. Fig. 11: Audio simulation control concept and SoundCar playback environment.

7 INTERACTIVE SOUND SIMULATION SYSTEM (H3S) An interactive driving simulation requires on-demand reproduction of vehicle sound and vibration components. Independent reproduction of required vehicle sounds is achieved through playback of an indexed and catalogued database. Wind and tire noise are primarily dependent upon speed, while engine noise depends upon the simultaneous variables of RPM and load. A basic vehicle sound simulation should contain, at minimum, the following sound components: Engine sound, dependent on engine type, rpm, and load Tire noise, dependent on tire model, vehicle speed, and road conditions Wind noise, dependent on air and vehicle speed Special sounds such as individual road bumps, dependent on speed Virtually moving sounds produced by other vehicles; dependent on direction, distance and speed relative to the driven vehicle Background sounds Instructions intended to be provided to the driver Sound Database for engine, tire and wind noise The engine, tire and wind noise database consists of recorded vehicle interior noise and vibration. Alternatively, synthesized interior noise and vibration from BTPS can be used. In the latter case it is also possible to interactively evaluate the influence of individual transfer paths. basic functionality of the interactive driving simulation system. Transferpath 1 Transferpath 2 Transferpath N Fig. 12: Combination of simulation system and SoundCar for complete virtual auditory environment. H3S MOBILE Speed, rpm, Load Sound Database Filter Filter Filter Airborne and Vibrations Engine Road/Tire Wind Pass By Background Airborne Driving Dynamic H3S mobile is the most recent sound simulation vehicle development (Fig. 13). It can be used under actual (fully realistic) driving conditions. The driving dynamics come from an actual vehicle drive. Filter Filter Filter + Throttle, Brake, Gear In order to achieve realistic playback of the powertrainrelated interior noise and vibrations, the database must include the complete load and rpm range of the engine. Run-up measurements from idle to maximum rpm (for example 7 rpm-65 rpm) for several load conditions (such as %, 3%, 6%, 1%, 15%, 2%, 3%, 4%, 5%, 75%, 1% throttle position) with a duration of at least 12 seconds each have proven to be suitable. For the tire and wind noise a coastdown from maximum vehicle speed to standstill is used. A realistic separation of tire and wind noise can be achieved via speeddependent FIR filters. In order to calculate such filters, the structure-borne excitation as well as the airborne noise radiation of all four tires is measured during the coastdown measurement. By using multiple coherence analysis, speed-dependent FIR filters can be calculated representing the separated tire and wind noise shares of the interior noise. The H3S software continuously generates the actual interior noise and vibrations controlled by the engine load and rpm and the vehicle speed. These values are calculated using a driving-dynamics model taking into account the user s requests (throttle/brake position, gear, etc.). The SoundCar playback system described above is used to provide a realistic playback of the generated sounds and vibrations. Fig. 12 illustrates the Fig. 13: H3S mobile The H3S hardware and software are placed in the trunk. The sound environment is generated and presented in real time via equalized headphones with additional ANR (Active Noise Reduction). In the H3S mobile it is possible to modify the acoustic behavior of the car, i.e. to change the engine, tire and/or wind noise while driving; however, the vibration cannot be changed. Therefore a car without conspicuous vibration should be used. Nevertheless, with H3S mobile, the most realistic playback is possible. This is essential for obtaining reliable data for sound evaluation [6].

8 APPLICATION EXAMPLES In the following sections two application examples using the new order-synthesis technique are described. INTERIOR NOISE SYNTHESIS In this example a sub-compact front-wheel-driven car was investigated. The interior noise was synthesized using the above-described BTPS methodology for the complete load and rpm range (%-1% throttle using 1 load steps, from idle condition to maximum engine rpm). Thus it was possible to calculate a complete dataset for the acoustic driving simulation system described in the previous section and to evaluate the interior noise interactively. The vehicle has a common engine installation using a pendulum suspension with the engine hanging in left and right engine mounts that are attached to the main front longitudinal body beams. A torque rod is introduced t o handle most of the static wind-up of the power unit during load. The dynamic rotation of the engine due to torque pulsations will result in forces transmitted through the main engine mounts as well as the torque rod. For the interior noise synthesis the structure-borne vibration transmitted through the engine mounts and the torque rod was taken into account for three directions (x-, y- and z-direction: front-to-rear, left-to-right and bottom-to-top, respectively). Furthermore additional structure-borne paths considered in the BTPS model were the left and right drive shaft as well as the air conditioning refrigerant pipe. Thus the model includes a total of 18 structure-borne transfer paths. The airborne noise radiation of the engine was acquired by 8 microphones around the engine and one microphone each at the intake orifice and the tailpipe, resulting in 1 airborne transfer paths. Measured Interior Noise 2 n/rpm L/dB(A)[SPL] Synthesized Interior Noise 2 n/rpm L/dB(A)[SPL] 6 7 Fig. 14: Spectrograms of the left artificial head channel of the measured (left) and synthesized (right) interior noise Fig. 14 shows the comparison of the left artificial head channel of the measured (left diagram) and the synthesized interior noise (right diagram) for a runup measurement (WOT, duration: 3s). The diagrams indicate good agreement, especially with respect to the noise pattern mainly characterizing the interior noise (2 nd and 4 th engine order, and half-orders leading to engine roughness). Further investigations with engine modifications ( for example a revised intake orifice) have proven the validity of the synthesis model. Therefore the described BTPS model covers the most relevant transfer paths and is suitable to evaluate the interior noise of the vehicle. VIBRATION DATA FROM EXCITATION Based on the existing BTPS model, the interior noise of a modified engine built into th e investigated vehicle could have been predicted. But since this engine was not available in hardware, the engine mount vibrations were simulated using Excite. The airborne noise radiation was assumed to remain unchanged due to the implemented modification. The order level and phase data for the half engine orders from.5 th order to 8.5 th order were simulated for steady-state conditions each 1 rpm from 1 to 6 rpm. Using the above-described order-synthesis technique these order level and phase data were used to calculate new input data for the interior noise synthesis. Measured Torque Mount Vibration n/rpm L/dB[m/(s^2)] Fig. 15: Measured torque mount vibration in x-direction Fig. 15 shows the measured torque mount vibration in x- direction of the original engine. This original input measurement was used to synthesize the new input data of the modified engine. Therefore the simulated order level and phase data were imported from Excite into the BTPS Software tool. Using order synthesis, the original order level and phase data were replaced by the simulated data. 5 2

9 Predicted result for new engine Input measurement Original engine Order synthesis Import Simulated order level and phase data e.g. order.5 to 8.5 Simulated vibrations New engine Fig. 16: General procedure of calculating new input time simulations Fig. 16 illustrates the general procedure of calculating new input data for the BTPS using order synthesis. The measured input time signal for a certain input channel (here: torque mount vibration, x-direction) is used as the basis for calculation since the engine cycles vs. time characteristic is identical to the other input channels. The original order level and phase data are then replaced by the simulated order level and phase data, resulting in a new input time signal. This new signal is time-synchronous to the other measured input signals. Therefore it can be used for the interior noise synthesis. data for BTPS using a combination of measurements and Synthesized Torque Mount Vibration The result of the order synthesis for the torque mount vibration in x-direction is shown in Fig. 17. This new input time signal is free of any audible artifacts and thus particularly suitable for auralization purposes. The synthesized interior noise contribution of this individual transfer path (torque mount x-direction) is shown in Fig. 18. The same procedure was carried out for the left and right engine mount and the torque mount, always in x-, y- and z-direction. The synthesized interior noise share of the sum of these 9 transfer paths is shown in Fig n/rpm L/dB[m/(s^2)] Fig. 17: Synthesized torque mount vibration in x- direction 5 2

10 Synthesized Interior Noise, Torque Mount Share Measured Interior Noise of Original Engine Synthesized interior Interior noise Noise of New Engine n/rpm L/dB(A)[SPL] Fig. 18: Synthesized interior noise share (artificial head, left) of the torque mount vibration in x- direction Synthesized Interior Noise, Structure Bourne Share n/rpm L/dB(A)[SPL] Fig. 19: Synthesized interior noise share (artificial head, left) of the structure-borne transfer paths: left, right engine mounts and torque mount in x-, y- and z-direction. Summing the original airborne noise shares as well as the additional transfer paths described above, the resulting interior noise for the modified engine can be calculated. In the given example the modification leads to higher levels of the half engine orders and thus to higher engine roughness (Fig. 2) n/rpm L/dB(A)[SPL] n/rpm L/dB(A)[SPL] 6 7 Fig. 2: Spectrograms of the left artificial head channel of the measured interior noise and the synthesized interior noise using simulated excitation data VIBRATION DATA FROM ADDITIONAL MEASUREMENTS Besides combining simulated and measured input data, the order-synthesis methodology also allows for combining several measurements. To verify the methodology, the engine was modified by deactivating the injection of one cylinder. Both the original and the modified engine were measured, recording all relevant airborne and structure-borne input channels. Therefore it was possible to perform an interior noise synthesis using the BTPS model described in the first application example for both the original and the modified engine. Fig. 21 shows the measured acceleration at the torque mount in x-direction for the original engine while Fig. 22 illustrates the measured vibrations at the same position for the modified engine. Due to deactivating the injection of one cylinder the half engine orders, especially the orders.5, 1 and 1.5 show increased levels. In order to combine those two measurements the order level and phase data of the modified engine were extracted from the measurement using the orderanalysis technique described in the previous section: the first step toward an adequate order synthesis. Such data can then be used to synthesize new input time data by replacing the order level and phase data of the input measurement of the original engine. Fig. 23 illustrates the general procedure of combining several measurements using the order-synthesis methodology

11 Measured Torque Mount Vibration Measured Torque Mount Vibration n/rpm L/dB[m/(s^2)] Fig. 21: Spectrogram of the measured torque mount vibration in x-direction of the original engine n/rpm L/dB[m/(s^2)] Fig. 22: Spectrogram of the measured torque mount vibration in x-direction of the modified engine 2 Input measurement Original engine Order synthesis Input measurement Modified engine Order analysis Extracted order level and phase data e.g. order.5 to 8.5 Fig. 23: General procedure for calculating new input time data for BTPS using data from several measurements In this example the half orders from.5 to 8.5 engine order were extracted from the measured engine and torque mount vibrations, for three directions. The result of the order synthesis for the torque mount vibration in x-direction is shown in Fig. 24. The order level and phase data of the new input time signal show good agreement to the measured data of the modified engine, while the rest of the signal corresponds well with the measured vibrations of the original engine. Again the new time data are free of any audible artifacts and thus particularly suitable for auralization. th th Torque Mount Vibration, Order Synthesis The synthesized interior noise contribution of the structure-borne transfer paths (left and right engine mount and torque mount, x-, y- and z-direction) for the modified engine is shown using measured (Fig. 25, left) and simulated (Fig. 25, right) data. The comparison shows good agreement with respect to the order levels of the half orders from.5 th order to 8.5 th order. It is also recognizable that there are some variations at higher frequencies, since these signal contents are based on different measurements n/rpm L/dB[m/(s^2)] Fig. 24: Spectrogram of the synthesized torque mount vibration in x-direction 5 2

12 Interior Noise, SB 5 Interior Noise, SB 5 simulation software for a hybrid approach combining simulated excitation signals with measured transfer paths from the different source locations (for example at the powertrain) to the receiver positions (such as the driver s ears). 2 n/rpm L/dB(A)[SPL] n/rpm L/dB(A)[SPL] Fig. 25: Spectrograms of the structure-borne share of the synthesized interior noise of the modified engine using measured input data (left) and data from several measurements (order synthesis, right) for the artificial head (left ear). The comparison of the synthesized interior noise sum (sum of airborne and structure-borne noise shares) in Fig. 26 based on measured (left diagram) and synthesized (right diagram) input data also shows a good agreement, although possible changes of the airborne noise radiation due to the modification were not taken into account in this example. However, if the implemented modification has great impact also to the airborne noise radiation of the engine and the airborne noise share has a relevant effect to the interior noise, it may also be necessary to take these effects into account. Interior Noise 2 n/rpm L/dB(A)[SPL] Interior Noise 2 n/rpm L/ db(a)[spl] Fig. 26: Synthesized interior noise of the modified engine using measured input data (left) and data based on a combination of simulation and measurements (order synthesis, right) for the artificial head (left ear). CONCLUSION It has been a challenging task to provide high-quality audible results from the simulated engine order level and phase values without artifacts. The described method based on order analysis and synthesis delivers these extraordinarily results using a high rpm resolution and a non-linear resampling technique. It also allows for merging different measurements within the scope of the Binaural Transfer Path Analysis and Synthesis. Instead of measuring the modified vehicle once again, only the paths of interest need to be considered, reducing project duration and cost. Finally, a dataset for an acoustic driving simulator, taking into account both sound and vibration, is available at a very early stage of vehicle development even without having a prototype, helping to increase understanding of customer responses and perceptions of target sounds. It is up to the sound quality engineer to monitor the quality of the vehicle-powertrain project. The described tools, including vehicle sound simulation, support this activity. REFERENCES 1. Genuit, K. and Bray, W.: A Virtual Car: Prediction of Sound and Vibration in an Interactive Simulation Environment, 21 SAE Noise & Vibration Conferenc e Proceedings, Traverse City. 2. Sottek, R.; Sellerbeck, P. and Klemenz, M.: An Artificial Head which Speaks from its Ears: Investigations on Reciprocal Transfer Path Analysis in Vehicles, Using a Binaural Sound Source, 23 SAE Noise & Vibration Conference Proceedings, Traverse City. 3. Fahy, F. J.: The vibro-acoustic reciprocity principle and applications to noise control. Acustica 81 (1995), Sottek, R.; Riemann, D. and Sellerbeck, P.: Virtual Binaural Auralisation of Vehicle Interior Sounds, Proceedings of the Joint Congress CFA/DAGA 4, Strasbourg Genuit, K. and Bray, W. R.: Prediction of sound and vibration in a Virtual Automobile, Sound & Vibration, July Paul, S.; Schulte-Fortkamp, B. and Genuit, K.: Soundquality and Soundscape: A New Qualitative Approach to Evaluate Target Sounds, in Proceedings of Internoise 24, Prague. Using presently-available tools, it is not only possible to predict the perceived sound quality of interior noise from measurements, but also to use data from acoustic