I. Field of the Invention
The present invention relates to a method for analyzing sound transmissions in a system from a sound source to a receiver.
II. Description of Related Art
Transfer path analysis (TPA) is a technique for comparing the sound transmission paths in a system from the sound source and to a receiver. TPA is particularly useful in automotive applications to compare the sound transmission paths between the various sound sources, such as the engine, transmission, etc., and the receiver, typically a user's ears in the passenger compartment.
In the previously known applications of TPA for automotive systems, each sound path between the sound source and the receiver is first identified. The sound transmission paths, furthermore, include both structure-borne sound paths and air-borne sound paths. For example, structure-borne sound paths would include the sound path from a sound source, such as the engine, through the engine mountings and to the engine chassis. Conversely, air-borne sound paths would include the sound path between a side (surface) of a sound source, such as an area of the engine surface, and through the air (by the direct sound radiation from the source) to the receiver.
The total sound pressure produced at the receiver is decomposed into each path contribution, which is the summation of sound pressures through the all structure-borne paths plus the summation of sound pressures through the all air-borne paths. The sound pressure through a structure-borne path is estimated by the force at the path in the source operation multiplied by the (structural) transfer function from the path to the receiver. The sound pressure through an air-borne path is estimated by the volume acceleration from the surface of the source structure in the source operation at the path multiplied by the (acoustic) transfer function from the path to the receiver.
One disadvantage of the previously known transfer path analysis utilized in such systems, and particularly automotive systems, is that only the magnitude of the noise (sound) received at the receiver was used to analyze the sound transmission paths. Conversely, the direction of the sound at the receiver for each sound path can not be considered. This, in turn, resulted in inaccurate or at least incomplete analysis of the sound transmission paths in the system between the sound source(s) and the receiver(s).
The present invention provides a method for analyzing sound transmission paths in a sound system in the frequency domain which overcomes the disadvantages and shortcomings of the previously known methods.
In brief, the present invention provides an improved method for analyzing sound transmission paths between a sound source(s) and a receiver(s) in a system which comprises the three-dimensional analysis of the sound radiation to the receiver. These sound transmission paths include both structure-borne sound paths, e.g. a sound path from the engine through an engine mounting to a passenger' ear in an automotive vehicle, as well as air-borne or acoustical paths, e.g. the sound path from a surface (side area) of the engine that directly radiates sound to the air to the passenger's ear.
The transfer function in terms of particle velocity vector for each sound path is first measured for each sound path preferably by using a particle velocity sensor at the receiver. Unlike the previously known transfer path analysis, since the particle velocity vector, rather than only the magnitude, at the receiver is measured, both the magnitude and direction of the sound at the receiver are measured in terms of the particle velocity vector.
For the structure-borne sound path, the transfer function vector is the particle velocity vector at the receiver divided by the force at the structure-borne sound path. For the air-borne path, the transfer function vector is the particle velocity vector at the receiver divided by the volume acceleration at the air-borne sound path.
However, since force is difficult in many cases to measure, the force at the paths may be estimated instead. Traditionally, the estimation is exercised as follows. After removing the source structure (e.g. engine), the accelerances (transfer function between the excitation force to the acceleration response) among the all structure-borne paths are measured, which make an accelerance matrix. The forces at the paths are estimated by the accelerations at the paths pre-multiplied by the inverse of the accelerance matrix. In the case of over-determination, matrix pseudo-inverse is used instead of matrix inverse.
The particle velocity (vector) transfer function from a path to the receiver multiplied by its associated force or volume acceleration at the sound path indicates the contribution of the path to the particle velocity (vector) at the receiver, and thus it enables to identify the dominant sound path(s) and direction(s) of the received particle velocity (vector) at the receiver.
Alternatively, a microphone array at the receiver is utilized to measure the radiated sound distribution (of the sound radiation surface) to the receiver. Both the particle velocity vector and the radiated sound distribution specify where the major part of the sound comes from (to the receiver). In the microphone array case, the standard transfer path analysis of sound pressure is implemented for each microphone of the microphone array. Then, the sound radiation distribution (contribution) to the receiver is identified.
A better understanding of the present invention will be had upon reference to the following detailed description when read in conjunction with the accompanying drawing, wherein like reference characters refer to like parts throughout the several views, and in which:
With reference first to
The vehicle 20 includes a sound source 22, such as a vehicle engine, which generates noise (sound) during operation. That noise is received by a receiver 24, e.g. a vehicle passenger, in the passenger compartment 26. Ideally, minimization, or at least reduction, of the noise (sound) received by the receiver 24 is desired.
In the sound transmission analysis method of the present invention, each sound pathway from the sound source 22 and to the receiver 24 is first identified. The sound paths include both structure-borne sound paths and air-borne or acoustic sound paths.
For example, still referring to
The identification of the sound paths also includes acoustical sound paths from the sound source 22 to the receiver 24. These acoustical sound paths include, for example, sides (surface areas) 34 and 36 of the engine. The acoustical sound paths differ from the structure-borne sound paths in that the acoustical sound paths create sound through air compression caused by surface (the sides 34 and 36) vibration of the engine 22 whereas the structural sound paths through the engine mounts 30 and 32 create sound through mechanical vibration of the structural components of the chassis 28 and compartment 26.
A three-dimensional particle velocity vector is determined which represents each sound path contribution to the particle velocity vector at the receiver. In other words, the particle velocity vector at the receiver is decomposed into each path contribution at each frequency:
Here, the superscript Op explicitly denotes a value in an operational condition of the source. For example, a particle velocity vector 40 corresponds to the sound path of the surface 34 while three-dimensional particle velocity vectors 42, 44 and 46 respectively correspond to the sound paths of the surface 36, engine mount 30, and engine mount 32. Furthermore, each vector 40-46 includes not only the magnitude of the contributed sound (particle velocity) to the receiver 24, but also the direction of the contributed sound (particle velocity) to the receiver 24.
In order to obtain the three-dimensional sound vectors 40-46, a particle velocity sensor 50 is preferably positioned at the receiver 24. Such a particle velocity sensor measures not only the magnitude but also the direction of the received sound (particle velocity).
Still referring to
Here, {right arrow over (V)}Path nOp is obtained by two different measurements. ({right arrow over (V)}Reciever/FPath n) is the structure-borne transfer function in terms of particle velocity vector from the path n to the receiver which is measured without the source operation. The other measurement is to obtain the force FPath nOp (n=3 and 4) in the actual source operation. Force sensors 64 and 66 are preferably mechanically connected in series with the sound path between the sound source 22 and the structural support, such as the engine mounts 30 and 32. In many cases it is not practical to install a force sensor 64 or 66 in series with the structure-borne sound path (mount). In these cases, any conventional alternative method, such as the matrix inversion method or the stiffness method, is implemented by installing accelerometer in parallel with the structure-borne sound path. Force (at each path) is generally a three-dimensional vector, and furthermore not only three-dimensional force but also three-dimensional moment (torque) needs to be considered for each path to be theoretically exact. Although only one of the six dimensions is shown with respect to force at structure-borne path, any of other five dimensions are taken into account whenever it is not negligible.
Here, N is the moment (torque) at the structure-borne path.
The transfer function ({right arrow over (V)}Receiver/FPath n) for each structure-borne sound path is determined in any conventional fashion, such as by using an excitation shaker or an impact hammer to excite each sound path.
Here, FPath n is the excitation force by the shaker or hammer at the path n, and {right arrow over (V)}Receiver is the resulting particle velocity vector at the receiver.
In order to obtain the particle velocity contribution from the air-borne paths, volume acceleration from the sound radiation surface of the source is used.
Here, ({right arrow over (V)}Receiver/QPath n) is the air-borne transfer function in terms of particle velocity vector from the path n to the receiver which is measured in a separate process without the source operation. The volume acceleration QPath n (n=1 and 2) in the source operation is ideally measured by volume acceleration sensors installed near the source surface 34 and 36. However, because it is often difficult, acceleration sensors are instead installed on the surface 34 and 36 each of which well represents the acceleration of the surface. The volume acceleration is obtained as the measured acceleration multiplied by the area of the surface.
The transfer function vector ({right arrow over (V)}Receiver/QPath n) for each air-borne sound path is determined in any conventional fashion, such as by using a volume acceleration source (speaker) to excite each sound path.
Here, QPath n is the excitation volume acceleration by a speaker at the path n, and {right arrow over (V)}Receiver is the resulting particle velocity at the receiver.
With reference now to
At step 102, the transfer functions in terms of particle velocity vector which are ({right arrow over (V)}Receiver/FPath n) and ({right arrow over (V)}Receiver/QPath n) are measured without the source operation. If necessary, the transfer functions between the structure-borne paths (APath n′/FPath n) are also measured without the source operation in order to estimate the force FPath nOp (by the matrix inversion method). Step 102 then proceeds to step 104.
At step 104, the source is operated in the actual operating conditions, and the responses {right arrow over (V)}ReceiverOp, QPath nOp and FPath nOp (and/or APath n′Op for the path force estimation, if necessary) are measured. Step 104 then proceeds to step 106.
At step 106, each operational response at the path is multiplied by the corresponding transfer function vector to find the contribution of the path to the particle velocity vector at the receiver. The resulting these terms
are compared to rank order the paths at each frequency. The magnitude of the particle velocity vector |{right arrow over (V)}Path nOp|=√{square root over (VPath nOp|hd x2+VPath nOp|y2+VPath nOp|z2)}, (n =1,2,3,4) may be first compared to identify the dominant path.
The measured particle velocity at the receiver {right arrow over (V)}ReceiverOp should be compared with the summation of the all contributions of the paths in order to validate the transfer path analysis model and measurement accuracy.
VReceiverOp≈{right arrow over (V)}Path 1Op+{right arrow over (V)}Path 2Op+{right arrow over (V)}Path 3Op+{right arrow over (V)}Path 4Op
If the error (equality) in the above equation is not acceptable, the identified source-paths-receiver model (the step 100) and/or the experimental measurements (the steps 102 and 104) need improvement before making a conclusion of the sound transmission path comparison.
By thereafter comparing the sound attributable to the sound path, the dominant sound path(s) and direction(s) are identified. Following such identification, improvements in the mechanical design of the vehicle or other sound system may be implemented which addresses and reduces the sound transmission to the receiver.
With reference now to
With reference now to
At step 202, the transfer functions in terms of sound pressure measured by each microphone of the microphone array which are (PReceiver|m/FPath n) and (PReceiver|m/QPath n) are measured without the source operation. Here, the subscript m (=1, . . . , M) denotes the microphone number of the microphone array, and M is the total number of the microphones in the array. If necessary, the transfer functions between the structure-borne paths are also measured without the source operation in order to estimate the force FPath nOp (by the matrix inversion method). Step 202 then proceeds to step 204.
At step 204, the source is operated in the actual operating conditions, and the responses PReceiverOp|m , QPath nOp and FPath nOp (and/or APath n′Op for the path force estimation, if necessary) are measured. Step 204 then proceeds to step 206.
At step 206, each operational response at the path is multiplied by the corresponding transfer function to find the contribution of the path to the sound pressure at each microphone at the receiver. The resulting these terms
are compared to rank order the paths at each frequency. The average of the sound pressure measured by the M microphones may be first compared to identify the dominant path.
The measured sound pressures at the receiver PReceiverOp|m (m=1, . . . , M) should be compared with the summation of the all contributions of the paths in order to validate the transfer path analysis model and measurement accuracy.
P
Receiver
Op|m≈PPath 1Op|m+PPath 2Op|m+PPath 3Op|m+PPath 4Op|m (m=1, . . . , M)
If the error in the above equation is not acceptable, the identified source-paths-receiver model (the step 200) and/or the experimental measurements (the steps 202 and 204) need improvement. The major sound radiation areas, such as 54, 56, 58 and 60, are determined from the M sound pressure values.
f(P1, . . . , PM)
where f=sound pressure mapping function
P=sound pressure (at the receiver)
M=number of microphones.
From the foregoing, it can be seen that the present invention provides an improved transfer path analysis for sound transmission systems, such as automotive vehicles, which accounts not only for the magnitude, but also for the direction of the sound at the receiver. As such, not only the identification of the dominant sound paths but also the direction of the sound at the receiver is identified so that the appropriate action may be taken for reduction in the sound transmission to the receiver as required.
It will also be understood that the sound transmission along any particular sound path may include multiple frequencies. As such, frequency analysis using mathematical techniques such as fast Fourier transforms may be required for a complete analysis of the sound system.
Having described our invention, many modifications thereto will become apparent to those skilled in the art to which it pertains without deviation from the spirit of the invention as defined by the scope of the appended claims.