WO1996013702A9

WO1996013702A9 - System for real time measurement of impedance spectra or transfer functions of acoustic and mechanical systems

Info

Publication number: WO1996013702A9
Application number: PCT/AU1995/000729
Authority: WO
Filing date: 1995-11-01
Publication date: 1996-07-04

Abstract

A system (11) for real time measurement of a transfer function of an acoustic or mechanical system includes, a cylindrical tube (12), with a horn (13) concentrically located within the cylindrical tube (12) and having an axis of symmetry coincident with the axis of symmetry of the cylindrical tube (12). The end (17) of the horn (13) having the smallest cross-section is butted onto a cylindrical pipe (27), which is provided with a small orifice (14) at one end. The horn (13) expands exponentially to a larger orifice (15) at the other end. A transducer (18) drives into the large orifice and is driven by a signal synthesised by a synthesiser (90) via a power amplifier (21). Located near the small orifice (14) of the cylindrical pipe (27) is the second transducer (20), such that the sound level is emitted from the small orifice (14) and the response signals of the system, into which this sound level is injected, can be monitored. The second transducer (20) is connected to the response analyser (91) capable of receiving and analysing the signals from the second transducer (20).

Description

SYSTEM FOR REAL TIME MEASUREMENT OF IMPEDANCE SPECTRA OR TRANSFER FUNCTIONS OF ACOUSTIC AND MECHaANICAL SYSTEMS Introduction

The present invention relates generally to an apparatus and method of real time measurement of transfer functions and acoustic impedance spectra of acoustic and mechanical systems. Background of the Invention

A transfer function of a system is defined as the ratio of the output (of the system) to the input (of the system). The input and output may be represented as spectra, and hence in this case the transfer function is a spectrum. The transfer function may be represented as a complex spectrum, or as the magnitude and phase spectra.

Acoustic impedance is a particular case of a transfer function where the "output" is indicative of acoustic pressure and the "input" is indicative of acoustic volume flow or volume velocity.

Impedance spectra contain many of the most important data about the acoustic performance of passive systems. In musical wind instruments, for instance, the peaks in the impedance spectrum indicate the frequencies of resonant modes which may cooperate in the coupled oscillation of the reed- bore system. In the case of the vocal tract, the rather broader and weaker resonances or formants can be identified from features in the impedance spectrum. In compound systems, the impedance spectra of two adjacent stages determine the transmission spectrum. All of the examples that we have given are air-filled cavities. However, the entire class of systems of musical instrument and mechanical vibrating systems should not be excluded from the examples.

The impedance spectrum and transfer functions are potentially useful to instrument manufacturers because they are one of the most important factors in an instrument's performance and because they can be objectively measured. The frequencies of the instrument's resonances, and the heights and widths of the impedance peaks corresponding to them can be correlated with the intonation of the instrument, with the ease of production of different notes and the quality of those notes (e.g. whether they are "bright" or "dark"; whether they have or lack "centre"). The impedance spectrum may be used as a quality control measurement for manufacture.

The impedance spectrum is quantitative and precise. In this sense it has an advantage over the assessment of the quality of an instrument by a good player. Except in extreme cases, it is not simple to determine from an impedance spectrum or transfer function whether an instrument is good or bad. This is best judged by an expert player (or a panel of such players). On the other hand, the impedance spectrum or transfer functions can determine whether two instruments are acoustically similar or not. One can therefore tell whether a given instrument is similar to or different from a reference instrument which was previously measured (and which may have been judged to be good or poor by a panel of players). This could be a useful procedure in instrument manufacture. For mass production, such a test could be used to determine whether each instrument fell within acceptable ranges of such variables as the frequency and height of impedance peaks and other features. In hand-made instruments, such a test could become part of a feedback loop: the maker could make gradual changes and test to see how the important parameters changed as a result.

There are also a number of instances where rapid measurements of the acoustic impedance of the vocal tract may be useful. The most obvious is in research into acoustic phonetics, but there are other potential applications as well.

The frequencies and relative heights of the resonances or formants of the vocal tract can be simply determined from the impedance spectrum. These data determine, to a large extent, the sound of spoken vowels. There is a reasonably good mapping between vowel sounds and the frequencies (fi^) of the two formants with lowest frequencies. Further.

SUBSTITUTE SHEET (RULE 26, there is a relatively simple correlation between (f 2) ^ant^ ^me degree of opening of the mouth and the tongue position, so it is relatively easy for someone to adjust (fi^) when the feedback is provided.

In most cases, people with normal hearing can determine the vowel sounds of their own languages by listening.

However, the profoundly deaf have considerable difficulty in learning to speak normally because they cannot use their ears in a feedback loop to match the sound that they produce with a target sound. While they can see the position of the lips of a speaker, this is inadequate feedback for vowel production because the internal shape of the mouth (position of tongue and palette) is also involved. The data (fi ^?) could close the feedback loop for such people.

The real time comparison of a transfer function of an acoustic or mechanical system with the transfer function of a reference system has many commercial applications including: i) in the production of musical instruments, to provide a feedback in manufacture for handmade instruments, quality control for mass production of instruments, and as a research tool in the design of better instalments: ii) in performance spaces, unwanted resonances may be predicted using architectural models for all geometric combinations of performers and listeners and to study the acoustic pathology of existing spaces: iii) in mechanical structures, unwanted resonances may be determined for all geometric combinations of source and site: iv) in speech pathology as hereinbefore described and language teaching, a visual display may be used to close the feedback loop to correct speech.

In the prior art. the impedance spectrum is obtained by scanning over the spectrum of frequencies one frequency at any one time. A source with an output impedance which is high compared to that of the system measured provides a volume velocity source at a given single frequency. A microphone measures the acoustic pressure in the system measured and the measurements are made to acquire the impedance spectrum. Prior art methods, while capable of high precision, are slow. Although slowness is not a major problem for many aspects of acoustical research, it does make the methods unattractive for some applications, and impossible to use in others.

For small systems such as the vocal tract and musical instruments, it is difficult to inject substantial acoustic power at wavelength greater than a few centimetres. Small high power transducers rarely have a flat frequency response and are usually frequency dependent and non-linear.

These frequency dependent responses are also compounded by the acoustic characteristics of the space in which the measurement is performed. In an effort to ameliorate the disadvantages of the prior art or at least to provide a commercially viable alternative to the prior art, it is proposed to provide an apparatus and method which enables the measurement of the entire impedance spectrum or transfer function of an acoustic or a mechanical system in real time as well as a method for calibrating such a system. Further, embodiments of the invention enable all measurements to be related to a reference system, and so are well suited to making comparisons and for eliminating effects due to the reverberant field. Summary of the Invention

According to a first aspect the invention provides a method of characterising in real time an acoustic, mechanical, electro-acoustic or electro-mechanical system under test comprising the steps of: generating an input signal having a plurality of selected discrete frequency components, injecting the input signal into the system under measurement, and measuring a response signal of the acoustic, mechanical, electro- acoustic or electro-mechanical system to the injected signal. In one application of the invention, the method further includes applying said method in order to compare a test system with a reference system. Preferably said method further includes the steps of: comparing the response spectrum of the reference system with a predetermined response spectrum. The method may further include adjusting an amplitude, a phase or frequency of each frequency component of the generated input signal and repeating the steps of injecting the signal and measuring the response until the response matches the desired response; so as to achieve a calibrated input signal which causes the reference system to produce the predetermined response spectrum: and subsequently injecting the calibrated input signal into the test system. measuring a response signal of the test system to the injected input signal, and determining a ratio of test system response to reference system response in order to analyse characteristics of the test system.

Typically, a test system may be any one of acoustic, mechanical, electro-acoustic or electro-mechanical systems and the reference system will be a similar system. The test system generally comprises the reference system with some pertubation applied. Examples of such systems are: 35a room as the reference system and a musical instrument within the room and the room itself as the test system; a model of an architectural structure as the reference system and a modified version of the same model of the architectural structure as the test system: however, the test system is not restricted to a pertubation of the reference system and the reference system need not be linked to the test system. Examples of this are: a first violin as the reference system and a second violin as the test system: a first amplifier and transducer as the reference system and a second amplifier and transducer as the test system.

In one form of the invention the injected input signal is selected such that the frequency components of the response signal have frequencies and amplitudes and/or phases directly correlating to the impedance spectrum of the system being measured.

According to a second aspect, the present invention provides an apparatus arranged to measure an acoustic impedance spectrum or a transfer function, comprising: signal generating means arranged to generate a signal having a plurality of selected, discrete frequency components, control means arranged to tailor the relative amplitudes and phase of the frequency components to control a frequency spectrum of the signal, generating means being arranged to transmit the acoustic or vibrating signal into a system under measurement with high power through a small aperture, and measurement means arranged to measure a frequency response of the system.

Preferably, the control means is also arranged to apply an adjustment to said acoustic signal the adjustment being related to the reciprocal of the amplitude of the Fourier components of the measured frequency response such that the frequency response to the adjusted acotistic signal has a predetermined frequency spectrum.

The preferred embodiment will also include display means arranged to display the measured frequency response and preferably the control means is a compviter also having a data storage means to store the data relating to the measured frequency response for the purpose of vising the stored data at a later date.

In one preferred form of the invention the measured frequency response of a reference system is stored in the data storage means and compared with the measured frequency response of a system under test.

Preferably the comparison is made by generating the spectrum representing the ratio of the response of the reference and test systems. In the preferred embodiment the comparison is performed by calibrating the input signal to provide a flat output spectrum having a unity amplitude for the reference system and then measuring the response of the test system to that same input.

Preferably, the signal generating means comprises transducer means having a high impedance acoustic outpvit driven by an electrical signal generator. In one form of the invention the transducer means comprises a speaker driving into an inverted horn to produce maximum energy transfer with high acoustic impedance which gives the transducer means an output impedance matched to the test system with maximal power output. In another form an impedance device such as a narrow tube filled with glass fibre is placed at the outpvit of the horn to increase the output impedance of the transducer system to further isolate the system under measurement from the signal generating means.

In another form of the invention the transducer means comprises a small magnet located in proximity to an electromagnetic coil and driven by an electrical signal to cause the magnet to vibrate, the vibrations being transmitted to a mechanical system to which the magnet is connected.

Preferably, the variovis embodiments of the invention use a synthesiser comprising a digital to analogue converter and a compviter to generate the electrical signal for driving the transducer means. Preferably the synthesiser will synthesise the signal by performing an inverse Fourier transform on a frequency spectrum selected for the input signal.

In one embodiment of the invention the measurement means comprises second transducer means arranged to detect an acoustic response signal and to produce an electrical analogue. The electrical analogue of the response signal is then preferably transformed into the frequency domain by Fourier transform, this step being performed preferably after the electrical signal has been converted to a digital signal vising an analogue to digital converter.

In another form of the present invention, the transfer function of the system under measurement is indicative of the acoustic gain of the system. Preferably the present invention includes a method of measurement of a transfer function of a mechanical or electro-mechanical system to which transducer means are linked for transmitting signals to and receiving signals from the mechanical or electro-mechanical systems, comprising the steps of: generating a predetermined reference signal having a plurality of selected discrete frequency components; driving the transducer means with said predetermined reference signal to thereby excite the mechanical system: receiving a response signal of the mechanical system via the transducer means; and analysing the response signals received by the transducer means.

In a preferred form of the present invention the transducer means comprises at least one or more transducers which include any of the following; a displacement transducer, a velocity transducer, a force transducer, a microphone, a speaker or the like.

In a preferred form of the invention, the analysis of the response signal involves first performing a Fovirier transform to obtain a frequency spectrum.

According to a third aspect the present invention provides a method of calibrating an apparatus for measuring an acoustic impedance spectrum or transfer function of a system comprising the steps of: a) generating an initial input signal having a plurality of discrete frequency components of selected frequency and amplitude and/or phase and injecting the initial input signal into the system: b) measuring an initial response signal of the system to the initial input signal; c) generating a correction signal by determining a ratio of a desired response signal to the initial response signal; and d) generating a calibrated input signal as the product of the initial input signal and the correction signal. In non-linear systems, or when rescaling is required after changing the input signal characteristic, the method above may not produce an input which results in an output matching the desired output to the degree required. In that case, the first stage calibrated input is injected into the system and the ratio of step c is again obtained. The product of this ratio and the first stage calibrated input is then prodviced to provide a second stage calibrated input. This process can be repeated as many times as is necessary to produce the required response.

This calibrated input signal can then be used to measure the reference system with a pertvibation applied or to measure a different system which is to be compared with the reference system.

Preferably, the selected discrete frequency components define a predetermined frequency spectrum and as are selected in accordance with the freqviencies of interest to be measured in the system. Brief Description of the Drawings

Embodiments of the present invention will now be described by way of example, with reference to the accompanying drawings, in which:-

Figure 1 is a schematic diagram of a method of calibration of one embodiment of the present invention. Figure 2 is a schematic diagram of the apparatus of a first embodiment of the present invention.

Figure 3 is a schematic diagram of a variation on the apparatus of Figure 2.

Figure 4 is a schematic diagram of the method of calibration of the embodiment of Figure 2.

Figure 5 is the embodiment of Figvire 2 vised in the measurement of a test system.

Figure 6 is a schematic diagram of the second embodiment of the present invention. Figure 7 is a schematic diagram of a third embodiment of the present invention.

Figure 8 is a schematic diagram of a fourth embodiment of the present invention. Figure 9 is a schematic diagram of the method of the embodiment of

Figure 7.

Figure 10 is a schematic diagram of a fifth embodiment of the present invention.

Figure 11 is a block diagram of a method of the preferred embodiment of the present invention.

Detailed Description of the drawings

A first embodiment of the present invention is described with reference to Figures 1-4.

Illustrated in Figvire 1 is a schematic block diagram of the procedvire of calibrating the apparatus of this embodiment. Figure 1 shows a synthesiser 90 comprising a compviter and digital-analogue converter (DAC) arranged to synthesise a predetermined frequency spectrum held in storage 95 into an electrical input signal. Preferably this frequency spectrum will be equal to a desired response spectrum of the calibrated system. The electrical input signal is fed through an amplifier 21 which drives a first transducer 18 to inject a reference signal in a reference system 40. A response signal of the reference system 40 is monitored by second transducer 20 which is connected to a response analyser 91 arranged to analyse the response signal into a corresponding response frequency spectrum. The response analyser 91 preferably comprises a computer, having storage means 92, and an analogue-digital converter (ADC), such that the response analyser 91 transforms the response signal into the corresponding response frequency spectrum and stores this information in the storage means 92. However, the response analyser 91 can also be a spectrum analyser with a storage means 92 to accomplish the same task. A signal processor 93 then calculates a compensating frequency spectrum as the prodvict of two sets of terms, the first of which is directly proportional to the amplitude of each frequency component of a desired frequency spectrum and the second of which is inversely proportional to the amplitvide of each freqviency component of the corresponding response frequency spectrum and directly proportional to the predetermined frequency spectrum. The compensating frequency spectrum is then held in a storage device 94 and used by the synthesiser to synthesise a compensated input signal. It should be noted that in the event that the desired response is flat each freqviency component of the compensating freqviency spectrvim is simply proportional to the reciprocal of the respective freqviency component of the corresponding response spectrum. In some cases, the desired frequency spectrvim may be the predetermined frequency spectrvim. Subsequently the compensating freqviency spectrum is synthesised, (Figure 1) by the synthesiser 90, into a compensation signal which is injected, via the amplifier 21. and the first transducer 18 into the reference system 40. The response signal, from the reference system 40 excited by the compensation signal through the first transducer 18. is detected by the second transducer 20 and analysed by the response analyser 91 into a recovered frequency spectrum. The ratio of the amplitude of each freqviency component of the recovered freqviency spectrum and the amplitude of each corresponding freqviency component of the predetermined freqviency spectrum is tested by a comparator 96 and if it is within a predetermined error margin of a selected value, then the apparatus is calibrated to the reference system 40. If the ratio is not within the predetermined error margin of the selected value, then we choose the calculated compensating frequency spectrvim as the predetermined freqviency spectrum, choose the recovered frequency spectrum as the corresponding frequency spectrvim and repeat the process of calculating corresponding frequency spectrum until the apparatus is calibrated to the reference system 40.

The preferred apparatus for the implementation of the method described above will now be described with reference to Figure 2. The apparatvis 11 comprises a cylindrical tvibe 12. with a horn 13 concentrically located within the cylindrical tube 12 and having an axis of symmetry coincident with the axis of symmetry of the cylindrical tube 12. The volume between the horn 13 and the cylindrical tube 12 is filled with foam 19 to keep the horn 13 rigid. The end 17 of the horn 13 having the smallest cross- section is butted onto a cylindrical pipe 27, which is filled with acoustic fibre padding 28 to increase the output impedance, and is provided with a small orifice 14 at one end. The horn 13 expands exponentially to a larger orifice 15 at the other end. The cylindrical tube 12 is sealed around the horn 13 at one end 17, such that only a small section of the horn 13 at the smallest cross-section protrudes from the sealed end 17 of the cylindrical tube 12. Located near the small orifice 14 of the cylindrical pipe 27 is the second transducer 20. such that the sound level emitted from the small orifice 14 and the response signals of the system, into which this sound level is injected, can be monitored. The second transducer 20 is connected to the response analyser 91 capable of receiving and analysing the signals from the second transdvicer 20. The computer 23 is connected to a digital-analogue converter 24 and is capable of generating a reference signal corresponding to a predetermined frequency spectrum, evaluated by or. inpvit into the computer 23. The computer 23 has a display 30 to display the spectra. An alternative arrangement is shown in Figure 3 in which the microcomputer 23 and A to D converter 24 are replaced by a spectrum analyser 25.

The digital-analogue converter 24 feeds the reference signal to an amplifier 21 connected to the digital-analogue converter 24 and. in turn the amplifier 21 drives the first transducer 18 with the reference signal. The first transducer 18. is enclosed in a chamber 16 which is filled with acovistic insulating (or damping) material 26 and is coupled to the horn 13 at the large orifice 15 such that the sound generated by the first transducer 18. when the system is driven by the reference signal, travels the length of the horn 13 and exits at the small orifice 14 end of the cylindrical pipe 27 adjacent to which the second transducer 20 is located. A high signal to noise ratio is achieved at the small orifice end 14 by driving the first transducer 18 at sufficient power.

In any measurement system, the available information is always finite. therefore one may only obtain a finite amount of information about a finite number of frequencies. (This is true not only of digital systems but also of systems in which the freqviency and amplitude appear to be continuous variables because of thermal and other noise and finite resolution.) Once one accepts that a measurement has only a finite number of frequencies, one can make large improvements in signal to noise ratio by selecting a set of frequencies, producing a signal all of whose energy is concentrated in those frequencies only, and by measuring those freqviency components only. As such the selection of a suitable predetermined freqviency spectrum depends on the system upon which measurements will be made. In this embodiment of the present invention the preferred predetermined freqviency spectrvim is selected by choosing a set of discrete equally spaced freqviency components, between an upper and lower frequency value, all with a same amplitude. The frequency difference between any two adjacent frequency components of the discrete equally spaced freqviency components is small and the freqviency difference is typically chosen in the infrasonic range of freqviencies. while the upper frequency value is defined by the range of frequency one choose to measure. Where ambient noise is predominant in certain freqviencies. there is no reason why the amplitvide of the respective freqviency components of the predetermined spectrvim cannot be boosted. For instance, a factory environment often has substantial low frequency noise and as svich it may be advantageous to use a predetermined frequency spectrum which has larger amplitudes at low freqviencies.

In some cases when noise at a particular frequency, or other signals are present it may be useful to remove them from the predetermined freqviency spectrvim. For instance, when the vocal tract is being studied while in vise, there will be substantial power in very narrow bands around frequencies that are multiples of a pitch freqviency. The pitch frequency is easily identified because it is not in the predetermined frequency spectrum and because it is usually the largest component in the spectrum, and all the frequencies lying within a small range of the multiples of the pitch freqviencies can be eliminated from the predetermined frequency spectrvim. The resonances of the vocal tract are many times broader than the widths of the spectral peaks of frequencies at multiples of the pitch frequency so very little information is lost. It is also possible to adjust the predetermined frequency spectrum so that none or few of its harmonics coincide with those of the voice pitch frequency.

Many other examples of predetermined freqviency spectra may be conceived and in general the predetermined freqviency spectrum may be defined with freqviency components, amplitude and phase of interest for a particular test system. Another example of a predetermined freqviency spectrum can be one in which the freqviency of the components of this spectrum are related by a logarithmic scale, and would be of use in cases where the frequency ranges of interest are large. The method of measurement is described by the following steps with reference to Figure 4 as follows: i) a predetermined frequency spectrvim 31 is synthesised to provide the reference signal by performing an inverse Fourier transform. The reference signal is then used to drive the first transducer 18 which is coupled to the large orifice end 15 of the horn 13 such that the sound generated propagates mainly throvigh the horn 13 and exits at the opposite end. into a reference system 40. The horn 13 functions as an impedance matching device feeding through an acoustic impedance to prodvice an acoustic sovirce of high impedance at the small orifice 14. The reference system 40 is preferably any system to which another system is to be compared including cavities, mechanical, electro-acoustic and electro-mechanical devices, ii) The second transducer 20 located near the output of the small orifice 14, monitors the response signals of the reference system 40. The response signals are analysed by Fourier transform into the corresponding response freqviency spectrvim 32. From the response freqviency spectrum 32, a compensating freqviency spectrum is calculated as the product of two sets of terms, the first of which is directly proportional to the amplitude of each freqviency component of a desired freqviency spectrum and the second of which is inversely proportional to the amplitude of each frequency component of the corresponding response frequency spectrum and directly proportional to the predetermined frequency spectrum 33. iii) The compensation freqviency spectrum 33 is then synthesised using the inverse Fovirier transform technique to produce a compensation signal which is then used to drive the first transducer 18 to produce the corresponding acoustic compensation signal which is transmitted through the small orifice 14 into the reference system 40. The second transducer 20 monitors the response signal of the reference system 40 to the corresponding acoustic compensation signal. This response signal, of the reference system 40. is analysed by a Fourier transform technique to produce a recovered frequency spectrum 34 of this response signal of the reference system 40; iv) If the ratio of recovered frequency spectrvim 34 to the predetermined freqviency spectrum 31 is constant, to a desired degree, then the apparatus hereinbefore described is calibrated relative to the reference system 40. It should be noted that, unless stated otherwise, a ratio of two spectra means the magnitude spectrum comprising the ratio of the amplitudes of respective frequency components and that a phase spectrvim comprises the difference between phases of the respective frequency components of two spectra.

If the ratio of recovered freqviency spectrvim 34 to the predetermined freqviency spectrum 31 is not constant, to the desired degree, then the compensation freqviency spectrum 33 is modified. The procedure outlined in the steps iii) and iv) is repeated until the recovered freqviency spectrvim 34 and the predetermined frequency spectrum 31 has a ratio which is constant to a desired degree. The compensation freqviency spectrum 33 is modified during each cycle of the repeated steps iii) and iv) by changing the amplitude of each frequency component of the compensation frequency spectrum 33 so as to result in a ratio between the amplitude of each component freqviency in the recovered spectrum 34 and the amplitude of each corresponding component frequency in the predetermined spectrvim 31 to be constant and equal to a predetermined value within a desired degree. The apparatus is now calibrated to the reference system 40.

The compensation frequency spectrvim 33, obtained when the apparatvis is calibrated, can be vised to determine the transfer function or the impedance spectrum of a test system 45 (shown in Figure 4 and 5). The compensation freqviency spectrvim 33 is synthesised. by the synthesiser 90. vising an inverse Fourier transform technique to produce a compensation signal, the compensation signal in turn drives the first transducer 18 to produce a corresponding acovistic compensation signal which is transmitted through the small orifice 14 into the test system 45. The second transducer 20 monitors the response signal of the test system to the corresponding acovistic compensation signal. The response signal of the test system 45 is analysed vising the Fourier transform technique into a frequency spectrum of the test system 45 vising a response analyser 91 which is similar to that of Figvire 2 but could equally take the forms shown in Figvire 3. The ratio between the transfer function of the test system 45 and the transfer function of the reference system 40. in the calibrated apparatvis. is the frequency spectrum of the test system 45. In particular if the transfer function of the reference system has substantially uniform amplitude, of each component freqviency, then the transfer fvinction of the test system is proportional to the response frequency spectrum of the test system 45. In the case where both the impedance of the test system 45 and the impedance of the reference system 40 are low compared to the outpvit impedance of the source, at the small orifice 14, then the transfer fvinction of the test system 45 is the acoustic impedance.

In a variation of the first embodiment of the present invention the second transducer 20 is not located near the small orifice 14, in which case the embodiment is particularly useful in measuring the transfer fvinction of large architectural acoustic structures.

A second embodiment of the present invention is illustrated in Figure 6 in which an electro-mechanical transdvicer is vised, comprising a small permanent magnet 50 fixed to a violin bridge 51 of a violin 52. Adjacent to the permanent magnet 50 is an electromagnetic coil 53 mounted independently of the violin bridge 51 and magnet 50. A receiving transducer is used to receive the response signals of the violin induced by the signals generated at the electro-mechanical transdvicer. The receiving transducer is an acovistic transducer 55 placed near the violin, bvit a force, velocity or displacement transdvicer 54 attached to the violin can perform an equal task of receiving the response signals.

In this embodiment the electro-mechanical transdvicer is connected to the synthesiser in substantially the same configuration as described in the first embodiment of the present invention.

A response analyser 91 (Refer Figures 2 & 3) is connected to the receiving transducer 20 to analyse the received response signal.

A predetermined input freqviency spectrvim is synthesised, vising an inverse Fourier technique, into an input reference signal which is used to drive the electromagnetic coil 53. The generated electromagnetic field signal drives the permanent magnet 50 setting up vibrations in the violin bridge 51 and hence in the violin 52 respectively. The vibrations set up in the violin bridge 51 and violin 52 continue to produce a response signal which is measured by the receiving transducer and represents the resonances of the violin 52. The response signal is then Fourier analysed into a response freqviency spectrum. The ratio of response frequency spectrum and the input freqviency spectrvim is the product of the transfer functions of the electro-mechanical transdvicer and of the violin 52.

The product of the transfer functions of this electro-mechanical transdvicer and violin can be compared to a similar product of the same electro mechanical transdvicer and a different violin. The comparison of the product of the transfer functions in this manner provides an objective and sensitive means for comparing the acoustic performance of any two or more violins. A third embodiment of the present invention is shown in Figure 7 in which an electro-mechanical transducer 60 comprising a permanent magnet 61. a cvip-shape cylindrical tube 63 attached to a shaft 64 and an electromagnetic coil 65 fixed to the outer surface of the cylinder of the cup- shape cylindrical tvibe 63. all of which are enclosed in a rigid container 62 so that only the shaft 64 protrvides through a small hole 66 in the rigid container 62. The permanent magnet 61 is fixed to the rigid container 62 and the cup-shape cylindrical tube 63 with the electromagnetic coil 65 fitting around a portion of the permanent magnet 61. svich that a current passing through the electromagnetic coil 65 would cause the coil 65 and cup-shape cylindrical tvibe 63 to move relative to the magnet 61 and rigid container 62. The motion of cup-shape cylindrical tube 63 moves the attached shaft 64. and therefore a changing current through the coil 65 would set the shaft 64 vibrating. The shaft 64 is attached to a mechanical system 67. A velocity or displacement transdvicer 68 is also attached to the mechanical system 67 to measvire a response of the mechanical system 67 to a stimulus provided by the electro-mechanical transducer 60. An acoustic transducer 69 is also placed in the vicinity of the mechanical system 67 to receive the acoustic response signals resulting from the stimulus provided by the electro-mechanical transducer 60. In this form of the present invention the electro-mechanical transducer is connected to the amplifier 21 and synthesiser 90 in substantially the same configuration as herein described in the first embodiment of the present invention. The response analyser 91 is connected to the velocity or displacement transducer 68.

The acoustic transdvicer 69 is also connected to the response analyser, so that the response signals can be measured independently from the velocity or displacement transdvicer 68.

A predetermined input frequency spectrum is synthesised. to produce an input reference signal which drives the electromagnetic coil 65 to vibrate relative to the permanent magnet 61. This vibration of the electromagnetic coil 65 translates to a vibration of the cup-shape cylindrical tube 63 and attached shaft 64 since the coil 65 is fixed to the cup-shape tube 63. The shaft 64 is coupled to the mechanical system 67 so that the vibrations of the shaft 64 are transmitted to the mechanical system 67. Accordingly, the mechanical system responds to the vibration producing a response signal of the mechanical system 67. This response signal of the mechanical system 67 is measured by the displacement transdvicer 68 and acovistic transdvicer 69 independently. The response signal received by the velocity, or displacement, transducer 68 and the acoustic transdvicer 69 is analysed into a response frequency spectrvim for each of the transducers 68. 69 independently. The ratio of the response frequency spectrum and the input freqviency spectrum is the prodvict of the transfer functions of the electro¬ mechanical transdvicer 60 and of the mechanical system 67.

The product of the transfer function of the electro-mechanical transdvicer 60 and the transfer fvinction of the mechanical system 67 can then be compared to a similar product of the same electro-mechanical transducer 60 and a different mechanical system. In this manner the transfer function of one mechanical system can be obtained relative to the transfer fvinction of another mechanical system. For example, the transfer fvinction of a violin belly (test system) in the process of construction could be compared with that of the belly of another (reference system) violin (judged to be good) and the maker could adjust the properties of the test system to match those of the reference system.

A fovirth embodiment of the present invention is described with reference to Figure 8. this embodiment being similar to the third embodiment of the present invention. However, in this embodiment the shaft 64 is attached to a clamp 74 which clamps the mechanical system 67 between a pair of piezoelectric crystals 72.73. The vibrations of the shaft 64 are transmitted to the mechanical system 67 via the clamp 74 and the piezoelectric crystals 72,73. The pair of piezoelectric crystals 72,73, which play the role of differential force transducers, are connected to the response analyser so that an independent measurement may be obtained of the force response signal resulting from stimulation by the vibrating shaft. The force response signal measured by the pair of piezoelectric crystals 72.73, is analysed by the response analyser into a force response frequency spectrum. For the situation where the velocity transdvicer 68 is used in conjvmction with the pair of piezoelectric crystals 72.73. the response signal received at the velocity transdvicer 68 is analysed to produce a velocity response frequency spectrum. The ratio of the amplitude of the force response freqviency spectrum to the corresponding freqviency component amplitudes of the velocity response spectrvim is a mechanical transfer function of the mechanical system 67. In the case where the velocity transducer 68 is attached to the mechanical system 67 at the point where the vibrations of the shaft 64 are transmitted to the mechanical system 67, via the clamp 74 and piezoelectric crystals 72. 73. the mechanical transfer fvinction is the mechanical impedance spectrum. An acoustic response signal produced by the mechanical system 67, when stimulated by the vibrating shaft 64 as described above, is received at acoustic transducer 69 located in the vicinity of the mechanical system 67. The acoustic transducer 69 is connected to the response analyser so that the received acoustic response signal may be analysed into a frequency spectrum of the acovistic response signal. A ratio of the amplitvide of each component frequency of the frequency spectrum of the acoustic response signal to the amplitude of the corresponding freqviency component of force response frequency spectrvim is a measure of the acovistic gain of the mechanical system 67.

Figvire 9 is a schematic of the method of measurement of the fovirth embodiment of the present invention. A synthesiser 90 produces a reference signal from a predetermined frequency spectrvim to drive, via the amplifier 21, the electro-mechanical transdvicer 60 to transmit the reference signal to the mechanical system 67. The force transducers (the piezoelectric crystals 72,73) monitor the mechanical system 67 and receive the force response signal from the response of the mechanical system 67. Simultaneously the velocity transducer 68, also monitoring the mechanical system, measures a velocity response signal. The force response signal and the velocity response signal are analysed by the response analyser 91 to produce the force response frequency spectrvim and the velocity response spectrum respectively. The ratio of the force response frequency spectrum and the velocity response frequency spectrvim is the mechanical transfer function of the mechanical system 67. The mechanical transfer fvinction data of a reference system is stored in the storage means for later comparison with a mechanical transfer function of a measured test system. When the force response signal is measured at the same point, on the mechanical system 67, as the velocity response signal, then the resulting mechanical transfer fvinction is the mechanical impedance of the mechanical system 67. A fifth embodiment of the present invention is described with reference to Figure 10. A reference transdvicer 80 is connected to a reference amplifier 81 (this combination is hereinafter referred to as the reference device) which preferably has a known characteristic response to electrical signals driving the reference amplifier 81 and reference transducer 80. The reference amplifier 81 is electrically coupled to a waveform synthesiser 82 which is arranged to generate a desired electrical signal to drive the reference device and to produce a corresponding acovistic signal. An acovistic transdvicer 85 is placed at a predetermined distance from the reference transducer 80, of the reference device, to monitor the corresponding acovistic signal. The acovistic transducer 85 is connected to a Fourier analyser 84, and a compviter 83, and the corresponding acovistic signal is analysed to find its freqviency spectrum.

A predetermined freqviency spectrum is synthesised by the waveform synthesiser 82 to drive the reference device to produce a corresponding acoustic response signal. This response signal propagates to the acovistic transdvicer 85. positioned a predetermined distance away from the reference transducer 80 of the reference device, and the response signal received at the acovistic transdvicer 85 is fed to the Fovirier analyser 84. The response signal is analysed, by the Fourier analyser 84. into a frequency spectrum of the response signal and the data of this freqviency spectrum is stored in a storage means which is preferably associated with the computer 83.

A compensating freqviency spectrum is then calculated in substantially the same way as that described in the first embodiment of the present invention and is subsequently synthesised by the waveform synthesiser 82 to drive the reference device to prod vice a compensated output signal. The compensated output signal is then detected by the acoustic transducer 85 and analysed, by the Fourier analyser 84, into a frequency spectrum of the compensated output signal. If a ratio of the amplitude of each frequency component of the frequency spectrum of the compensated output signal and the amplitude of the corresponding frequency component of the predetermined frequency spectrvim is within a predetermined error margin of a selected value then the compensating frequency spectrum is calibrated, to the reference amplifier 81 and reference transducer 80, and shall be described hereinafter as the calibrated input freqviency spectrum. If the ratio of the amplitvide of each frequency component of the frequency spectrum of the compensated output signal to the amplitude of the corresponding freqviency component of the predetermined freqviency spectrum is not within the predetermined error margin of the selected value, then the steps of calculating the compensation frequency spectrvim are repeated vmtil one is satisfied that the compensation freqviency spectrum is the calibrated input freqviency spectrum.

A test amplifier 100 and a test transducer 101 may then be compared with the reference device by connecting them in place of the reference amplifier 81 and the reference transdvicer 82 in the apparatus described in this embodiment. The calibrated compensation frequency spectrvim is synthesised by the waveform synthesiser 82 to drive the test amplifier 100 and test transdvicer 101 to prodvice an acovistic signal. The acoustic signal produced by the test transdvicer 101 propagates across the predetermined distance and is received at the acoustic transducer 85 to be analysed by the Fovirier analyser 84 into a frequency spectrum of the acoustic signal. A ratio of the amplitvide of the frequency spectrum of the acoustic signal and the amplitude of the corresponding freqviency of the calibrated input freqviency spectrvim is an indication of the deviation of the test system (test amplifier 100 and test transducer 101) relative to the reference device.

Figure 11 is a block diagram of a method according to a preferred form of the present invention. An input signal 110 having a predetermined frequency spectrvim which is selected to have a set of discrete freqviency components with selected amplitude, phase and freqviencies. This input signal 110 is fed. via a power matching means 111 (as an example the horn of the first embodiment of the present invention), into a reference system 112.

A reference output signal as measured is a response of the reference system 112 to the input signal 110. The input signal 110 is also fed. via the power matching means 111. into a test system 113. A test output signal is measured as a response of the test system 113 to the inpvit signal 110. The reference output signal and the test output signal is put through a comparator 114 which compares the test ovitput signal with the reference output signal and hence characterises the test system 113 relative to the reference system 112. Typically, the selected discrete frequency components are chosen, svich that the frequency difference falls within the infrasonic range (0-15Hz) and any discrete freqviencies up to an upper discrete freqviency value of 20kHz. However, the method is not restricted to these freqviencies and freqviencies above the 20 kHz value may be chosen. Preferably, a freqviency range in which one performs the characterisation of the test system according to the embodiments of the present invention depends upon the frequency range of interest for a particular test system. It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments withovit departing from the spirit or scope of the invention as broadly described. The present embodiments are. therefore, to be considered in all respects as illustrative and not restrictive.

Claims

CLAIMS:

1. A method of characterising, in real time, an acoustic, mechanical, electro-acoustic or electro-mechanical system vmder test comprising the steps of: generating an inpvit signal having a plurality of selected discrete frequency components, injecting the inpvit signal into the system vmder measurement, and measuring a response signal of the acovistic. mechanical, electro- acoustic or electro-mechanical system to the injected signal.

2. The method of claim 1 wherein the system vinder test is compared with a reference system by comparing respective characterisations of the two systems.

3. The method of claim 2 further including the step of: comparing the response spectrum of the reference system with a predetermined response spectrvim.

4. The method of claim 3 further including the steps of adjusting an amplitude, a phase or freqviency of each frequency component of the generated inpvit signal and repeating the steps of injecting the signal and measuring the response of the reference system until the response matches the desired response, so as to achieve a calibrated input signal which causes the reference system to produce the predetermined response spectrum: and subsequently injecting the calibrated inpvit signal into the test system: measuring a response signal of the test system to the injected inpvit signal: and determining a ratio of test system response to reference system response in order to analyse characteristics of the test system.

5. The method according to any one of claims 2, 3 or 4 wherein the test system and the reference system are similar acoustic systems.

6. The method according to any one of claims 2, 3 or 4 wherein the test system and the reference system are similar mechanical systems.

7. The method according to any one of claims 2, 3 or 4 wherein the test system and the reference system are similar electro-acoustic systems.

8. The method according to any one of claims 2, 3 or 4 wherein the test system and the reference system are similar electro-mechanical systems.

9. The method of any one of the preceding claims wherein the test system comprises the reference system to which a pertvibation is applied.

10. The method of claim 9 wherein the reference system is a room and the test system is the room and a musical instrument within the room.

11. The method of claim 9 wherein the reference system is a model of an architectural structure and the test system is a modified version of the same model of the architectviral structure.

12. The method as claimed in any one of claims 1-8 wherein the reference system is a first musical instrument and the test system is a second musical instrument.

13. The method of claim 12 wherein the reference system is a violin having benchmark characteristic and the test system is a violin to be compared with the benchmark.

14. The method as claimed in any one of claims 1-8 wherein the reference system is a first benchmark amplifier and transducer and the test system is a second amplifier and transdvicer to be compared with the benchmark.

15. The method as claimed in any one of the preceding claims wherein the injected input signal is selected such that the frequency components of the response signal have freqviencies and amplitudes and/or phases directly correlating to an impedance spectrvim of the system being measvired.

16. The method as claimed in any one of claims 1 - 15. wherein the comparison is made by generating a spectrvim representing the ratio of the responses of the reference and test systems.

17. The method of claim 16 wherein the comparison is performed by calibrating the inpvit signal to provide a flat ovitput spectrum having a unity amplitvide for the reference system and then measuring the response of the test system to that same input.

18. An apparatus arranged to measvire an acovistic impedance spectrum or a transfer fvinction. including: signal generating means arranged to generate a signal having a plurality of selected, discrete freqviency components; the generating means being arranged to transmit the acoustic or vibrating signal into a system under measurement with high power throvigh a small apertvire. control means arranged to tailor the relative amplitudes and phase of the frequency components to control a freqviency spectrvim of the signal: and measurement means arranged to measure a frequency response of the system.

19. The apparatus as claimed in claim 18 wherein the control means is also arranged to apply an adjustment to said acovistic signal the adjustment being related to the reciprocal of the amplitvide of Fovirier components of the measvired freqviency response such that the freqviency response to the adjusted acovistic signal has a predetermined frequency spectrum.

20. The apparatvis of claim 18 or 19 further including display means arranged to display the measvired frequency response.

21. The apparatus of claim 18. 19 or 20. wherein the control means is a computer having data storage means to store the data relating to the measured freqviency response for the pvirpose of using the stored data later.

22. The apparatvis of claim 21 wherein the measvired frequency response of a reference system is stored in the data storage means and comparison means are arranged to compare the measured frequency response of the reference system with the measure frequency response of a system vinder test.

23. The apparatvis as claimed in any one of claims 18-22, wherein the comparison means is arranged to generate a spectrvim representing the ratio of the measured responses of the reference and test systems.

24. The apparatus as claimed in any one of claims 18-23, wherein the control means is arranged to adjust the generated signal such that the measvired response of the reference system has a flat spectrum, and the same signal is applied to the test system.

25. The apparatus as claimed in any one of claims 18-24, wherein the generating means includes transducer means selected from a displacement transdvicer. a velocity transducer and a force transdvicer.

26. The apparatvis as claimed in any one of claims 18-25 wherein the signal generating means includes transducer means having a high impedance acoustic outpvit driven by an electrical signal generator.

27. The apparatvis as claimed in claim 25 wherein the transducer means includes a speaker driving into an inverted horn with an outpvit impedance matched to the test system to produce maximum energy transfer with high acoustic output impedance.

28. The apparatvis as claimed in claim 27 wherein an impedance device is placed at the ovitput of the horn to increase the output impedance of the transducer system to fvirther isolate the system under measvirement from the signal generating means.

29. The apparatus as claimed in claim 28 wherein the impedance device is a narrow tube filled with glass fibre.

30. The apparatus as claimed in claim 26 wherein the transdvicer means comprises a small magnet located in proximity to an electromagnetic coil driven by an electrical signal to cavise the magnet to vibrate, the vibrations being transmitted to a mechanical system to which the magnet is connected.

31. The apparatus as claimed in any one of claims 18-30, wherein the signal generating means includes a synthesiser comprising a digital to analogue converter and a computer to generate the electric signal for driving the transdvicer means.

32. The apparatus as claimed in claim 31 wherein the synthesiser is arranged to synthesise the signal by performing an inverse Fourier transform on a frequency spectrum selected for the inpvit signal.

33. The apparatus as claimed in any one of claims 18-32. wherein the measurement means includes second transdvicer means arranged to detect an acoustic response signal and to produce an electrical analogue.

34. The apparatus of claim 33 wherein Fovirier transform means are arranged to transform the electrical analogue of the response signal into the freqviency domain.

35. The apparatvis of claim 34 wherein analogue to digital conversion means are arranged to convert the electrical signal to a digital signal and the Fovirier transform means operates on the digital system.

36. The apparatvis of claim 18-35, wherein the transfer function of the system vinder measurement is indicative of the acoustic gain of the system.

37. A method of measvirement of a transfer function of a mechanical or electro-mechanical system to which transdvicer means are linked for transmitting signals to and receiving signals from the mechanical or electro- mechanical systems, including the steps of: generating a predetermined reference signal having a plurality of selected discrete frequency components: driving the transdvicer means with said predetermined reference signal to thereby excite the mechanical system: receiving a response signal of the mechanical system via the transducer means: and analysing the response signals received by the transducer means.

38. The method of claim 37 wherein the analysis of the response signal involves first performing a Fourier transform to obtain a freqviency spectrum.

39. The method of claim 38 wherein the transducer means includes at least one or more transducers selected from a displacement transducer, a velocity transducer , a force transdvicer. a microphone, and a speaker.

40. A method of calibrating an apparatus for measuring an acoustic impedance spectrum or transfer function of a system comprising the steps of: a) generating an initial inpvit signal having a plurality of discrete freqviency components of selected freqviency and amplitude and/or phase and injecting the initial input signal into the system: b) measuring an initial response signal of the system to the initial inpvit signal: c) generating a correction signal by determining a ratio of a desired response signal to the initial response signal: and d) generating a calibrated inpvit signal as the product of the initial input signal and the correction signal.

41. The method of claim 40 further including the steps of, when after changing the input signal characteristic, the outpvit does not match the desired outpvit to the required degree; e) injecting the calibrated input to prodvice a resultant response signal into the system: f) generating a correction signal by determining a ratio of the desired response signal to the resultant response signal to obtain a further correction ratio: g) generating the prodvict of the further correction ratio and the calibrated input to provide a fvirther calibrated inpvit; and h) repeating the correction process of steps e) to g) as many times as is necessary to produce the reqviired response.

42. The method of claim 41 wherein a final calibrated input signal is used to measure the reference system with a pertubation applied.

43. The method of claim 42 wherein a final calibrated input signal is vised to measure a different system which is to be compared with the reference system.

44. The method as claimed in any one of claims 40-43. wherein the selected discrete frequency components define a predetermined freqviency spectrum and are selected in accordance with the frequencies of interest to be measured in the system.