EP1685741B1

EP1685741B1 - Sonic emitter arrangements

Info

Publication number: EP1685741B1
Application number: EP04798520A
Authority: EP
Inventors: Alastair Sonaptic Limited SIBBALD
Original assignee: Sonaptic Ltd
Current assignee: Cirrus Logic International UK Ltd
Priority date: 2003-11-18
Filing date: 2004-11-16
Publication date: 2013-04-03
Anticipated expiration: 2024-11-16
Also published as: GB0326807D0; EP1685741A1; WO2005051037A1; GB2408405A

Description

The present invention relates to sonic emitter arrangements, and it relates more particularly, though not exclusively, to such arrangements including miniature loudspeakers, such as microspeakers, and to their incorporation into portable electronic devices such as mobile cellular phones, digital cameras, portable games consoles and hand-held computers, or into miniature loudspeaker enclosures, such as earphones. The invention also encompasses devices, such as portable electronic devices, and miniature loudspeaker enclosures incorporating such emitter arrangements.
Portable electronic devices, such as those mentioned above, are becoming increasingly popular. For example, it is now commonplace for mobile phones and digital cameras to incorporate music players using the MP3 format, and some of the individual technologies are converging to create hybrid devices such as mobile phones combined with digital cameras or gaming consoles.
Portable electronic devices in general are well-suited for use with personal earphones or headphones, because they are often used in public places. Furthermore, it has not been possible (or worthwhile) hitherto for the housings of such devices to incorporate small loudspeakers which provide anything more than an extremely basic listening experience. There are several reasons for this.
Firstly, loudspeakers tend to add significantly to the physical size and cost of the end product. Secondly, the acoustical output quality of any loudspeaker is critically dependent on the way in which it is built and mounted. Furthermore, because of the restricted space available in a hand-held portable electronic device, for example, the loudspeakers must be placed relatively close together; typically less than 10 cm apart, or even less than 5 cm in a mobile cellular phone. Thus, when stereophonic audio material is played, the stereo effect is lost because the left and right channels are being reproduced from virtually the same point in space, whereas stereo is intended for playback on more widely spaced loudspeakers (typically about 2 metres apart) in order to create a spatial "sound image".
In order to address this emerging market for portable electronic devices with more sophisticated audio capabilities, loudspeaker manufacturers have recently developed extremely compact loudspeakers, known as "microspeakers", whose dimensions are similar to those of the driver units used in earphones. Such microspeakers are typically less than 20 mm in diameter and less than 5 mm in thickness. Despite their small dimensions, however, such microspeakers have significant power output capability, often with power ratings of several hundreds of milliwatts.
Microspeakers have been built into certain types of mobile phones by several manufacturers, but they are usually attached simply to the inner face of the phone's housing with their front surfaces exposed via a mesh or grille, or by several small holes in the housing. Although this is adequate for transmitting simple audio, such as ring-tones, to the listener, it is not adequate for delivering more sophisticated audio performance, such as 3D-positional audio for games, or stereo expansion for ring-tone and music playback.
The physical requirements for rendering 3D-positional audio via microspeakers are more demanding than they are for rendering stereo. Firstly, 3D-audio based on HRTF (Head-Related Transfer Function) processing relies on precise spectral filtering for its effect, and anything which introduces peaks or troughs into the system frequency response will degrade the perceived 3D-audio effects.
Secondly, an intrinsic and critical element of all loudspeaker-based 3D-audio is transaural crosstalk cancellation, as is described fully in GB 2,340,005 . This relates to the natural acoustic crosstalk which occurs around the head when an individual is listening to a pair of stereo loudspeakers. When the left-channel signal is emitted by the left-hand loudspeaker, it travels not only to the left ear, but also, a little later in time, crosses to the right ear (and vice versa). The brain recognises the high degree of correlation between the two signals, the primary signal and the crosstalk signal, and then correctly attributes their source to the left-hand loudspeaker, hence the sound is perceived to emanate from the left-side loudspeaker.
Without this crosstalk signal, the left-channel signal would have been delivered only to the left ear of the listener, just as it would have been using headphones. This feature, of delivering each sound channel only to its respective ear, is necessary for the correct operation of HRTF-based 3D-audio.
The presence of the transaural crosstalk signal inhibits 3D-audio effects, and so it must be cancelled by generating a signal which is equal in magnitude, and opposite in polarity, from the opposite loudspeaker, as is described in GB 2,340,005 .
In order to achieve adequate (say, 90%) crosstalk cancellation, the cancellation signal must match the crosstalk signal in magnitude and phase within fairly precise limits, about 3 dB of amplitude and ±20° of phase. This means that the relative time-of-arrival of the signals at the listener's ears must be synchronised very carefully. Anything which interferes with the integrity of the left and right-channel signals will degrade the crosstalk cancellation, and hence degrade the effectiveness of the perceived 3D-audio. During playback on portable devices, where the loudspeakers might only be 40 mm or so apart, the timing must synchronise to within a few microseconds for optimum effect. These constraints, and other considerations, create the following requirements for the successful rendering of 3D-audio using microspeakers.

1. The frequency response of the speakers should be relatively flat and smooth, without significant notches or peaks.
2. The high-frequency (HF) response of the speakers should not be substantially impaired or degraded in any way, especially for the satisfactory reproduction of music and MIDI-related sources. Ideally, the frequency response should extend, without significant reduction in amplitude, up to 10 kHz and preferably beyond.
3. The sound for each channel must be emitted from a single point-source, or as close to this as possible. If the sound source is relatively large, then the wavefronts will be distributed over the emitting area, and the sound wave will effectively be "dispersed" in time when it arrives at the listener's ears. This makes transaural crosstalk-cancellation extremely difficult or even impossible.
4. There must be minimal secondary emission, that is, sound emission which occurs from locations other than the primary microspeaker acoustic output port. For example, some sources of secondary emission in cell-phones include: (1) the casing and faceplate if the microspeaker is not well isolated from them; (2) the rear of each speaker, sound from which actually propagates through the alternate channel speaker, and (3) various holes in the casing if the rears of the speakers are not properly enclosed. There should be no additional emission ports other than the primary one.
5. The left- and right-channel sources must be placed as far apart as is practical in order to maximise the time-of-arrival difference between the left and right ears.
6. The left- and right-channels should be well matched in all respects, including loudspeaker matching with respect to both phase and amplitude.

There are two additional, physical requirements for deploying microspeakers in a slim casing, such as may be used (for example) for a cell-phone, to render 3D-audio. Firstly, it is required that the form-factor be suitable for the available space, and especially that the total thickness profile of the component package is adequately thin, typically less than 8 mm, in order to fit into the available depth of housing. Secondly, particularly for casings of the "clam-shell" or "ear flip member" type which can be opened and closed, it is necessary for the phone to reproduce 3D-audio successfully in both the open and closed positions.
One feature of all mobile electronic devices is the very limited space which is available for the user interface, primarily the graphics display, keypad and other controller devices; the graphics display in particular taking priority when housing space is allocated during design. Accordingly, there is little or no space for loudspeakers, however small, to be mounted on the front panel, facing the listener.
The various alternative options for incorporating microspeakers into a mobile phone include (a) mounting microspeakers to either side of the device, facing outwards to the listener's left and right sides, respectively; and (b) mounting the microspeakers internally of the device body, and delivering the audio output via a conduit, or pair of conduits, to respective output vents.
Of the alternative options, side mounting is the easier to implement, for example by mounting the microspeakers in sealed pods on opposite sides of the device. However, although the pods might be small, the overall additional bulk makes the phone body somewhat unsightly and the pods also detract from the smoothness of the body, making it less easy to move the phone into and out of a pocket, or a holster.
Internal mounting is thus a favoured option although, as stated previously, the mounting arrangements for the microspeakers are critical to the resultant performance, and a number of operational and design-related criteria must be complied with, depending on the audio performance required. Also, as mentioned above and as discussed in more detail hereinafter, particular considerations arise in relation to cell-phones which are designed in the form of clam-shell structures, comprising a pair of hinged body units (effectively, a body with a corresponding lid) which are closed together when not in use, and which are unfolded for use when required.
In any event, if internally-mounted microspeakers are to be used, it is axiomatic that a conduit of some sort must be used to convey the audio energy to the outside world. For microspeakers mounted on an internal chassis or frame within an outer housing, suitable conduits can in principle be formed within the housing to which the front (sound-emitting) surfaces of the microspeakers are exposed, and emission apertures can be cut into the housing, at some small distance from the microspeakers, to act as sound outlet ports, linked via the conduit to the microspeaker.
Problems arise with such arrangements however, since the conduits and any "dead-space" volumes adjacent the microspeaker form resonant acoustic cavities. Such cavities generally exhibit Helmholtz-resonator characteristics (described in Appendix 1), and thus create undesirable peaks and troughs in the emitted sound spectrum. Moreover, the housing itself is exposed directly to a considerable amount of sound energy present in the cavities, to which it is partly transparent, and hence the cell-phone housing itself becomes an undesirable secondary emission source for both microspeakers, further reducing the sound quality, and significantly impairing the effectiveness of 3D-positional audio.
One method of reducing the peaks and troughs caused by the resonant conduit would be to simply fill the cavities with sound damping material. Such an approach, however, inevitably results in the absorption of a considerable amount of the sound energy indiscriminately across the spectrum, and substantially reduces the emitted volume.
Various prior-art proposals have sought to address the acoustic limitations which result when microspeakers and other transducers, such as piezoelectric monomorphs, are mounted into a housing adjacent to one or more emission apertures; such proposals mostly involving attempts to insert peaks and troughs, and/or to re-locate unwanted peaks and troughs, into selected areas of the sound spectrum, in an attempt to compensate for the peaks and troughs associated with the aforementioned Helmholtz-type cavities and/or to lift or depress an amplitude versus frequency characteristic which falls below or exceeds, respectively, a desired sound level. Such approaches are acceptable where the principal interest lies in a limited part of the spectrum (e.g. for the reproduction of ring tones and the like) or where the overall amplitude-frequency characteristic does not need to be controlled to the extent required for accurate 3D-positional performance.
Such proposals are disclosed, for example, in WO 83/023304 ; US 5,953,414 ; US 6,324,052 and WO 02/340030 , none of which successfully addresses the problem of controlling the overall amplitude-frequency characteristic to an extent sufficient to render music and/or 3D-positional audio performance of acceptable quality.
WO 2004/030402 , which was published later than the priority date claimed herein, and is thus referred to hereinafter as "the intermediate document" describes a twin-resonant acoustic structure for incorporating a microspeaker into a cell-phone housing with the intention of providing good alerting performance and extended frequency response in the voice frequency range (between 300 Hz and 3400 Hz). This proposal recognises the problem which arises when a microspeaker is coupled via a conduit to an opening in the housing, in that the frequency response becomes dominated by a resonant peak, such that the mid-frequency and high-frequency responses are very strongly attenuated, and discloses the use of a first "forward tuning volume", which comprises, in effect, a Helmholtz cavity, adjacent the face of a loudspeaker. This volume is coupled via a passage both to an opening in the housing, and to a second forward tuning volume, in fluid communication with the passageway, lying between the first forward tuning volume and the opening in the housing.
Without the second forward tuning volume, the amplitude-frequency characteristic attributable to this arrangement, however, contains a large resonant peak at about 3.5 kHz, followed by a significant trough at about 5.2 kHz, and exhibits progressive high-frequency (HF) attenuation. These large spectral perturbations are not acceptable for 3D-positional audio, nor for music reproduction, both of which require a fairly flat spectral response.
When the second forward tuning volume is added to the acoustic structure, another resonant peak is added to the response in the region of 6 kHz to 7 kHz (which is its purpose), but the spectral trough at 5 kHz still remains and the HF performance decreases even further. Thus, although this expedient can be used to increase the apparent volume of an alerting tone in the 6 kHz to 7 kHz region, these additional spectral fluctuations and the further HF reduction would further degrade the reproduction characteristics needed for 3D-positional audio and music.
It is an object of this invention to provide sonic emitter arrangements, comprising one or more miniature transducers (such as microspeakers) disposed within a housing, together with associated control means capable of so controlling the amplitude versus frequency characteristics of emitted sound as to maintain amplitude excursions attributable to conduits and/other means for conveying the sound from the transducers out of the housing to an extent that renders such arrangements capable of emitting sound of acceptable quality for the reproduction of music and/or for 3-D positional audio imaging.
From one aspect there is provided a sonic emitter arrangement comprising at least one sonic transducer encased within a housing dimensioned to be portable or wearable; said transducer having a sonic emission surface, and the arrangement further comprising at least one conduit linking said emission surface to an outlet through which sound produced by said transducer can be emitted from said housing; wherein at least one dimension of part at least of the length of said conduit is flared so as to increase toward said outlet, thereby to influence the amplitude versus frequency characteristic of sound emitted from said outlet.
According to the invention, said housing encases first and second sonic transducers; said transducers each having a respective sonic emission surface, and wherein the arrangement further comprises a respective conduit linking each said sonic emission surface to a respective emission outlet through which sound produced by said transducer can be emitted from said housing; and wherein each said conduit is flared so as to increase toward its respective outlet, thereby to influence the amplitude versus frequency characteristic of sound emitted from said outlets.
Another example, utilising a single sonic transducer, provides first and second conduits linking the transducer's sonic emission surface to respective emission outlets through which sound produced by said transducer can be emitted from said housing; and wherein each said conduit is flared so as to increase toward its respective outlet, thereby to influence the amplitude versus frequency characteristic of sound emitted from said outlets.
In the arrangements described in both of the two immediately preceding paragraphs, it is preferred that substantially identical flaring is applied to each of said conduits.
The flaring may be applied over part only of the length of a conduit. In such circumstances, it is preferred (though not essential) that such flaring occurs adjacent the outlet.
The flaring is preferably smooth and may conform to a substantially linear profile or follow an exponential or other curvilinear form, though it may alternatively, or in addition, incorporate one or more discrete steps.
Arrangements of the kind described in the foregoing paragraphs are efficient, can be implemented at relatively low cost, and can be readily adapted to a wide range of different transducer types and sizes, though it is preferred that a miniature loudspeaker, such as a microspeaker, is used.
In further preferred forms of the invention, the arrangement comprises one or more acoustic resonant absorbers linked to said emission surface and/or to at least one conduit, in order to further influence said characteristic. By this means, fine-tuning and/or additional compensation for unwanted characteristics can be achieved.
In one embodiment, at least one of the acoustic resonant absorbers comprises a Helmholtz resonator, such resonators being relatively simple to construct and exhibiting reliable performance over a wide range of operating conditions.
In another embodiment, at least one of the acoustic resonant absorbers comprises a quarter-wavelength tube, channel or groove device, or plurality thereof; such resonant devices being relatively simple either to mould in plastic materials or to cut in metallic materials, and therefore well-suited to mass-production means.
Where quarter-wavelength absorber tubes are utilised, it is preferred to provide a distributed array of quarter-wave channels conforming substantially to a concentric elliptical array pattern.
Any arrangement in accordance with the invention may conveniently be fabricated, at least in part, from superposed laminar components; such components being of metal, plastics or any other material which can be readily machined, moulded or otherwise formed to the required tolerances and which, when assembled, exhibits suitable acoustic performance.
Where such laminar components are used, edge-emission conduits can conveniently be formed by opening an otherwise enclosed aperture, such as an aperture located substantially centrally within a plate, through to an edge of the plate by removal of the plate material, bearing in mind that other laminae will lie above and below the plate in question, thereby defining part at least of the conduit.
If desired, any one or more of the resonant absorbers may be fabricated of, or may contain or have associated therewith, acoustic damping material of any convenient kind, such as cotton fibre wool or tissue paper.
Arrangements according to any embodiment of the invention may be incorporated into electronic devices such as mobile telephones, digital cameras, mobile games consoles or portable sound and/or multimedia equipment.
In such circumstances, the said housing usually serves as the overall housing for the invention. Typically, such a device incorporates two or more such arrangements, preferably matched in performance, permitting the device to exhibit sophisticated audio performance characteristics, such as 3D-positional audio or enhanced stereophonic sound.
One or more arrangements according to any example of the invention may be incorporated into portable and/or wearable loudspeaker enclosures such as earphones, intended to be worn by a listener.
It will be appreciated from the foregoing statements of the invention that the invention calls for at least one dimension of part at least of a conduit acoustically coupling a microspeaker to an emission outlet to be flared so as to increase toward said outlet, thereby to influence the amplitude versus frequency characteristic of sound emitted.
The principle of deploying a flared conduit, as used in the present invention, should not be confused with other audio applications in which flared devices are known. For example, a flared trumpet device was used in the earliest gramophone machines as a mechanical transformer to transmit efficiently the mechanical vibrations from a small needle into a large-area, resonant and reflective surface, thus increasing its loudness. Also, flared horn-type arrangements are widely used to increase the efficiency of loudspeakers used in public-address systems, by acting as an acoustic transformer to match the impedance of the piston-like diaphragm to that of the air.
In contrast to these known applications of flared acoustic devices for impedance matching and the like, the present invention employs a flared topography for a very different reason, namely to create a minimally resonant emission conduit linking the emission surface of the microspeaker with the emission aperture, in order to provide a desired, flat frequency response emission characteristic.

Minimal Resonance Principle

Embodiments of the present invention were devised by designing and constructing the relevant elements of a Helmholtz-type conduit (as defined in Appendix 1 hereinafter) so as to minimise its resonant properties, according to the following two principles devised by the inventor.
Firstly, it was decided to reduce the magnitude of the resonant peak by minimising the Q factor of the resonance, within practical constraints, in order to minimise the "insertion response"; a term which is defined hereinafter. Secondly, it was decided to locate the spectral peak associated with the residual resonance at as high a frequency as would be practically possible, in order to avoid the HF attenuation related to the prior-art Helmholtz-type cavities.
By inspection of equations (1) and (2) in Appendix 1, it can be deduced that both of these objectives can be achieved by the following principles:

1. Minimising the internal volume of the conduit, V;
2. Minimising the length factor, L; and
3. Maximising the sonic emission area, S.

Conformance to these principles is not straightforward, however, because these parameters are mutually interdependent. For example, in a conduit of constant rectangular section, if it were required to increase the emission area, S, then this could be achieved by increasing the thickness of the conduit. An increase in the thickness of the conduit would proportionately increase the emission area, S. However, it would also proportionately increase the volume, V, and this would create a conflict in satisfying the above principles, in which S must be maximised, and V must be minimised. This conflict also occurs if the width of the conduit were to be increased.
Embodiments of the present invention conceived to satisfy the minimal resonance principles, above, utilise a "flared emitter" structure. By minimising the cross-sectional area of the emission conduit adjacent the microspeaker, and maximising its area at the point of emission, then the internal volume, V, is minimised whilst maximising the sonic emission area, S.
One implementation of these principles is to minimise the frontal volume adjacent the microspeaker, with a small, constant cross-sectional area in this region, from which is provided a flared conduit to the emission aperture. Another approach is to provide a progressive increase in cross-sectional area throughout the length of the conduit between the microspeaker and the emission aperture. By this means, the emission conduit less resembles a Helmholtz resonator, but rather becomes more like an open tube structure, and is minimally resonant.
In practice, for an orthogonal emission arrangement, it will be appreciated that the length factor, L, cannot be less than the radius of the microspeaker, or thereabouts.
In order that the invention may be clearly understood and readily carried into effect, certain embodiments thereof will now be described, by way of example only, with reference to the accompanying drawings, of which:

Figures 1(a) and 1(b) show an idealised representation of the relationship between a microspeaker and an outlet slot, or port, for the acoustic emissions therefrom;
Figures 2(a) and 2(b) show a flip-top, or "clam-shell", type of mobile phone in closed and open conditions respectively;
Figures 3(a) and 3(b) are similar to Figures 2(a) and 2(b) respectively, but illustrate a convenient disposition and configuration for outlet ports conducive to use with embodiments of the present invention;
Figure 4 shows, in exploded diagrammatic perspective, part of an arrangement, in accordance with one example of the invention, utilising a plate formed with a flared emission slot;
Figure 5 shows an amplitude-frequency response characteristic measured in relation to an arrangement of the kind shown in Figure 4, and a characteristic, similarly measured, for a microspeaker alone for comparison;
Figure 6 shows an idealised representation of the relationship between a microspeaker and an outlet slot, or port, to illustrate measurements of insertion loss in acoustic emissions;
Figure 7 shows an insertion loss characteristic measured in relation to an arrangement of the kind shown in Figure 4;
Figures 8(a) and 8(b) show a flared emitter plates incorporating respectively a Helmholtz resonator and quarter-wave stub resonators, each intended individually to replace the plate shown in Figure 4, in arrangements in accordance with second and third embodiments respectively of the invention;
Figure 9 shows an amplitude-frequency response characteristic measured in relation to an arrangement including a plate of the kind shown in Figure 8(a), and a characteristic, similarly measured, for a microspeaker alone for comparison;
Figure 10 shows an insertion loss characteristic measured in relation to an arrangement including a plate of the kind shown in Figure 8(a);
Figure 11 shows an amplitude-frequency response characteristic measured in relation to an arrangement including a plate of the kind shown in Figure 8(b), and a characteristic, similarly measured, for a microspeaker alone for comparison;
Figure 12 shows an insertion loss characteristic measured in relation to an arrangement including a plate of the kind shown in Figure 8(b);
Figure 13 shows a dually flared emitter plate, incorporating an array of quarter-wave stub resonators, intended to replace the plate shown in Figure 4, in arrangements in accordance with a fourth embodiment of the invention and utilising a single microspeaker; and
Figure 14 defines certain parameters used herein in describing Helmholtz resonator characteristics.

The invention aims to provide an efficient sonic emitter structure with physical and acoustic properties that are especially well-suited (inter alia) to application in cell-phones and other portable devices for the reproduction of stereophonic music and, especially, 3D-positional audio.
In terms of physical properties, the invention permits construction of the components and their assembly into a thin, substantially planar configuration, and provides sonic emission from an aperture, or port, 10 formed in one edge of its structure and orthogonally oriented with respect to the emission plane of an internal microspeaker 20 as shown in Figure 1, in which L-L' represents the propagation (sound emission) axis, and C-C' represents the central emission axis of the microspeaker. This is especially beneficial for use in clam-shell type cell-phones where the flat form factor enables deployment in a thin structure, in pairs (Figure 1(b)), and where the edge emission property, being front-back symmetrical, lends itself to operation in both open and closed positions.
This latter feature is an important attribute of certain embodiments of the invention. The clam-shell phone structure (see Figures 2 and 3) comprises a pair of hinged body units (a body 100 with a corresponding lid 200) which are closed together when not in use, and which are unfolded for use when required. If a pair of microspeakers were to be mounted inside the upper surface of the lid unit 200 directly behind respective sound emission apertures 300 and 400, the apertures would be exposed towards the listener when held in the hand, in the closed mode, as shown in Figure 2(a). However, with the lid 200 opened in order to use the cell-phone, then the emission apertures 300 and 400 would face away from the listener (Figure 2(b)). Furthermore, the emission apertures would be occluded by the lid 200 itself, thus reducing the perceived volume and high-frequency content. This represents a major problem for the use of video games on a clam-shell type phone, in which the user will wish to use the phone in open mode, in order to see the visual display, but also will require 3D audio playback from the microspeakers. This problem is not solved by incorporating the speakers into the lower, body unit, because the user's hand is likely to occlude any emission apertures which would be present therein.
However, by engineering the embodiment of the invention of Figure 1(b) into a cell-phone body so as to provide edge emission from a pair of rectangular slit apertures such as 500 formed in opposing sides of the lid structure 200, as depicted in Figure 3, then the emission is symmetrical in both the open and closed positions, as shown in Figures 3(a) and 3(b) respectively, which is ideal for the reproduction of 3D-sound in both circumstances.
A further valuable property of the invention relates to the minimal acoustic dispersion of the propagated signal. Acoustic dispersion occurs when there is not a single, well-defined path from the source to the listener, as is the case when the sound emitting device is not a point source. If it were, then the acoustic path length to the listener's ear would be well-defined. In practice, however, microspeaker sound is emitted from a finite area. If this area were divided into a matrix of elemental areas, then each element would have a slightly different path length to the listener's ear. This variation in path lengths represents the acoustic dispersion range. For 3D-audio, as described previously, it is important to ensure that the wavefronts arrive separately and synchronously at the listener, and hence with minimal dispersion.
However, as described already, and illustrated in Figure 1, a feature of embodiments of the invention is that the aperture 10, from which sonic emission occurs, can usefully be formed as a narrow, rectangular slit, typically 2 mm or less in width and 10 mm or so in length. Consequently, in a clam-shell cell-phone configuration, if the two apertures such as 500 in Figure 3 are arranged and oriented so as to lie along the opposing, lateral edges of the lid 200, as described above for front-back symmetry, they also present the listener with a pair of suitably narrow sound sources, maximally separated by the width of the clam-shell lid 200. The resultant vertically oriented emission slits such as 500 are near perfect for 3D-audio reproduction, owing both to the maximal separation and to the minimal acoustic dispersion associated with the very narrow emission area.
Figure 4 shows, in exploded diagrammatic form, the principal components of a sonic emitter arrangement in accordance with one example of the invention. The arrangement is, in this example, constructed in laminar form from overlaid elements of aluminium sheet stock, and incorporates a 16 mm diameter, 2 mm thick microspeaker unit 20 (Foster type 364870), with an active emission surface area of 9 mm diameter located centrally thereof, mounted with epoxy sealant into an aluminium flange plate 22 of thickness 4 mm, measuring 28 x 28 mm and formed with a central aperture of diameter 16.2 mm.
A plate 30 is secured to the plate 22; the plate 30 being formed as shown with a conduit configured as a flared slot 32 to provide a gradual transition from a 9 mm wide central aperture, overlying the active transducer area, to a 16 mm wide emitting aperture open at one edge of the plate. In order to avoid any impedance discontinuities, a linearly flared profile was adopted for the slot 32 in this example, as shown. Such a conduit is referred to hereinafter for convenience a "9-16 flared emitter", referring to the 9 mm central aperture diameter and the 16 mm length of the emitting edge, respectively.
It is important to note that flaring characteristics other than linear (for example exponential or other curvilinear characteristics) can be used if preferred. Moreover, the flaring need not be smooth, and thus steps or other discrete dimensional variations can be used instead of, or in addition to, smooth flaring profiles. Still further, and in relation to any flaring characteristic used, the flare need not extend all of the way from the central aperture to the edge of the plate 30. In some embodiments, the width of the slot 32 initially is maintained at 9 mm for a predetermined distance away from the central aperture; the flare then commencing and continuing to the edge of the plate 30.
In the orthogonal emission format (as shown in Figure 1), where the conduit is relatively flat and wide, there is clearly more scope to flare the width dimension of the conduit, rather than the height dimension. It is not important which of the two dimensions is flared (or indeed both might be flared); what is important is that the cross-sectional area of the conduit be increased along at least a part of its path from the microspeaker surface to the outlet port, and that there is no substantial decrease in cross-sectional area at any point (because this would create an increase in the Helmholtz-like resonant properties of the conduit).
Also secured to the plate 22 is a rear enclosing cavity 40 of volume about 2 ml. The cavity 40 is coupled to the rear surface of the microspeaker 20, and comprises an 18 mm diameter tube 42 which, in this embodiment, is 7.9 mm in length and is provided with flanges 46, 48 at either end, and a plain sealing plate 44 secured to that flange (48) of the tube 42 remote from the microspeaker 20. In general, that surface of the microspeaker 20 which is intended to be the sound-emitting surface is herein designated the "front" surface; the other (parallel) major surface of the microspeaker correspondingly being designated as the "rear" surface.
The upper surface of the plate 30 is covered with a simple flat blanking or capping plate 34.
In this example, most of the laminar components are made of aluminium sheet stock of thickness 2 mm, and the use of such 2 mm thick material for the 9-16 flared emitter plate 30 provides an emission area of 32 mm². For this constant thickness plate, the intrinsic volume is linked to the aperture area, because they are both dependent upon the thickness of the plate.
Figure 5 shows both the measured, on-axis amplitude-frequency response characteristic 31 of the arrangement shown in Figure 4, and thus incorporating a conduit comprising the 9-16 flared emitter 30, against, as a reference, the original on-axis response 21 of the microspeaker 20 without the flared emitter in place. This characteristic (31) compares favourably with those attributed to the proposals of the aforementioned intermediate document, particularly in respect of the performance required for stereophonic music and 3D-positional audio. Compared with the characteristics attributed to the aforementioned proposals, the main resonant peak has been reduced in magnitude and distributed over a wider frequency range. Moreover, in an arrangement according to this example of the present invention, the principal peak now lies at 5 kHz, well above the voice-band (300 Hz to 3.4 kHz) as opposed to lying in the vicinity of 3.2 kHz. Compared to the aforementioned proposals, the invention also improves, by around 3 dB, the gain in the important region extending from 1 kHz to 1.5 kHz.
Equally importantly, the high-frequency response associated with use of the flared emitter is usefully sustained to 10 kHz and beyond. As can be seen from Figure 5, at 10 kHz, the response is only about 6 dB less than the reference microspeaker itself.
In terms of acoustic properties, therefore, use of this embodiment of the invention provides an excellent high frequency response, suppressing the Helmholtz resonance otherwise present intrinsically in its structure.
It will be appreciated that a complete arrangement configured for stereophonic sound reproduction and/or 3D-positional audio will comprise a housing that incorporates a pair of constructions such as that shown in and described with reference to Figure 4, and that the emission slits for the two conduits will typically be disposed to either side of the housing, as shown, for example, in Figures 1 and 3. It will also be appreciated that, in such circumstances and order to provide matched performance from the two microspeakers, the two conduits will in general be configured to exhibit substantially identical flarings.
In another embodiment of the invention, described hereinafter, the arrangement also provides a relatively flat, smooth frequency response by means of the additional integration of a resonant absorber device.
In contrast to prior-art devices and other proposals intended to provide a signal boost at specific, narrow regions of the spectrum in order to compensate for an otherwise poor HF response, or to accentuate a ring alert, certain embodiments of the present invention seek to provide a relatively flat frequency response over an extended part of the audible frequency spectrum.
In contrast to prior-art devices and other proposals employing resonant volumes and chambers to amplify the sonic output at certain frequency ranges by resonance, the present invention uses a resonance-suppressing structure.
In contrast to proposals employing Helmholtz-type resonant cavities to amplify the transmitted signals at certain frequencies, some embodiments of the present invention use a resonant cavity structure to selectively suppress, by absorption and/or attenuation, residual device resonance at certain frequencies. Furthermore, the resonant cavity structure is not restricted to a Helmholtz-type configuration, but one preferred embodiment of the present invention employs an array of quarter-wave stubs as a resonant absorber array.

Insertion Response

A simple method of defining clearly the properties of the present invention, and for quantifying its beneficial characteristics, relates to the concept of the "Insertion Loss" of a component. The Insertion Loss of a device in a transmission circuit is defined as the difference between the transmission response without the device in place, and the transmission response with the device in place. More rigorously, we can define an "Insertion Response", which will define the change in the transmitted spectrum owing to the presence of a device (rather than just a simple gain-factor at a single, specified frequency).
This concept of an Insertion Response is a powerful method for characterising the acoustic properties of a device, because the resultant data are independent of the drive transducer. Figure 6 shows a simple method for measuring the Insertion Response of a microspeaker emission conduit. First, as shown in Figure 6(a), a microspeaker is mounted on to a small, fixed sealed rear enclosure having a volume that is typical of the final application (for cell-phones, this is typically 2 ml). The microspeaker characteristics are measured by conventional impulse or frequency sweep methods in an anechoic chamber, using a reference-grade microphone in an on-axis (C-C') position (Figure 6(a)) directed at an emission outlet, and typically at a distance of about 5 or 10 cm. Next, as shown in Figure 6(b), an emission conduit is attached to the front surface of the microspeaker enclosure and the measurements are repeated. Here, L-L' represents the on-axis direction of propagation, now orthogonal to the microspeaker central axis, C-C'. The Insertion Response is calculated by subtracting the second response from the first, yielding on-axis results, and can be plotted as a function of gain (dB) against frequency, as illustrated later. In all of the examples described later, Insertion Responses have been calculated using the on-axis data for both the emission conduit measurements and the reference microspeaker measurement, where this is defined to be aligned along the central emission axis of the transducer or conduit.
Embodiments of the invention permit simultaneous adherence to the following operational criteria:

1. The Insertion Gain characteristic should be smaller than +10 dB within the voice spectrum range of 300 Hz to 3400 Hz.
2. The Insertion Loss characteristic should be smaller than -10 dB in the frequency range 500 Hz to 10 kHz (and, ideally, to about 15 kHz).

From practical experimentation, it has been found that these criteria provide acceptable quality music reproduction and the effective rendering of 3D-positional audio on a cell-phone platform; and Figure 7 shows at 33 an insertion response curve for the arrangement of Figure 4.
It is a further objective of the invention to produce a still higher-quality sonic emitter by further reducing or even eliminating the still-present, residual resonant peak present in the emitted spectrum of the flared emitter. An obvious approach is to insert damping material into the conduit. However, as already noted, this expedient significantly reduces the emitted volume level across the spectrum, and the microspeaker cannot be simply driven harder to compensate for the reduction in sound volume if it is already operating near its maximum output capability.
A further embodiment of the invention therefore utilises a compensating resonant absorber cavity, linked to the flared emitter conduit 30, 32; tuned to the unwanted residual resonant peak, and having an appropriate Q factor. The compensating resonator absorbs acoustic energy specifically at the relevant resonant frequency and to the desired spectral profile, without reducing the sound output across the spectrum, as would happen if damping material were simply introduced into the cavity. The inventor has discovered that either a Helmholtz-type resonant absorber or, alternatively, one or more miniature quarter-wave tube-type resonant structures can be used to reduce or eliminate residual resonant peaks in the response of the flared emitter.
Accordingly, an arrangement comprising a compensated structure of this kind using a Helmholtz-type absorber was fabricated using stacked laminar components as shown in Figure 4, but with the plate 30 replaced by a plate 50 shown in Figure 8(a), in order to address the residual spectral peak at about 5 kHz in the characteristic 31 of Figure 5.
The new "integrated" sonic emitter plate 50 was fabricated by integrating a flared emission conduit and a suitable Helmholtz absorber into the same plate, and is therefore readily manufactured. The Helmholtz parameters to absorb the residual 5 kHz resonance were calculated, and a corresponding circular cavity 54 and a linking channel 56 were incorporated into the plate 50, which also was formed with a flared 9-16 emitter conduit 52, with the Helmholtz absorber neck coupled to the flared conduit adjacent the microspeaker surface, opposite the path leading to the emission aperture, as shown in Figure 8(a). The circular cavity 54 is 7 mm in diameter, and is linked to the flared conduit by a 2 mm wide channel 56, which is 2 mm deep (the thickness of the plate).
A cotton damping material (not shown) was used to fill the 7 mm diameter absorbing cavity 54 in order to match the Q factor of the residual resonance, thereby providing a damped resonant cavity; a resonant absorber.
The 9-16 flared emitter 50 with integral Helmholtz absorber, as shown in Figure 8(a), was inserted into the arrangement of Figure 4 in place of the plate 30 and was characterised using the same method as for the previous devices, with the results being shown in Figure 9.
Figure 9 shows both the measured frequency response 51 of the flared emitter with integrated Helmholtz absorber, and, as before, the original, on-axis response 21 of the microspeaker without the flared emitter in place as a reference. Although the response is not perfectly flat, it more closely resembles the "pure" response of the reference microspeaker on its own. Figure 10 depicts at 53 the Insertion Response of the flared emitter with integrated Helmholtz absorber.
This characteristic has proven excellent for the reproduction of music and for rendering 3D-positional audio; the HF response extending beyond the plot shown, and the overall response being 800 Hz to 15.8 kHz ±4 dB, and totally free from any sharp peaks or troughs. At 10 kHz, the response is only about 2 dB less than the reference microspeaker level. By comparison, a response attributed to the Helmholtz resonator-based arrangement proposed in the intermediate document was about 30 dB below this reference point at the same frequency.
In order to further simplify the mass production of the invention by dispensing with the need to use cotton damping material (or alternative damping materials), research was carried out to discover whether one or more quarter-wave tube absorbers could be used as an alternative to the Helmholtz-type resonant absorber, above.
This was done by milling narrow, rectangular channels of rectangular profile into the upper surface of a 9-16 flared emitter plate, such that when the capping plate 34 was added to the surface, the channels formed closed tubes ("stubs"), with open ends exposed to the flared conduit, adjacent the microspeaker surface, and opposite to the path leading to the emission aperture. The resulting plate 60, as shown in Figure 8(b), comprises a 9-16 flared emitter slot 62, as before, and in addition a plurality of stub absorbers 64a to 64e. The stub absorbers could, of course alternatively be formed in the underside of the capping plate 32. As a further alternative, some stub absorbers could be formed in the plate 60 and some in the plate 32, or part of some or each of the stub absorbers could be formed in juxtaposed, facing surfaces of each plate.
Use of this principle was found to be successful, with resonant absorption occurring at predictable frequencies. It was also found that, if the channels were more than several square millimetres in cross sectional area, the absorption was greater than required and not well matched to the Q factor of the residual resonance. Also, these relatively large section area channels often introduced artefacts in the response in the form of side-lobe accentuation adjacent the absorption frequency. By reducing the cross-sectional area of the quarter-wave channels to 1 mm² or less, it was found that the resistive losses caused by fluid interaction with the sidewalls became relatively greater, thus reducing the Q factor and the artefacts. However, the frequency band over which absorption took place was also reduced to several hundred Hertz, whereas it was required to effect absorption over about 1 kHz, bandwidth, as indicated by the peak in Figure 9.
Accordingly, a distributed array of five quarter-wave channels 64 (a) to 64(e) was milled onto the surface of a 9-16 flared emitter 60, 62, as shown in Figure 8(b). Each stub channel 64 was 1 mm in width, and 0.7 mm in depth. In order to expose all of the channels 64 correctly to the emission conduit adjacent the microspeaker, and to pack them in the restricted space, a concentric elliptical array pattern was used. Initially, the lengths of the channels were calculated so as to provide absorption at 500 Hz intervals from 4.1 kHz to 6.1 kHz. This was successful, but a slightly improved performance was obtained by adjusting the length of one of the channels (empirically) to a higher frequency, owing to the interaction between them. The resulting five-stub array was made according to the following frequency and length parameters, where the stub reference letters are shown in Figure 8(b).

Stub Reference Frequency (Hz) Length (mm)

64(d) 6100 14.1

64(b) 6989 12.3

64(a) 5100 16.8

64(c) 4600 18.6

64(e) 4100 20.9
This proved to be very successful, as the results shown in Figure 11 demonstrate.
Figure 11 shows both the measured frequency response 61 of the flared emitter 60 with integrated quarter-wave array absorber 64a to 64e, and as before, as a reference, the original, on-axis response 21 of the reference microspeaker without the integrated flared emitter in place.
The response 61 follows closely the "pure" response 21 of the microspeaker on its own. It is excellent for the reproduction of music and for rendering 3D-positional audio. The HF response extends beyond the plot shown, being 800 Hz to 15.30 kHz ±3 dB, and is totally free from any sharp peaks or troughs. Figure 12 shows at 63 the insertion loss response of the flared emitter with integrated quarter-wave array absorber.
It will of course be appreciated in relation to embodiments of the invention utilising modified emitter plates such as 50 (Figure 8(a)) and 60 (Figure 8(b)) that, as mentioned above in relation to the basic construction of Figure 4, a complete arrangement configured for stereophonic sound reproduction and/or 3D-positional audio will comprise a housing that incorporates a pair of microspeakers, each associated with a respective flared and compensated conduit, and that the emission slits for the two conduits will typically be disposed to either side of the housing, as shown, for example, in Figures 1 and 3. It will also be appreciated that, in such circumstances and order to provide matched performance from the two microspeakers, the two conduits will in general be configured to exhibit substantially identical flaring and resonant absorption characteristics.
In certain circumstances, it may be desirable for each plate, such as 50 or 60, to incorporate more than one resonant absorber. In such circumstances, the resonant absorbers may or may not be of the same kind; i.e. a Helmholtz-type absorber and one or more quarter-wavelength channels may be coupled to the same flared conduit.
In general, it is noted in relation to the invention that:

1. The parameters of the sonic emitter can be selected and adapted so as to enable its integration into a wide range of differing device body sizes and shapes.
2. Although aluminium was used for device fabrication in the above examples, it is equally effective to use other materials, such as plastics materials, although it is much preferred to use a rigid plastic material rather than a soft plastic (such as polystyrene) in order to minimise secondary emission.

It will be appreciated that the invention may be implemented in other forms and utilised in other applications than those specifically described herein, and that accordingly the examples used herein are not intended to limit the scope of the claims hereof. For example, embodiments of the present invention discussed hitherto in this specification have been principally focussed upon the use of arrangements utilising a pair of microspeakers, each producing sound intended to emerge from a respective output aperture so as to provide stereophonic and 3D sound capabilities of acceptable quality. In some other embodiments of the invention, however, aimed at meeting different criteria, useful results have been obtained by coupling sound from a single microspeaker via respective conduits to two or more emission apertures. In such embodiments, it is usual that both (or all) apertures are coupled to the front surface of the microspeaker. Such embodiments of the invention find application, for example, to mobile telephones intended to be operable "hands-free" and which thus may be placed on a stand or other support so that a user can (for example) conduct a call or listen to music, using the telephone, whilst doing something else.
Figure 13 shows a plate member 70 for use in such an embodiment of the invention; the plate member 70 being generally similar to plate member 60 of Figure 8(b), but formed with dual flared conduits 72 and 74, each linking the output surface of a single sonic emitter, such as a microspeaker (not shown), to a respective emission slot region 76, 78. The flared conduits 72 and 74 are merged in a centrally located region 73 which overlies the emission surface (not shown) of the microspeaker. It will be appreciated that, apart from the fact that two conduits (72, 74) are used to link a single sonic transducer to a pair of emission slots, the construction of arrangements using plates of the kind shown at 70 in Figure 13, and their incorporation into electronic devices utilising them, is substantially the same as for other embodiments of the invention disclosed herein.
Referring again to Figure 13, it will be observed that the plate 70 also includes an array 80 of quarter-wave stub absorbers; the array in this case comprising five stubs, 80(a) to 80(e), of different lengths, split for convenience into groups of two and three disposed respectively to either side of the aforementioned central region 73 at which the dual conduits merge.
In one embodiment, the merged central region 73 of the conduits 72, 74 is 8 mm in width, and the width of each conduit then flares linearly and smoothly to about 20 mm at its respective emission aperture region. In that embodiment, and in order to suppress an intrinsic resonance of the system, occurring at about 4360 Hz, the array 80 of five quarter-wave stubs, split into groups as aforesaid for convenience, was integrated into the plate. Each stub comprised a channel 1.0 mm wide and 0.8 mm deep, but with their respective lengths matched to the required absorption frequencies in accordance with the following table:

Stub Reference Frequency (Hz) Length (mm)

80(a) 3360 25.5

80(b) 3860 22.2

80(c) 4360 19.7

80(d) 4860 17.6

80(e) 5360 16.0
It may be of assistance to the reader to provide certain information about the properties of Helmholtz resonators and quarter-wave tubes, and thus the following appendices are furnished for convenience.

Appendix 1: Helmholtz Resonator Properties

The properties of the Helmholtz resonator are described in the literature, for example by: Kinsler et al. Fundamentals of Acoustics (3rd edition), pp. 225-228; L E Kinsler, A R Frey, A B Coppens and J V Sanders, John Wiley and Sons, New York (1984): ISBN 0-471-02933-5.
Diagrams of a theoretical Helmholtz resonator, without and with flanges, are shown in Figures 14(a) and 14(b) respectively. The resonator comprises a rigid walled cavity 1 of volume V, connected to the ambient by way of a neck 2 having length L, and cross sectional area S. The structure behaves as a resonant system when the dimensions of all of these parameters are significantly smaller than the wavelengths λ under consideration. For audio waves in the relevant range for the present invention (say 500 Hz to 6 kHz), the wavelengths lie between 680 mm and 57 mm. It is assumed that the neck constriction is sufficiently short that all fluid particles may be assumed to move in phase when actuated by a sound pressure wave.
Referring to the aforementioned publication of Kinsler et al., if λ>>L, then the mass of air (or other fluid) in the neck moves as a unit, represented by an inertance. Similarly, if λ>>V^⅓, then the volume of air V represents a compliance element (with associated stiffness), and if V>>S^½, the opening radiates sound as a simple source, corresponding to a resistance element. This acoustical system is directly analogous to a mechanical oscillator, in which its inertance, compliance and resistance correspond to the mass, compliance and resistance of, for example, a damped, sprung piston, and also to the inductance, capacitance and resistance of an L-C-R resonant electrical circuit.
When sound energy is incident on to the entrance to the tube, the mass of air in the neck moves back and forth against the compliance of the internal fluid; resonance occurs when the reactance approaches zero. The structure is characterised by its resonant frequency, ω₀ (radians/sec), [or f₀ (Hz)] and the dimensionless "quality factor" or Q, by the following two equations. $ω_{0} = C \sqrt{\frac{S}{LʹV}}$
$Q = 2 π \sqrt{V {(Lʹ / S)}^{3}}$
Here, the factor L' is used for the effective length of the neck, rather than the physical length L because of its radiation-mass loading.
According to Kinsler et al., at low frequencies a circular opening of radius a is loaded with a radiation mass equal to the fluid contained in a cylinder of area πa², and length 0.85a, if terminated in a wide flange 3 (Figure 1 (b)), and 0.6a if not flanged (Figure 1 (a)). From this, it can be shown that the relationship between L' and L is as follows. $Lʹ = L + 1.7 a (outer end flanged)$
$Lʹ = L + 1.5 a (outer end not flanged)$
If the neck opening is a circular opening of radius a then, in the limit when L approaches 0, L' approaches a value of 1.7a (flanged) or 1.5a (not flanged).
The Q factor of a Helmholtz resonator can be defined as the quotient of the resonant frequency, ω₀, and the width of the resonant peak at the half-power points (ω₁ and ω₂) lying at the -3 dB points flanking the peak at ω₀. $Q = \frac{ω_{0}}{ω_{2} - ω_{1}}$
Kinsler et al. also note, in the publication referenced above, that the resonator acts as an amplifier, in effect, with Q representing the ratio of the pressure amplitudes inside and outside the cavity, and hence gain factor, G (dB), is given by: $G = 20 \log Q$
Hence a typical Q value of 96, obtained experimentally, would represent a gain factor of 40 dB.
As has been noted in the description of the invention, by consideration of equation 2, the Q factor of the emitting cavity can be reduced either by reducing V, reducing L, or increasing S, or by any combination of or, preferably, all of these. In addition to the Q factor, another important feature of the resonant peak of a resonant system is its overall shape, and this is governed by the presence of any associated non-reactive impedances, such as resistive components, and their relative values. Their effect is to dampen the resonance and limit the range of the impedance of the system at resonance, such that it does not approach an infinitely large value or a zero value. Accordingly, it is possible to have two resonant systems featuring an identical Q value, but which possess differing spectral profiles. One might display a resonant curve in the frequency domain featuring concave-upwards flanks to the resonant peak, for example, whereas the other might exhibit flanks having a concave-downwards profile. In one aspect of the present invention, resistive elements are introduced in order to refine and match the shape of the resonant profile of a compensating cavity to that of an emitting cavity. The equations described herein can be extended to include such resistive elements and model resonant system behaviour in even more detail.

Appendix 2: Quarter-Wave Tube Resonator Properties

The acoustic properties of various closed- and open-tube (or pipe) configurations are well known to those skilled in the art. The relevant configuration here is the "single-ended, closed, quarter-wave tube", or "λ/4 tube". For frequency F₀, this is a tube that is one-quarter of a wavelength in length, being closed at one end and open at the other. When exposed to a sonic sound field at frequency F₀, some of the incident wave energy travels in to, and along, the tube, and then it and is reflected back from the closed end. After a time period corresponding to the passage of one half of a wavelength, the incident wave at the tube entrance will have undergone a 180° phase shift. Consequently, destructive interference occurs when it interacts with the original, reflected wave emerging from the quarter-wave tube; it is resonant at frequency F₀. Sound energy primarily in the region of that one specific wavelength interacts with the tube.
The amount of energy interaction between the sound field and the tube is dependent on the relative cross sectional area of the λ/4 tube. The "Q" factor is determined by the resistive losses relating to frictional interaction between the fluid medium (air) and the tube walls, and also by thermal energy loss to the walls of the tube. If the tube diameter is made very small, then the resistive losses by both mechanisms increase, and the "Q" factor decreases. Accordingly, by making very narrow diameter tubes (in the form of rectangular section grooves), the inventor has found that it is possible to manufacture resonant absorbers having characteristics that can be controlled very accurately, and which can be used advantageously for compensating the residual resonance of a minimally resonant conduit.
For example, where L is the length of the tube and C is the speed of sound, under ambient conditions, the absorption frequency, F₀, is given by: $F_{0} = C / 4 L$

and hence an absorbing quarter-wave tube for, say, a frequency of 5 kHz, would be required to have a length of 17.2 mm.

Claims

A portable electronic device, comprising:
a housing, having a width and a thickness;

first and second sonic transducers located within the housing, each sonic transducer having a respective sonic emission surface;

first and second sonic emission outlets, connected to the first and second sonic transducers respectively;

first and second conduits, respectively connecting the first sonic transducer to the first sonic emission outlet and the second sonic transducer to the second sonic emission outlet;

characterized in that:
the first and second sonic transducers are microspeakers;

the first and second sonic emission outlets comprise narrow rectangular slits, each having a width of 2mm or less in the direction of the thickness of the housing, and located in respective opposite lateral edges of the housing such that they are orthogonally oriented with respect to the sonic emission surfaces of the sonic transducers, and are separated by the width of the housing; and

the first and second conduits are flared, in that the cross-sectional areas of the conduit are increased along at least a part of their length, with no substantial decrease in cross-sectional area at any point.
A device according to claim 1, wherein substantially identical flaring is applied to each of said first and second conduits.
A device according to claim 2, wherein the flaring applied to each conduit is substantially linear.
A device according to claim 2, wherein the flaring applied to each conduit is substantially exponential or is otherwise curvilinear.
A device according to claim 2, wherein the flaring applied to each conduit is substantially smooth.
A device according to claim 2, wherein the flaring applied to each conduit contains at least one step or other discontinuity.
A device according to any of claims 2 to 6, wherein the flaring is applied over part only of the length of each conduit.
A device according to any preceding claim comprising at least one acoustic resonator means respectively linked to each conduit.
A device according to claim 8 wherein each acoustic resonator means comprises at least one Helmholtz resonator.
A device according to claim 8 or claim 9, wherein each acoustic resonator means is fabricated of, or contains or has associated therewith, acoustic damping material.
A device according to any of claims 8 to 10 wherein each acoustic resonator means comprises at least one quarter-wavelength tube, channel or groove device.
A device according to claim 11, wherein each acoustic resonator means comprises a distributed array of quarter-wavelength channels conforming substantially to a concentric elliptical array pattern.
A device according to claim 11 or claim 12, wherein each acoustic resonator means comprises at least one quarter-wavelength channel of cross-sectional area 1 mm² or less.
A portable electronic device according to any preceding claim, wherein the device comprises a cellular telephone of the claim-shell kind.