U.S. patent application number 12/250245 was filed with the patent office on 2013-04-18 for vehicle accessory microphone.
The applicant listed for this patent is G. Bruce Poe, Alan R. Watson. Invention is credited to G. Bruce Poe, Alan R. Watson.
Application Number | 20130094663 12/250245 |
Document ID | / |
Family ID | 40534223 |
Filed Date | 2013-04-18 |
United States Patent
Application |
20130094663 |
Kind Code |
A9 |
Watson; Alan R. ; et
al. |
April 18, 2013 |
VEHICLE ACCESSORY MICROPHONE
Abstract
A microphone assembly includes one or more transducers (2210)
that are positioned in one or more housings. A preprocessing
circuit (2215) includes a inverted comb filter (2245) for
eliminating predetermined frequencies between harmonics of the
human voice in a predetermined frequency range. A processing
circuit (2220) coupled to the preprocessing circuit (2215) is used
for outputting an electrical signal such that the transducers
(2210) used in combination with the processing circuit (2220) very
effectively cancels noise. The microphone assembly can be employed
in a vehicle accessory such as a vehicular mirror.
Inventors: |
Watson; Alan R.; (Buchanan,
MI) ; Poe; G. Bruce; (Hamilton, MI) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Watson; Alan R.
Poe; G. Bruce |
Buchanan
Hamilton |
MI
MI |
US
US |
|
|
Prior
Publication: |
|
Document Identifier |
Publication Date |
|
US 20090097674 A1 |
April 16, 2009 |
|
|
Family ID: |
40534223 |
Appl. No.: |
12/250245 |
Filed: |
October 13, 2008 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10492490 |
Apr 9, 2004 |
7447320 |
|
|
12250245 |
|
|
|
|
10076158 |
Feb 14, 2002 |
6882734 |
|
|
10492490 |
|
|
|
|
11539751 |
Oct 9, 2006 |
7443988 |
|
|
10076158 |
|
|
|
|
09444176 |
Nov 19, 1999 |
7120261 |
|
|
11539751 |
|
|
|
|
11551757 |
Oct 23, 2006 |
8224012 |
|
|
09444176 |
|
|
|
|
10634065 |
Aug 4, 2003 |
7130431 |
|
|
11551757 |
|
|
|
|
09724119 |
Nov 28, 2000 |
6614911 |
|
|
10634065 |
|
|
|
|
PCT/US00/31708 |
Nov 17, 2000 |
|
|
|
09724119 |
|
|
|
|
Current U.S.
Class: |
381/86 |
Current CPC
Class: |
H04R 1/083 20130101;
B60R 2011/0005 20130101; H04R 2499/13 20130101; B60R 11/0247
20130101; B60R 2011/0033 20130101; B60R 1/12 20130101; B60R
2001/1284 20130101 |
Class at
Publication: |
381/86 |
International
Class: |
H04B 1/00 20060101
H04B001/00 |
Claims
1. A vehicle accessory comprising: an accessory housing for
attaching to a vehicle; at least one transducer carried by the
accessory housing; and a microphone interface circuit electrically
coupled between the transducers and a remote processing circuit
located remote from the vehicle accessory where the microphone
interface circuit includes an inverted comb filter for eliminating
predetermined frequencies between harmonics of the human voice in a
predetermined frequency range.
2. The vehicle accessory of claim 1, wherein the accessory housing
is a mirror housing.
3. The vehicle accessory of claim 2, the microphone interface
circuit further comprising: a noise level detector for adjusting
the amplitude of the signal output by the inverted comb filter.
4. The vehicle accessory of claim 3, wherein the noise level
detector samples audio from an analog-to-digital converter that
receives audio from the at least one transducer.
5. The vehicle accessory of claim 1, wherein the inverted comb
filter utilizes a fast Fourier transform for determining a
fundamental audio frequency of audio received at the at least one
transducer.
6. The vehicle accessory of claim 5, wherein the fundamental audio
frequency is used to generate a plurality of adaptive filter
coefficients for determining the predetermined frequencies that are
to be eliminated.
7. The vehicle accessory as in claim 1, wherein audio from the
inverted comb filter is converted to analog for use by the remote
processing circuit.
8. A vehicle accessory comprising: an accessory housing for
attaching to a vehicle; a first transducer carried by the accessory
housing and generating a first audio signal; a second transducer
carried by the accessory housing and generating a second audio
signal; and a microphone interface circuit electrically coupled
between the transducers and a remote processing circuit located
remote from the vehicle accessory, the microphone interface circuit
including an inverted comb filter for eliminating noise located
between harmonics of the human voice in a predetermined frequency
range.
9. The vehicle accessory of claim 8, wherein the accessory housing
is a mirror housing.
10. The vehicle accessory of claim 9, the microphone interface
circuit further comprising: a noise level detector for adjusting
the amplitude of the signal output by the inverted comb filter.
11. The vehicle accessory of claim 10, wherein the noise level
detector samples audio from an analog-to-digital converter that
receives audio from the first transducer and second transducer.
12. The vehicle accessory of claim 8, wherein the inverted comb
filter utilizes a fast Fourier transform for determining a
fundamental audio frequency of a user's voice.
13. The vehicle accessory of claim 12, wherein the fundamental
audio frequency is used to generate a plurality of adaptive filter
coefficients for determining the predetermined frequencies that are
to be eliminated by the inverted comb filter.
14. A vehicle accessory comprising: an accessory housing for
attaching to a vehicle; a first transducer carried by the accessory
housing and generating a first audio signal; a second transducer
carried by the accessory housing and generating a second audio
signal; a high frequency boost circuit coupled to the first
transducer and second transducer for receiving the first audio
signal and second audio signal and for boosting the frequency
response at high frequencies to compensate for the effect of noise
in the vehicle; and a microphone interface circuit electrically
coupled between the first transducer and second transducer and a
remote processing circuit located remote from the vehicle accessory
where the microphone interface circuit includes an inverted comb
filter for eliminating noise located between harmonics of the human
voice in a predetermined frequency range.
15. The vehicle accessory of claim 14, wherein the accessory
housing is a vehicular mirror housing.
16. The vehicle accessory of claim 14, the microphone interface
circuit further comprising: a noise level detector for adjusting
the amplitude of a signal output by the inverted comb filter.
17. The vehicle accessory of claim 16, wherein the noise level
detector samples audio from an analog-to-digital converter that
receives audio from the first transducer and second transducer.
18. The vehicle accessory of claim 14, wherein the inverted comb
filter utilizes a fast Fourier transform for determining a
fundamental audio frequency of the user's voice in the frequency
domain.
19. The vehicle accessory of claim 18, wherein the fundamental
audio frequency is used to generate a plurality of adaptive filter
coefficients for determining the predetermined frequencies that are
to be eliminated by the inverted comb filter.
20. The vehicle accessory of claim 14, wherein the high frequency
boost circuit boosts the frequency response such that a frequency
response curve associated with the first transducer and second
transducer is not flat and independent of the effect of the
vehicle.
21. The vehicle accessory of claim 14, wherein the high frequency
boost circuit boosts the frequency response between about 3.5 and 5
KHz.
22. A vehicle accessory comprising: an accessory housing for
attaching to a vehicle; a plurality of transducers carried by the
accessory housing and generating a plurality of respective audio
signals; a microphone interface circuit electrically coupled
between the plurality of transducers and a remote processing
circuit located remote from the vehicle accessory, the microphone
interface circuit including at least one inverted comb filter for
eliminating predetermined frequencies between harmonics of the
human voice in a predetermined frequency range; and a noise level
detector for adjusting the amplitude of a signal output of the
inverted comb filter.
23. The vehicle accessory of claim 22, wherein the accessory
housing is a rearview mirror assembly housing.
24. The vehicle accessory of claim 22, wherein the at least one
inverted comb filter utilizes a fast Fourier transform for
determining a fundamental audio frequency of a user's voice in the
frequency domain.
25. The vehicle accessory of claim 24, wherein the fundamental
audio frequency is used to generate a plurality of adaptive filter
coefficients for determining the predetermined frequencies that are
to be eliminated by the inverted comb filter.
26. The vehicle accessory of claim 22, wherein the microphone
interface circuit further comprises: an digital-to-analog converter
for converting audio adjusted in amplitude from the at least one
inverted comb filter into an analog signal for use by the remote
processing circuit.
Description
[0001] This application claims priority to U.S. Pat. No. 6,882,734;
U.S. Pat. No. 7,120,261; U.S. Pat. No. 6,614,911; and U.S.
application Ser. No. 10/492,490.
BACKGROUND OF THE INVENTION
[0002] The present invention pertains to microphones, and more
particularly to a microphone associated with a vehicle accessory
such as a rearview mirror assembly or the housing of a rear vision
display device.
[0003] It has long been desired to provide improved microphone
performance in devices such as communication devices and voice
recognition devices that operate under a variety of different
ambient noise conditions. Communication devices supporting
hands-free operation permit the user to communicate through a
microphone of a device that is not held by the user. Because of the
distance between the user and the microphone, these microphones
often detect undesirable noise in addition to the user's speech.
The noise is difficult to attenuate. Hands-free communication
systems for vehicles are particularly challenging due to the
dynamically varying ambient noise that is present. For example,
bi-directional communication systems such as two-way radios,
cellular telephones, satellite telephones, and the like, are used
in vehicles, such as automobiles, trains, airplanes and boats. For
a variety of reasons, it is preferable for the communication
devices of these systems to operate hands-free, such that the user
need not hold the device while talking, even in the presence of
high ambient noise levels subject to wide dynamic fluctuations.
[0004] Bi-directional communication systems include an audio
speaker and a microphone. In order to improve hands-free
performance in a vehicle communication system, a microphone is
typically mounted near the driver's head. For example, a microphone
is commonly attached to the vehicle visor or headliner using a
fastener such as a clip, adhesive, hook and loop fastening tape
(such as VELCRO brand fastener), or the like. The audio speaker
associated with the communication system is preferably positioned
remote from the microphone to assist in minimizing feedback from
the audio speaker to the microphone. It is common, for example, for
the audio speaker to be located in a vehicle adaptor, such as a
hang-up cup or a cigarette lighter plug used to provide energizing
power from the vehicle electrical system to the communication
device. Thus, although the communication system designer knows the
position of the audio speaker in advance, the position of the
microphone is unknown as the user can position the microphone where
they choose. The position of the microphone relative to the person
speaking will determine the level of the speech signal output by
the microphone and may affect the signal-to-noise ratio. The
position of the microphone relative to the audio speaker will
impact on feedback between the speaker and microphone. Accordingly,
the performance of the audio system is subject to the user's
installation of the microphone. Additionally, the microphone will
typically include a wire, which if it is mounted to the surface of
the vehicle interior, will not be aesthetically pleasing.
Alternatively, if the wire is to be mounted behind the interior
lining, the vehicle interior must be disassembled and then
reattached so that the wire can be hidden, which may result in
parts that rattle loudly or hang loosely from the vehicle
frame.
[0005] One potential solution to avoid these difficulties is
disclosed in U.S. Pat. No. 4,930,742, entitled "REARVIEW MIRROR AND
ACCESSORY MOUNT FOR VEHICLES", issued to Schofield et al. on Jun.
5, 1990, which uses a microphone in a mirror mounting support.
Although locating the microphone in the mirror support provides the
system designer with a microphone location that is known in
advance, and avoids the problems associated with mounting the
microphone after the vehicle is manufactured, there are a number of
disadvantages to such an arrangement. Because the mirror is
positioned between the microphone and the person speaking into the
microphone, a direct unobstructed path from the user to the
microphone is precluded. Additionally, the location of the
microphone on the windshield detrimentally impacts on microphone
design flexibility and overall noise performance of the
microphone.
[0006] U.S. Pat. Nos. 5,940,503, 6,026,162, 5,566,224, 5,878,353,
and D 402,905 disclose rearview mirror assemblies with a microphone
mounted in the bezel of the mirror. None of these patents, however,
disclose the use of acoustic ports facing multiple directions nor
do they disclose microphone assemblies utilizing more than one
microphone transducer. The disclosed microphone assemblies do not
incorporate sufficient noise suppression components to provide
output signals with relatively high signal-to-noise ratios, and do
not provide a microphone having a directional sensitivity pattern
or a main lobe directed forward of the housing and attenuating
signals originating from the sides of the housing.
[0007] It is highly desirable to provide voice recognition systems
in association with vehicle communication systems, and most
preferably, such a system would enable hands-free operation.
Hands-free operation of a device used in a voice recognition system
is a particularly challenging application for microphones, as the
accuracy of a voice recognition system is dependent upon the
quality of the electrical signal representing the user's speech.
Conventional hands-free microphones are not able to provide the
consistency and predictability of microphone performance needed for
such an application in a controlled environment such as an office,
let alone in an uncontrolled and noisy environment such as an
automobile.
[0008] Commonly-assigned U.S. Pat. No. 6,882,734 and PCT
Application Publication No. WO 01/37519 A2 also disclose various
embodiments of rearview mirror-mounted microphone assemblies. In
those embodiments, at least one microphone transducer is typically
aimed at the driver of the vehicle. This usually results in the
microphone assembly being visibly mounted to the top or bottom
surface of the mirror housing. Such visibility raises certain
styling concerns as well as performance issues when used in certain
environments and in combination with digital signal processing
circuits.
[0009] It has long been desired to provide improved microphone
performance in devices such as communication devices and voice
recognition devices that operate under a variety of different
ambient noise conditions. Communication devices supporting
hands-free operation permit the user to communicate through a
microphone of a device that is not held by the user. Because of the
distance between the user and the microphone, these microphones
often detect undesirable noise in addition to the user's speech.
The noise is difficult to attenuate. Hands-free communication
systems for vehicles are particularly challenging due to the
dynamically varying ambient noise that is present. For example,
bi-directional communication systems such as two-way radios,
cellular telephones, satellite telephones, and the like, are used
in vehicles, such as automobiles, trains, airplanes and boats. For
a variety of reasons, it is preferable for the communication
devices of these systems to operate hands-free, such that the user
need not hold the device while talking, even in the presence of
high ambient noise levels subject to wide dynamic fluctuations.
Accordingly, there is a need for a microphone for a vehicle
providing improved hands-free performance and preferably enabling
voice recognition operation when a digital signal processing
circuit is utilized.
BRIEF DESCRIPTION OF THE DRAWINGS
[0010] The subject matter that is regarded as the invention is
particularly pointed out and distinctly claimed in the claim
portion that concludes the specification. The invention, together
with further objects and advantages thereof, may best be understood
by reference to the following description taken in conjunction with
the accompanying drawings, where like numerals represent like
components, and in which:
[0011] FIG. 1 is a top plan view illustrating a vehicle with a
portion of the roof cut away;
[0012] FIG. 2 is a front, bottom and left side perspective view
illustrating a rearview mirror assembly and fragmentary mirror
support used in the vehicle of FIG. 1;
[0013] FIG. 3 is a top exploded view illustrating a microphone
assembly used in the mirror according to FIG. 2;
[0014] FIG. 4 is a bottom plan view illustrating the microphone
assembly according to FIG. 2;
[0015] FIG. 5 is a bottom plan view illustrating a transducer mount
in the microphone assembly according to FIG. 3;
[0016] FIG. 6 is cross-sectional view taken along plane 6-6 in FIG.
4 illustrating the microphone assembly according to FIG. 3;
[0017] FIG. 7 is a top plan view illustrating the microphone
assembly according to FIG. 5 with the circuit board removed to view
show the transducers in transducer mount;
[0018] FIG. 8 is a circuit schematic partially in block diagram
form illustrating a circuit employed with the microphone assembly
of FIGS. 3-7;
[0019] FIG. 9 is a top plan view schematic representation
illustrating the sound channel for the transducers of the
microphone assembly according to FIGS. 1-7;
[0020] FIG. 10 is a top plan view schematic representation
illustrating the sound channel for an alternate transducer
arrangement for the microphone assembly;
[0021] FIG. 11 is a top plan view schematic representation
illustrating the sound channel for another alternate transducer
arrangement for the microphone assembly;
[0022] FIG. 12 is a circuit schematic partially in block diagram
form illustrating a circuit for use with the microphone according
to claim 11;
[0023] FIG. 13 is a circuit schematic partially in block diagram
form illustrating an auto-calibration circuit for use with the
microphone assembly;
[0024] FIG. 14 is a flow chart representing operation of the
controller of FIG. 12;
[0025] FIG. 15 is a cross-sectional view of the microphone
according to FIG. 10 taken along the longitudinal axis of the
microphone;
[0026] FIG. 16 is a perspective view of a microphone assembly
constructed in accordance with another embodiment of the present
invention;
[0027] FIG. 17 is an exploded perspective view of a microphone
assembly shown in FIG. 16;
[0028] FIG. 18 is a front isometric view of an embodiment of a
rearview mirror assembly constructed in accordance with another
embodiment of the present invention;
[0029] FIG. 19 is a rear isometric view of an embodiment of a
rearview mirror assembly shown in FIG. 18;
[0030] FIG. 20 is a side elevation of the rearview mirror assembly
shown in FIGS. 18 and 19;
[0031] FIG. 21 is an exploded perspective view of a microphone
assembly constructed in accordance with another embodiment of the
present invention;
[0032] FIGS. 22A-22D are plots of polar patterns at different
frequencies as obtained from a microphone assembly constructed in
accordance with the present invention with a cover over the
transducers;
[0033] FIGS. 23A-23D are plots of polar patterns at different
frequencies as obtained from a microphone assembly constructed in
accordance with the present invention without a cover over the
transducers;
[0034] FIG. 24 is a side elevational view of a portion of a
rearview mirror assembly having a deflector, a fine turbulence
generator and a microphone assembly according to another embodiment
of the present invention;
[0035] FIG. 25 is a top view of the portion of the rearview mirror
assembly having the deflector, the fine turbulence generator and
the microphone assembly that are shown in FIG. 24;
[0036] FIG. 26 is a rear view of the portion of the rearview mirror
assembly having the deflector, the fine turbulence generator and
the microphone assembly that are shown in FIGS. 24 and 25;
[0037] FIG. 27 is an electrical circuit diagram in block form
showing an embodiment of a microphone processing circuit of the
present invention;
[0038] FIG. 28A is an electrical circuit diagram in schematic form
showing an exemplary high pass filter that may be used in the
circuit shown in FIG. 27;
[0039] FIG. 28B is an electrical circuit diagram in schematic form
showing an exemplary all-pass phase shifter that may be used in the
circuit shown in FIG. 27;
[0040] FIG. 28C is an electrical circuit-diagram in schematic form
showing an exemplary summing circuit that may be used in the
circuit shown in FIG. 27;
[0041] FIG. 28D is an electrical circuit diagram in schematic form
showing an exemplary three-pole high pass filter that may be used
in the circuit shown in FIG. 27;
[0042] FIG. 28E is an electrical circuit diagram in schematic form
showing an exemplary buffer circuit that may be used in the circuit
shown in FIG. 27;
[0043] FIG. 29A is a plot of three frequency response curves of a
second order microphone assembly with sound originating from three
different directions;
[0044] FIG. 29B is a plot of a frequency response curve of the
second order microphone processing circuit shown in FIG. 27 but
without the all-pass phase shifter;
[0045] FIG. 29C is a plot of four frequency response curves of the
second order microphone processing circuit shown in FIG. 27 with
sound originating from four different directions;
[0046] FIG. 30 is block diagram illustrating a microphone system
constructed in accordance with the present invention;
[0047] FIG. 31 is a process diagram for the digital signal
processor shown in FIG. 30 according to a first embodiment;
[0048] FIG. 32 is an exemplary plot of a FFT of an audio signal
received from a typical transducer while receiving both noise and a
user's speech;
[0049] FIG. 33 is a graph of an ideal inverted comb filter for
filtering the audio signal whose FFT is illustrated in FIG. 32;
[0050] FIG. 34 is a process diagram for the digital signal
processor shown in FIG. 30 according to a second embodiment;
[0051] FIG. 35 is a simplified electrical schematic of a prior art
microphone assembly coupled to an electronic assembly;
[0052] FIG. 36 is a simplified electrical schematic of a microphone
assembly coupled to an electronic assembly through a microphone
interface, according to an embodiment of the present invention;
[0053] FIG. 37 is a simplified electrical schematic of a microphone
assembly coupled to an electronic assembly through a microphone
interface, according to another embodiment of the present
invention;
[0054] FIG. 38 is a simplified electrical schematic of a microphone
assembly coupled to an electronic assembly through a microphone
interface, according to yet another embodiment of the present
invention;
[0055] FIG. 39A is an elevational view of the front of a rearview
mirror assembly constructed in accordance with an alternative
embodiment of the present invention;
[0056] FIG. 39B is an elevational view of the rear of the rearview
mirror assembly shown in FIG. 39A;
[0057] FIG. 39C is an elevational view of the side of the rearview
mirror assembly shown in FIGS. 39A and 39B;
[0058] FIG. 39D is a plan view of the top of the rearview mirror
assembly shown in FIGS. 39A-39C;
[0059] FIG. 40 is a graph showing four plots representing the
output of various microphone assemblies when the vehicle defroster
is running at full speed;
[0060] FIGS. 41A-41D are perspective views showing a rearview
mirror incorporating a microphone assembly in accordance with
another embodiment of the present invention;
[0061] FIG. 42 is a plan view of the top of another embodiment of
the microphone assembly of the present invention, shown with the
housing in outline form;
[0062] FIGS. 43A and 43B are perspective views of a housing that
may be used for a microphone assembly constructed in accordance
with the embodiment of the present invention shown in FIG. 42;
[0063] FIG. 44 are plots of polar patterns for both of the two
transducers as obtained from a microphone assembly constructed in
accordance with the embodiment of the present invention shown in
FIGS. 42, 43A and 43B;
[0064] FIGS. 45A and 45B are perspective views of an alternative
housing that may be used for a microphone assembly constructed in
accordance with the embodiment of the present invention shown in
FIG. 42;
[0065] FIG. 46 are plots of polar patterns for both of the two
transducers as obtained from a microphone assembly constructed in
accordance with the embodiment of the present invention shown in
FIGS. 42, 45A and 45B;
[0066] FIG. 47 is a plan view of the top of another embodiment of
the microphone assembly of the present invention, shown with the
housing in outline form;
[0067] FIGS. 48A and 48B are perspective views of a housing that
may be used for a microphone assembly constructed in accordance
with the embodiments of the present invention shown in FIGS. 47 and
49;
[0068] FIG. 49 is a plan view of the top of another embodiment of
the microphone assembly of the present invention, shown with the
housing in outline form;
[0069] FIGS. 50A-50E show a rearview mirror assembly incorporating
a microphone assembly in accordance with another embodiment of the
present invention;
[0070] FIG. 51 is an electrical circuit diagram in block form
showing a noise cancellation circuit that may be used with the
present invention; and
[0071] FIG. 52 is an electrical circuit diagram in block form
showing an alternative noise cancellation circuit that may be used
with the present invention.
[0072] FIG. 53A is an elevational view of the front of a rearview
mirror assembly incorporating a microphone assembly in accordance
with a first embodiment of the present invention;
[0073] FIG. 53B is an elevational view of the rear of the rearview
mirror assembly incorporating a microphone assembly in accordance
with the first embodiment of the present invention;
[0074] FIG. 53C is an elevational view of one side of the rearview
mirror assembly incorporating a microphone assembly in accordance
with the first embodiment of the present invention;
[0075] FIG. 53D is a plan view of the top of the rearview mirror
assembly incorporating a microphone assembly in accordance with the
first embodiment of the present invention;
[0076] FIG. 53E is an elevational view of one side of the rearview
mirror assembly in partial cross-section taken along line E-E in
FIG. 53B;
[0077] FIG. 54A is an elevational view of the rear of the rearview
mirror assembly incorporating a microphone assembly in accordance
with the second embodiment of the present invention;
[0078] FIG. 54B is an elevational view of one side of the rearview
mirror assembly incorporating a microphone assembly in accordance
with the second embodiment of the present invention;
[0079] FIG. 54C is a plan view of the top of the rearview mirror
assembly incorporating a microphone assembly in accordance with the
second embodiment of the present invention;
[0080] FIG. 54D is an elevational view of one side of the rearview
mirror assembly in partial cross-section taken along line D-D in
FIG. 54C;
[0081] FIG. 55 is an elevational view of the front of the
microphone assembly of the second embodiment of the present
invention;
[0082] FIG. 56 is a plan view of the top of the microphone assembly
of the second embodiment of the present invention;
[0083] FIG. 57 is an elevational view of one end of the microphone
assembly of the second embodiment of the present invention;
[0084] FIG. 58 is a cross-sectional view of the microphone assembly
shown in FIG. 55 taken along line VII-VII;
[0085] FIG. 59 is a cross-sectional view of the microphone assembly
shown in FIG. 55 taken along line VIII-VIII;
[0086] FIG. 60 is a cross-sectional view of the microphone assembly
shown in FIG. 55 taken along line IX-IX;
[0087] FIG. 61 is a schematic top view of the microphone assembly
of the second embodiment of the present invention;
[0088] FIG. 62 is an exploded perspective view of a portion of the
microphone assembly of the second embodiment of the present
invention;
[0089] FIG. 63 is a polar plot taken of a rearview mirror assembly
of the present invention having a microphone assembly with a low
acoustic resistance windscreen;
[0090] FIG. 64 is a comparative noise plot taken of two different
rearview mirror assemblies of the present invention, one having a
microphone assembly with a low acoustic resistance windscreen and
the other one having a microphone assembly with a very high
acoustic resistance windscreen;
[0091] FIG. 65 is a polar plot taken of a rearview mirror assembly
of the present invention having a microphone assembly with a very
high acoustic resistance windscreen, but with no acoustic dam;
[0092] FIG. 66 is a polar plot taken of a rearview mirror assembly
of the present invention having a microphone assembly with a very
high acoustic resistance windscreen and an acoustic dam;
[0093] FIG. 67 is a polar plot taken of a rearview mirror assembly
of the present invention in a first orientation at 250 Hz and
having a microphone assembly with a very high acoustic resistance
windscreen and an acoustic dam;
[0094] FIG. 68 shows various polar plots taken of a rearview mirror
assembly of the present invention in a first orientation at various
frequencies between 300 Hz and 2 kHz and having a microphone
assembly with a very high acoustic resistance windscreen and an
acoustic dam;
[0095] FIG. 69 shows various polar plots taken of a rearview mirror
assembly of the present invention in a first orientation at various
frequencies between 3 Hz and 6 kHz and having a microphone assembly
with a very high acoustic resistance windscreen and an acoustic
dam;
[0096] FIG. 70 shows various polar plots taken of a rearview mirror
assembly of the present invention in a second orientation at
various frequencies between 300 Hz and 1 kHz and having a
microphone assembly with a very high acoustic resistance windscreen
and an acoustic dam;
[0097] FIG. 71 shows various polar plots taken of a rearview mirror
assembly of the present invention in a second orientation at
various frequencies between 3 kHz and 6 kHz and having a microphone
assembly with a very high acoustic resistance windscreen and an
acoustic dam;
[0098] FIG. 72 shows various polar plots taken of a rearview mirror
assembly of the present invention in a second orientation at
various frequencies between 6.5 Hz and 8 kHz and having a
microphone assembly with a very high acoustic resistance windscreen
and an acoustic dam; and
[0099] FIG. 73 is a schematic view of a microphone assembly of the
present invention; and
[0100] FIG. 74 is a perspective view of the microphone assembly
shown in FIG. 73.
DETAILED DESCRIPTION OF THE INVENTION
[0101] The microphone assemblies of the present invention are
associated with an interior rearview mirror and have superior
performance even in the presence of noise. The microphone
assemblies enhance the performance of hands-free devices with which
they are associated, including highly sensitive applications such
as voice recognition for a telecommunication system, by improving
the signal-to-noise ratio of the microphone assembly output. The
microphone assemblies eliminate mechanically induced noise and
provide the designer with significant freedom with respect to
selection of the microphone assembly's sensitivity, frequency
response and polar pattern. Additionally, circuitry can be provided
for the transducer to generate an audio signal from the transducer
output that has a high signal-to-noise ratio.
[0102] As seen in FIG. 1, a vehicle 100 includes an interior
rearview mirror assembly 101 by which the vehicle operator 103
(illustrated in phantom) can view a portion of the road behind the
vehicle 100 without having to turn around. The rearview mirror
assembly 101 is mounted to the vehicle windshield 105, or the
vehicle's headliner, via a mirror mounting support 104, in a
conventional manner that facilitates electrical connection of the
rearview mirror to the vehicle's electrical system and permits
driver adjustment of the mirror-viewing angle.
[0103] The rearview mirror assembly 101 according to a first
embodiment is enlarged in FIG. 2. The mirror assembly 101 includes
an elongated housing 206 pivotably carried on mirror support 104.
The mirror 202 may be any conventional interior rearview mirror,
such as a prismatic mirror of the type used with a mirror housing
manually adjustable for daytime and nighttime operation, or a
multiple element mirror effecting automatic reflectivity
adjustment, such as an electrooptic or electrochromic mirror. The
elongated housing 206 may be of any conventional manufacture such
as integrally molded plastic.
[0104] The rearview mirror assembly 101 further includes a
microphone assembly 208 that is preferably mounted to the housing
206 at a location visible to the vehicle driver 103 or at a
position which is direct line of sight between the speaker's mouth
and the microphone. It is advantageous for the microphone assembly
208 to be positioned on the mirror housing 206 as the mirror
assembly is movably carried on the support 104. The driver 103
(FIG. 1) will typically adjust the position of the mirror 202 and
housing 206 to reflect images visible through the rear window 109
of the vehicle 100. When making such an adjustment for viewing
angle, the driver 103 adjusts the mirror 202 toward their eyes by
moving housing 206, which will simultaneously direct the front of
microphone assembly 208 toward the driver. However, the microphone
assembly could be mounted in other vehicle accessories, such as a
visor, an overhead console, a vehicle trim component such as a
headliner or an A-pillar cover, a center console, an on-windshield
console, or the like.
[0105] A first embodiment of the microphone assembly 208 will now
be described in greater detail with respect to FIGS. 3-7. The
microphone assembly includes a microphone housing 300, a transducer
mount 302, a first transducer 304, a second transducer 306, and a
circuit board 308. The microphone housing 300 (FIGS. 3 and 4) is
generally cylindrical, having a round foot print and a low profile,
although the housing could have a generally square foot print, an
elongated elliptical or rectangular foot print, or any other shape
desired by the microphone designer. The microphone housing 300
includes front ports 312 that face the driver 103 and rear ports
314 that face away from the driver 103. The ports 312 and 314
provide a sound passage through the microphone housing. The ports
312, 314 can have any suitable opening shape or size. The housing
also includes posts 316, 317 used to hold the microphone assembly
208 together, as described in greater detail herein below. A rail
318 on the inside surface of housing 300 is shaped to receive a
portion of mount 302. When received in the rail, mount 302 is
positioned with the transducer 304 and 306 sound channels properly
aligned with the ports 312, 314. The housing also includes mounting
tabs 320 for insertion into openings (not shown) in the lower
surface of housing 206. For example, the tabs can be generally
L-shaped in profile for insertion into the housing 300. After tabs
320 are inserted into housing 206, the microphone housing 300 is
locked to the mirror housing 206 by rotating the microphone to a
locked position, thereby securing the microphone assembly 208 on
the housing assembly 101. Alternately, the tabs 320 can be elongate
snap connectors that slide into an opening (not shown) in the
bottom surface of the mirror housing and snap into engagement with
the inside surface of the mirror housing 206 after fall insertion.
The microphone housing 300 can be integrally molded plastic,
stamped metal, or of any other suitable manufacture.
[0106] The transducer mount 302 is configured such that it is
pressed into the housing 300 and is slightly compressed between
circuit board 308 and housing 300. The transducer mount provides
acoustic seals for the transducers 304 and 306, and with the
circuit board 308 and housing 300, defines acoustic channels, or
sound passages, to the front and rear faces of the transducers 304,
306, as described in greater detail below. The mount 302 includes
webs 324 between walls 332 and webs 325 between walls 333 that
extend outwardly from the core of mount 302 to provide sound
passages, and also help to position mount 302 in the housing 300.
Projections 326, 327 are located on opposite ends of mount 302 to
help position mount 302 in housing 300. Openings 328, 329 are
provided in the webbing 324, 325 of mount 302 for passage of posts
316, 317. Cylindrical wells 330, 331 are provided in the core of
transducer mount 302 for receipt of transducers 304, 306,
respectively. Each of the wells 330, 331 includes a terminating
wall 501 (FIG. 5) against which the front faces 500 of the
transducers 304, 306 sit. The terminating walls 501 each include a
channel 506, 508 that extends radially outward from the center of
the well, which is the location of the front transducer aperture.
The mount 302 can be of any suitable manufacture, such as a molded
elastomer. In particular the mount 302 is resilient and
non-conductive, and provides acoustic isolation. For example, the
transducer mount 302 can be manufactured of urethane commercially
available from Mobay, Inc.
[0107] The transducers 304 and 306 are preferably substantially
identical. The transducers include a front aperture 502 which
passes sound to the front surface of a transducer diaphragm and
openings 337 (FIG. 3) in the back face that port sound to the back
surface of the transducer diaphragm. The transducers include
electrical leads 336 on the back face thereof for electrical
connection to the conductive layer of circuit board 208. The
transducers 304 and 306 can be any suitable, conventional
transducers, such as electret, piezoelectric, or condenser
transducers. The transducers may be, for example, electret
transducers such as those commercially available from Matsushita of
America (doing business as Panasonic), and may advantageously be
unidirectional transducers. If electret transducers are employed,
the transducers can be suitably conditioned to better maintain
transducer performance over the life of the microphone assembly
208. For example, the diaphragms of the transducers 304, 306 can be
baked prior to assembly into the transducers.
[0108] The circuit board 308 has a conductive layer, on surface
334, etched and electrically connected to the transducer leads 336
of transducers 304, 306. The microphone leads 340 are connected to
the transducer leads 336 by a circuit 800 (FIG. 8) mounted to the
conductive layer of circuit board 308. Although circuit 800 can be
mounted on the circuit board 308 in the microphone housing, it will
be recognized that the circuit 800 can alternatively be mounted on
a printed circuit board in the mirror housing 206, and further that
in the case of an electrooptic mirror, such as an electrochromic
mirror, the circuit 800 can be mounted on a common circuit board
with the mirror electrical components, or the circuit 800 and the
mirror electrical components can be mounted on separate circuit
boards within the housing 206. The electrical connection of the
microphone leads 340, the transducer leads 336, and the components
of circuit 800, are preferably by electrical traces in the
conductive layer of the circuit board, formed by conventional means
such as etching, and vias extending through the dielectric
substrate of the printed circuit board. The circuit board includes
holes 350 and 352 for receipt of posts 316 and 317 on microphone
housing 300. The posts 316, 317 are heat staked to the circuit
board substrate after the posts are inserted through holes 350 and
352 to secure the connection of the circuit board to the housing
300 and insure that the microphone assembly provides acoustically
isolated sound channels between the transducers 304, 306 and the
ports 312, 314, as described in greater detail herein below.
[0109] To assemble the microphone assembly 208, the transducers 306
and 308 are mounted on the circuit board 308 by conventional means,
such as by soldering transducer leads 336 to the conductive layer
334 of circuit board 308. It is envisioned that the transducer
leads can alternatively be elongated posts that extend through vias
in the printed circuit board, that the surface 360 can be a
conductive layer, and that the components of circuit 800 can be
located on surface 360 of the printed circuit board, connected
between the transducer leads 336 and the microphone leads 340.
Regardless of how the transducers 304 and 306 are mounted on the
circuit board 308, the circuit board mounted transducers are
pressed into the cylindrical wells 330, 331 in the mount 302. When
fully inserted in the wells, the front faces 500 (FIG. 5) of the
transducers 304, 306, are positioned against the terminating wall
501 of the wells 330, 331. The wall 501 of each of the wells 330,
331 includes a channel 506, 508 aligned with the openings 502 in
the front face of the transducers 304, 306.
[0110] The partial assembly comprising mount 302, transducers 304,
306 and circuit board 308, is pressed into the housing 300. FIG. 7
illustrates the microphone assembly 208 with the printed circuit
board 308 removed. The back surfaces of the transducers 304, 306,
having multiple openings 337 and transducer leads 336, are visible
from the open end of the cylindrical wells 330,331. When the
transducers 304, 306 are fully inserted in the well, such that the
front face 500 of the transducers are juxtaposed with the wall 501
terminating the well, a chamber is formed between the back surface
of each of the transducers 304, 306 and the circuit board 308, as
best shown in FIG. 6. A wall of the mount circumscribes the
periphery of the transducer 306, 307, and a short channel 371, 373
extends from the well 330, 331 to the aperture 370, 372. The
circumscribing wall provides an acoustic seal with the circuit
board 308. Apertures 370, 372 connect the chamber, between each of
the transducers 304, 306 and the circuit board 308, with the
channels 510, 512, respectively. The chamber behind each of the
transducers provides a sound passage from the back openings 337 of
the transducers through channels 371, 373, 510, and 512 and ports
312, 314. When the mount 302 is fully inserted in the housing 300,
the sound passages extending from the front face of each of the
transducers to ports 312 and 314 are defined by the housing 300 and
the mount 302. The sound passages extending from the back face of
each of the transducers to ports 312 and 314 are defined by the
housing 300, mount 302 and circuit board 308.
[0111] In particular the front opening 502 of transducer 306 is
connected to the front ports 312 of the microphone housing 300 via
the sound passage 506 as best shown in FIG. 6. The rear face
openings 337 of the transducer 306 is acoustically coupled to the
rear ports 314 via sound channel 373, aperture 372 and channel 510.
Transducer 304 is coupled to the front ports 312 and the rear ports
314 in the same manner, but in the opposite phase. In particular,
the front face of transducer 304 is acoustically coupled to the
rear ports 314 via acoustic channel 508 (FIG. 5). The rear face
openings 337 of the transducer 304 are acoustically coupled to the
front ports 312 via channel 371, aperture 370, and channel 512.
Signals originating from the front of the microphone assembly,
which is the surface of the microphone assembly facing the driver,
enter the front of transducer 306 and the back of transducer 304,
whereas sound originating from the rear of the microphone assembly
enter the front face of transducer 304 and the back face of
transducer 306. Omni-directional sounds will be detected equally by
the transducers, at opposite phases.
[0112] As illustrated in FIG. 6, the center axes C of the
transducers 304, 306 are oriented at an angle of 90 degrees with
respect to the longitudinal axes L.sub.B and L.sub.F of the
channels 506, 508, 510, 512. Thus, the acoustic outputs from the
two transducers lie on a common axis facing in opposite directions
and perpendicular to the center axis C of the transducers.
[0113] The transducers 304 and 306 are electrically coupled to an
operational amplifier 802 (FIG. 8) of circuit 800. In particular,
transducer 306 is coupled to the inverting input of the operational
amplifier 802 and transducer 304 is coupled to the non-inverting
input of the operational amplifier. Resistor R8, connected between
the transducer 306 and the inverting input of the operational
amplifier 802, is preferably a potentiometer to permit manual
balancing of the transducers. Alternatively, the resistor R12
connected between transducer 304 and the non-inverting input of the
operational amplifier, or both resistors R10 and R12, can be
implemented by potentiometers. It is also envisioned that a
variable gain amplifier with an associated manually adjustable
potentiometer can be inserted in one or both of the paths between
transducers 304, 306 and operational amplifier 802. The operational
amplifier may be implemented using any suitable operational
amplifier, such as the TLC271 operational amplifier available from
Texas Instruments, Inc. The manually adjustable potentiometer R8 is
provided for varying the gain of the transducer path to permit
adjustment of the signal level from transducer 306 such that both
transducer 304, 306 paths produce the same signal gain (i.e., the
signal gain through both transducers is equal). By providing
identical gain through both transducers, omni-directional noise
detected by both transducers will be completely cancelled at the
output of the operational amplifier 802.
[0114] Acoustic signals generated by the vehicle driver, such as
the driver's speech, will be input to the front of transducer 306
and the back of transducer 304, such that the speech will be
present in the audio signal at the output of operational amplifier
302. Sound from the sides of the microphone assembly will be
cancelled by the transducers 304, 306 and the operational amplifier
802. The most intense noise in a vehicle tends to originate from
the sides and/or front of the vehicle. The microphone assembly 208
mounted on the rearview mirror 206, including amplifier 802, will
significantly reduce noise as the bi-directional microphone
assembly is not responsive to noise originating from the sides of
the vehicle when mounted in the mirror assembly 101 which is
generally aligned with the longitudinal axis of the vehicle.
Furthermore, mechanical noise, such as that originating in the
rearview mirror assembly 101, will be detected by both transducers
304, 306 equally, and thus will be cancelled out by the operational
amplifier 802.
[0115] The output of the operational amplifier 802 is input to a
3-pole high pass filter and unity gain follower 804, having a
cut-off at approximately 100-300 Hz, and preferably at 150 Hz. The
filter removes noise below the voice frequency. Terminals 340 are
coupled to the vehicle's electrical circuitry, which may for
example include voice recognition circuitry, a cellular
transceiver, a two-way radio, or any other control circuitry. The
transistors Q1 and Q2 can be implemented using any suitable
commercially available transistor elements, such as FFB2227
commercially available from Fairchild Semiconductor.
[0116] In summary, the bi-directional microphone assembly 208 is
very responsive to voice signals from the driver 103 located in
front of the mirror assembly 101, as signals from the front of the
mirror will sum in operational amplifier 802. As a consequence,
on-axis sound will experience a gain and the microphone assembly
will have a high signal-to-noise ratio. It is envisioned that a
gain of approximately 6 dB can be achieved by bi-directional
microphone assembly 208. The microphone is highly directional, such
that off-axis sound is attenuated, and even nulled, by the
microphone. Further, the bi-directional microphone assembly 208 can
employ any type of directional transducer, so long as identical
transducers are employed.
[0117] The bi-directional microphone assembly 208 is schematically
illustrated in FIG. 9, and alternate embodiments are schematically
illustrated in FIGS. 10 and 11. As described above, the
bi-directional microphone assembly 208 includes transducer 306,
having its front face opening ported to the front ports 312 through
channel 506 and its back face openings ported to the back ports 314
through channels 370, 371 and 510, and transducer 304, having its
front face ported to the rear ports 314 through channel 508 and its
rear face ported to the front port 312 through channels 372, 373
and 512. The bi-directional microphone assembly 208 thus has
transducers mounted on the same lateral axis, but at opposite
phases. An alternative to the bi-directional microphone assembly
208, is the hyper cardioid microphone assembly 1000 illustrated in
FIG. 10. The hyper cardioid microphone assembly 1000 includes a
front transducer 1002 having its front face acoustically coupled to
port 1004 through channel 1005 and its back face acoustically
coupled to port 1006 through channel 1009. The front face of a rear
transducer 1008 is acoustically coupled to ports 1010 through
channel 1011 and the rear face of transducer 1008 is acoustically
coupled to port 1006 through channel 1012. The transducers are
electrically coupled to an operational amplifier in the same manner
that the transducers 304 and 306 are electrically coupled to
operational amplifier 802. However, unlike bi-directional
microphone assembly 208, for which identical transducers are
selected, the transducers 1002 and 1008, and the variable gain
balance circuit 802, are selected and operated such that the front
transducer 1002 produces a greater sensitivity than the back
transducer 1008 while maintaining a null of the vibration created
signals.
[0118] The microphone assembly 1000 may be advantageous in
applications wherein the noise incident on the microphone assembly
is generally random and omni directional, or in an environment
where the front lobe of the microphone needs to be larger to
accommodate off-axis noise sources. Microphone assembly 1000 will
be better suited for use in vehicles where the person speaking,
such as the driver, is not positioned in front of the rearview
mirror assembly, because the bi-directional microphone 208 may
attenuate the speech from the person speaking. As noted above, the
most intense noise in a vehicle originates from the side of the
vehicle, which the bi-directional microphone assembly 208 mounted
to the mirror assembly 101 will better reject than the hyper
cardioid microphone assembly 1000. Another problematic
environmental condition better resolved by the bi-directional
microphone assembly 208 than the hyper cardioid microphone assembly
1000, is small room reverberation effect. Reverberation causes
noise, with a wavelength long relative to room dimensions, such
that it is omni-directional. Microphone assembly 208, having two
identical transducers, will effectively null omni-directional
components, such that all the reverberating noise will be
cancelled. The hyper cardioid microphone assembly 1000 will not
completely cancel such reverberation noise, due to the differential
on-axis sensitivity for the front and rear transducers 1002,
1008.
[0119] Whereas bi-directional microphone assembly 208 requires
matched transducers such that the noise is cancelled, the hyper
cardioid requires transducers producing different on-axis
sensitivity. In particular, the transducer sensitivity differential
for transducers 1002 and 1008 needs to be 5 to 15 dB, and may for
example be 10 dB. The transducer control and damping values, which
should be considered for the hyper cardioid microphone assembly
1000, will not be important for the bi-directional polar microphone
assembly 208 so long as the transducers are the same. So long as
identical transducers are provided, the out of phase and the
omni-directional contents, such as mechanical vibration,
reverberations, sound having a frequency such that it is
non-directional, will null, in microphone assembly 208. The hyper
cardioid microphone assembly 1000 requires two different
sensitivities from the front and back transducers 1002 and 1008.
The transducers must be carefully selected to have the desired
sensitivity differential. Microphone assembly 1000 preferably uses
higher quality transducers for the front and back transducers 1002,
1008, so that the desired performance can be achieved and
sustained, than need be used for the bi-directional microphone
assembly 208.
[0120] A second order microphone assembly 1100 according to another
alternate embodiment is disclosed in FIG. 11. The microphone
assembly 1100 includes transducers 1102 and 1112. The front face of
transducer 1102 is coupled to a port 1104 through an acoustic
channel 1106. The rear face of transducer 1102 is acoustically
coupled to port 1110 through channel 1108. The front face of rear
transducer 1112 is coupled to port 1110 through channel 1114. The
rear face of transducer 1112 is coupled to port 1116 through
channel 1118. The transducers 1102 and 1112 are electrically
coupled to a circuit 1200 (FIG. 12). The sound from the front
transducer 1102 is input to the non-inverting input of an
operational amplifier 802. The signal from transducer 1112 is input
to a time delay 1202 prior to being input to the amplifier 802. The
time delay circuit 1202 introduces a time delay equal to the time
period required for sound to travel distance D2, which is the
distance from the center of the front transducer 1102 to the center
of the rear transducer 1112. The delayed signal is input to the
inverting input of the operational amplifier 802 through
potentiometer R8.
[0121] In operation, the signals originating from the front of the
microphone assembly 1100 will reach the rear transducer 1112 a
short time period after reaching the front transducer 1102. This
time delay is equal to the time required for sound to travel from
the center of the front transducer 1102 to the center of the rear
transducer 1112. Since the signal entering the rear transducer is
electronically delayed in time delay circuit 1202 by an amount
equal to the time period required for sound to travel distance D2,
the rear signal will arrive at the inverting input of the
operational amplifier 802 delayed by a time period equal to twice
the time required for sound to travel distance D2. Sound
originating from the rear, however will reach front transducer 1102
delayed by a time period equal to the time required for sound to
travel distance D2. Because the signal from the rear transducer
1112 signal is delayed electronically, in delay 1202, by a time
period equal to the time required for sound to travel distance D2,
the signal originating from the back sensed by both transducers
1102 and 1112 will be input to both the non-inverting and inverting
inputs of the operational amplifier 802 at the same time, such that
they are cancelled by the amplifier 802.
[0122] Accordingly, a null is provided for signals originating from
the rear of the microphone assembly. It will be recognized that the
greater distances D1 and D2 for the second order microphone
assembly 1100, the greater the sensitivity of the microphone
assembly. Additionally, for every distance D2, there is a crossover
frequency above which the difference in phase no longer adds to the
output, such that the highest upper frequency desired sets the
maximum distance D2. Above the crossover frequency, the microphone
will lose its directional properties and suffer frequency response
anomalies. It is envisioned that the maximum distance D2 for the
second order microphone assembly 1100 will be between 0.75 and 1.4
inches, and may, for example, be approximately 1 inch.
[0123] One issue with respect to this implementation, is the phase
shift that will occur. In particular, the higher the frequency, the
greater the phase shift that the signal will experience between the
front transducer and the rear transducer. Low frequency signals
will experience little phase shift, whereas high frequency signals
will experience a large phase shift. Since acoustic sensitivity
increases with additional phase shift, low frequency sensitivity
will be very low. However, because the signals of interest are
voice signals, which are relatively high frequency signals, the
signals of interest will not be significantly affected by this
phase shift. Additionally, it is envisioned that equalization
techniques can be used to compensate for the phase shift and low
frequency roll-off in bass sensitivity of the microphone 1100. The
front and back transducers 1102 and 1112 achieve a second order
directional function by their spacing. Additionally, the two
transducers face the same direction, such that the front face of
both the front and rear transducers port forwardly and the back of
both the front and rear transducers port rearwardly. The
transducers 1102 and 1112 are spaced by a distance D2, which is a
dimension close to D1 of the front transducer 1102, and may also be
a dimension close to the D3 for the rear transducer 1112. The
greatest output from the microphone will occur responsive to
on-axis sound in front of the microphone assembly 1100, where the
arrival delay is doubled as is the signal strength.
[0124] The vibration null and additional acoustic advantages of
microphone 208 can be gained for the microphone assemblies 1000 and
1100 by using four transducers, as illustrated in FIG. 11 for
microphone assembly 1100. In particular, optional transducers 1120
and 1130 are provided in addition to transducers 1102 and 1112. The
rear face of transducer 1120 is coupled to the front port 1122 via
channel 1124 and the front face of transducer 1120 is coupled to
port 1128 via channel 1126. The front face of rear transducer 1130
is coupled to rear port 1134 via channel 1136 and the back of
transducer 1130 is coupled to port 1128 via channel 1132. The front
transducers 1102 and 1120 are connected to opposite inputs of the
operational amplifier without delay so as to cancel
omni-directional noise. The rear transducers 1112 and 1130 are
similarly connected to opposite inputs of the operational
amplifier, after being delayed by the time period required for
sound to travel distance D2, so as to cancel omni-directional
noise. Using two pairs of transducers, each pair will achieve a
bi-directional pattern and be devoid of vibration noise. In
particular, nulls will occur at 90, 180, 270 degrees. The one main
lobe of the microphone assembly 1100 is narrow and forwardly
directed, being narrower than the bi-directional microphone
assembly 208 forward lobe, and having better off-axis noise
cancellation.
[0125] An automatic balancing circuit 1300 (FIG. 13) can be used in
place of, or in addition to, the manual balancing potentiometer R8.
Automatic balancing circuit includes a controller 1302 coupled to
receive the output of transducer 304 and variable gain amplifier
1304. The controller generates a gain control signal applied to a
variable gain amplifier 1304.
[0126] In operation, the controller monitors the signal levels
output by the transducer 304 and the variable gain amplifier 1304,
as indicated in blocks 1402 and 1404 of FIG. 14. The controller
monitors for the presence of speech in step 1406. If speech is
present, the controller does not adjust the gain of the variable
gain amplifier 1304. If speech is not present, the controller
determines whether the output of the variable gain amplifier 1304
is equal to the output of transducer 304, in step 1408. If it is
not equal, the gain of variable gain amplifier 1304 is adjusted in
proportion to the difference between the signal level at the output
of transducer 304 and the signal level at the output of amplifier
1304, as indicated in step 1410. The output of the variable gain
control will thus be equal to the signal level at the output of
transducer 306, thereby providing noise cancellation. Variation in
the relative performance of the transducers 304, 306 over time or
temperature can thus be compensated automatically by the automatic
gain control circuit 1300.
[0127] The microphone assemblies 1000 and 1100 can be manufactured
in the same manner as the microphone assembly 208, but with
different spatial relations for the transducers. For example,
whereas the transducers 304 and 306 of microphone assembly 208 are
positioned laterally an equal distance from the front and back
ports 312, 314, the transducers 1002 and 1008 are positioned one
behind the other between the front and back ports 1004, 1010, and
may for example be positioned along the longitudinal axis of the
microphone assembly 1000, through which the cross section of FIG.
15 is taken.
[0128] In particular, the microphone assembly 1000 includes an
elastomeric transducer mount 1506 into which transducers 1002, 1008
are mounted. The front of transducer 1002 ports through channel
1005 and the rear of transducer 1008 ports through chamber 1510 and
channel 1006. The front face of rear transducer 1008 ports through
channel 1011 and the rear surface ports through chamber 1510 and
channel 1006. A substantially rigid microphone housing 1512
encloses the transducer mount 1506, and includes mechanical
connectors 1504 for connection to the mirror housing 206, as well
as bottom, front and rear ports for sound to enter the microphone
for passage to the transducers. The connectors 1504 can be snap
connectors or connectors that rotate into engagement with the
mirror housing in the same manner as connectors 320. The transducer
mount 1506 provides acoustic seal with the transducers 1002, 1008,
and the circuit board 1502.
[0129] FIGS. 16 and 17 show an alternative structure for microphone
subassembly 1600. Microphone subassembly 1600, as illustrated,
includes an electronic portion 1641, which includes a first
microphone transducer 142 and a second microphone transducer 1644
mounted to a printed circuit board 1645.
[0130] Microphone transducers 1642 and 1644 may be mounted facing
one another or facing away from one another with their central axes
aligned coaxially. By mounting microphones 1642 and 1644 to face
opposite directions, the sensed pressure waves caused by the
vibrations are sensed 180 degrees out of phase from one another. By
mounting the microphone subassembly to the vehicle such that the
common central axis of the transducers is generally aligned with
the driver's mouth, the assembly effectively cancels the noise
produced by mechanical vibrations of windshield 20 and the rearview
mirror assembly of the vehicle while increasing the gain of the
driver's speech. A microphone processor circuit adds the outputs
from the two transducers to one another thereby nulling any
vibration-induced noise.
[0131] As shown in FIG. 17, transducers 1642 and 1644 may be
mounted on their sides and the subassembly may include acoustic
ports that are 90 degrees relative to the mechanical axes of the
transducers. This allows both of the natural transducer front ports
to face the redirected front port of the assembly.
[0132] According to another embodiment, the inventive microphone
assembly utilizes two micro-phone transducers facing in opposite
directions. The output of the rear facing transducer preferentially
receives noise signals while the output of the forward facing
transducer preferentially receives voice signals. Via appropriate
electronic processing the presence of significant voice signals can
be determined. During periods when there are no significant voice
signals, output can be reduced with no harm to voice quality. If
this processing is done on a frequency band basis, noise dominated
bands can be removed with no harm to voice quality since those
bands containing significant voice signals will be passed into the
output with no alteration.
[0133] Microphone transducers 1642 and 1644 are mounted sideways
through holes formed in printed circuit board 1645. Portions of
transducers 1642 and 1644 extend below the bottom surface of
circuit board 1645 and portions also extend above a top surface of
printed circuit board 1645. Mounting the transducers in this
orientation and position relative to the circuit board provides
several advantages. First, the electrical contacts on the
transducers may be directly soldered to traces on the printed
circuit board. This avoids the need for manually connecting wires
to the transducer contacts and subsequently manually connecting
those wires to the circuit board. Thus, the transducers may be
mounted to the circuit board using conventional circuit board
populating devices.
[0134] Another advantage of mounting the transducers such that they
extend above and below the surfaces of the printed circuit board is
that one side of the circuit board may include a conductive layer
serving as a ground plane. Such a ground plane may shield the
transducers from electromagnetic interference (EMI) that may be
produced by other components within the rearview mirror assembly or
in other components within the vehicle. Such EMI can introduce
significant noise into the signal delivered by the transducers. In
a preferred embodiment, each transducer is mounted in a circuit
board having a conductive ground plane facing the acoustically
active portion of the transducer while the circuit components are
mounted to the opposite side.
[0135] As shown in FIGS. 16 and 17, microphone subassembly 1600
further includes an acoustic cup 1650 having a pair of central
recesses 1652 and 1654 arranged to accept the portions of
microphones 1642 and 1644, respectively, that extend below the
bottom surface of printed circuit board 1645. Microphone
subassembly 1600 further includes a plurality of ports 1655
disposed about the peripheral bottom portion of acoustic cup 1650.
Microphone subassembly 1640 further includes a cloth 1658, which
serves as a windscreen and protects the microphones from the
external environment. Cloth 1658 is preferably made of a
hydrophobic material and is secured to cup 1650 across ports 1665
to keep water from reaching microphones 1642 and 1644.
[0136] Microphone subassembly 1600 also includes the outer
microphone housing 1660 formed in the shape of a cup with a
plurality of acoustic ports 1665 disposed about the bottom and
sides of the housing. Ports 1665 are preferably aligned with ports
1655 of acoustic cup 1650. Housing 1660 preferably includes one or
more posts 1666a-1666c that aligns and mates with grooves
1656a-1656c in acoustic cup 1650 and grooves 1646a-1646c of printed
circuit board 1645. The posts and grooves serve to align ports 1655
and 1665 while also ensuring that the microphone transducers cannot
rotate or change orientation within housing 1660. Housing 1660
further includes a plurality of tabs 1662a-1662c that resiliently
engage the peripheral edge of an aperture formed in mirror housing
206 (FIG. 2). Mirror housing 206 would preferably include
corresponding slots for receiving resilient tabs 1662a-1662c to
ensure that microphones 1642 and 1644 are optimally aligned
relative to the vehicle.
[0137] While the microphone subassembly is shown in FIG. 2 as being
mounted to the bottom of the mirror housing, it should be noted
that the preferred location is actually on the top of the mirror
housing. An example of a rearview mirror assembly having a
microphone subassembly 1600 mounted on the top of the mirror
housing is shown in FIGS. 18-20. Microphone subassemblies mounted
on a mirror housing receive not only direct sounds from the driver,
but also sounds reflected off the windshield. When the microphone
subassembly is mounted on the bottom of the mirror housing, there
is more of a time difference between the arrival of the direct
sound and the reflected sound than when the microphone subassembly
is mounted on the top of the mirror housing. When the arrival times
are far enough apart, the resulting combination produces a
frequency response that has a series of frequencies with no output.
The series, when plotted, resembles a comb, and hence is often
referred to as the "comb effect."
[0138] Mounting the microphone subassembly on top of the mirror
housing avoids the comb effect in the desired pass band. As shown
in the side view in FIG. 20, the distance between the windshield
and the top of the mirror housing is much smaller than that at the
bottom of the mirror housing and thus the reflected sound adds
correctly to the direct sound creating a louder, but otherwise
unaffected, version of the direct sound. The end result being a
higher signal-to-noise ratio and better tonal quality. These are
very important attributes in hands-free telephony and vocal
recognition in an automotive environment.
[0139] A problem with mounting the microphone subassembly to the
top of the mirror housing results from the fact that the microphone
assembly is closer to the windshield. When the windshield defroster
is activated, a sheet of air travels upward along the windshield.
Thus, when the microphone subassembly is placed on top of the
mirror housing, it is exposed to more airflow as the air from the
defroster passes between the mirror housing and the window past the
microphone subassembly. This airflow creates turbulence as it
passes over the microphone subassembly, which creates a significant
amount of noise. To solve this problem, a deflector 1670 extends
upward from the rear of mirror housing 1630 so as to smoothly
deflect the airflow from the defroster over and/or beside
microphone subassembly 1600 so that it does not impact the
transducers or create any turbulence as it passes over and around
the microphone subassembly. Because the airflow primarily would
enter the rear of the microphone subassembly, the deflector may be
designed to redirect the air with minimal impact on the frequency
response of the microphone subassembly. This is important for high
intelligibility in the motor vehicle environment. With no direct
air impact and the avoidance of turbulence near the microphone
subassembly, mounting the microphone subassembly on the top of the
mirror housing can offer superior resistance to airflow-generated
noise.
[0140] As an additional measure, a signal may be transmitted over
the vehicle bus or other discrete wire or wireless communication
link, which indicates that the windshield defroster has been
activated. This signal could be received and processed by the
microphone processor and used to subtract an exemplary noise
waveform that corresponds to that detected when the windshield
defroster is activated. Alternatively, when the system determines
that the driver is speaking into the microphone and that the
windshield defroster is activated, the system will temporarily turn
down or tarn off the defroster, or otherwise produce a synthesized
speech signal advising the driver to turn down the defroster. The
voice recognition circuitry within the mirror may also be utilized
for purposes of recognizing noise generated by the defroster such
that the system will be able to either advise the driver to turn
the defroster down or off or to perform that task
automatically.
[0141] In addition to recognizing the sound produced by the
windshield defroster, the microphone may also be used to recognize
the sources of various other sounds and hence subtract them from
the sound received while the driver is speaking. For example, the
microphone may be used to detect low pass response to determine
whether the vehicle is moving. Additionally, the microphone may be
used to recognize other events, such as a door closing or whether
the air bags have been inflated. Upon detecting that the air bags
have been inflated, the telematics rearview mirror assembly may be
programmed to call 911 and to transmit the vehicle location in a
distress signal.
[0142] FIG. 21 shows an exploded view of a microphone assembly 1700
constructed in accordance with another embodiment of the present
invention. Microphone assembly 1700 includes a pair of transducers
1702 disposed in apertures 1704 at opposite ends of a transducer
boot 1706. Transducer boot 1706 includes an inner cavity 1708 by
which the front surfaces of transducers 1702 are acoustically
coupled and to a forward-facing port 1710 in boot 1706. Transducer
boot 1706 is mounted in an aperture 1712 of a circuit board 1714.
Thus, a portion of transducer boot 1706 extends below circuit board
1714 while the remaining portion is positioned above circuit board
1714 with port 1710 extending out and resting upon the upper
surface of circuit board 1714. An advantage to using a transducer
boot 1706 or a similar structure for holding the transducers is
that the transducers may be oriented with respect to the transducer
boot, inserted into the transducer boot, and held by the transducer
boot prior to insertion to the circuit board. Following insertion,
the transducer contacts may then be soldered to the circuit board
leads. The transducer boot preferably has pegs and other details
that facilitate appropriate registration with the circuit board in
an auto-insertion apparatus. Thus, rather than requiring the
auto-insertion apparatus to somehow grasp a cylindrical transducer
and attempt to appropriately align and register the transducer with
the circuit board for subsequent soldering, the auto-insertion
apparatus would merely need to utilize the details provided in the
transducer boot to provide such proper alignment.
[0143] Microphone assembly 1700 further includes a boot cover 1720.
Boot cover 1720 includes a forward opening 1722 that extends over
the protruding port 1710 of transducer boot 1706 so as to allow
port 1710 to extend and open outside of boot cover 1720. Boot cover
1720 further includes a pair of tapered side walls 1724 that slope
farther apart toward the rear of transducer boot 1720 where a rear
opening 1726 is provided. In this manner, an acoustic port is
provided at the rear of the microphone assembly, which is
acoustically coupled via the tapered side walls 1724 to the rear
surfaces of transducers 1702.
[0144] Microphone assembly 1700 further includes a windscreen 1730,
which is preferably a hydrophobic and heat-sensitive
adhesive-coated fabric. Windscreen 1730 is adhesively attached to
the underside of a microphone assembly cover 1732 so as to extend
across ports 1734 provided in cover 1732. Cover 1732 is preferably
tightly bonded about circuit board 1714 to provide a
water-impervious enclosure for transducers 1702.
[0145] Microphone cover 1732 is shown in FIG. 21 as having a
generally square shape. It should be noted, however, that cover
1732 may be a rectangle or other shape and the size and shape of
apertures 1734 may be changed so as to adjust the directionality of
the microphone. Further, the acoustic resistivity of windscreen
1730 may be varied to also vary the directionality and polarity of
the microphone assembly. Specifically, the acoustic resistivity of
windscreen 1730 may be increased to at least about 1 acoustic
.OMEGA./cm.sup.2 and preferably has an acoustic resistivity of at
least about 2 acoustic .OMEGA./cm.sup.2.
[0146] To illustrate the effect of adjusting the acoustic
resistivity of the windscreen and the size and positioning of the
ports in the microphone housing cover, the polar patterns were
plotted for the microphone assembly with and without the cover and
windscreen surrounding the microphone transducers at four different
frequencies, which are plotted in FIGS. 22A-22D and in FIGS.
23A-23D. The polar patterns (FIGS. 22A-22D) were plotted with the
cover and windscreen in place, and then, the cover and windscreen
were removed and the polar patterns were plotted for the same four
frequencies, which are shown in FIGS. 23A-23D. Specifically, the
polar patterns shown in FIGS. 22A and 23A show the microphone
characteristics at 250 Hz, the polar patterns shown in FIGS. 22B
and 23B were taken at 500 Hz, the polar patterns shown in FIGS. 22C
and 23C were taken at 1000 Hz, and the polar patterns shown in
FIGS. 22D and 23D were taken at 2000 Hz. As apparent from a
comparison of the respective polar patterns, the rear lobe that is
present when the cover is not provided over the transducers is
effectively eliminated by appropriately configuring the cover and
windscreen.
[0147] While it has been typical in conventional microphones to
minimize the acoustic resistivity of a windscreen by increasing the
porosity of the windscreen, the microphone assembly of the present
invention advantageously utilizes a windscreen with a higher
acoustic resistivity by decreasing the porosity of windscreen and
yet obtaining not only better water-resistant properties, but to
also improved the acoustic characteristics for the microphone
assembly. The reduction of the rear lobe of the polar pattern of
the microphone assembly is particular advantageous when the
microphone assembly is mounted on a rearview mirror assembly since
significant noise may be introduced from the windshield defroster
where such noise is typically to the rear and sides of the
microphone assembly.
[0148] When the microphone transducers are sealed in separate
housings having their own cover and windscreens, the cover ports
and acoustic resistivity of the windscreens may be different for
the different transducers so as to compensate for any effects
experienced by the transducers as a result of the positioning of
the transducers on the vehicle accessory. For example, when one
transducer is mounted closer to the face of the rearview mirror,
its polar pattern is different from that of a transducer spaced
farther from the mirror surface. Thus, by selecting an appropriate
cover design and windscreen resistivity, the effects of the
differences resulting from the positioning of the transducers may
be compensated such that the transducers exhibit substantially
similar polar patterns and other characteristics. While the
windscreen has been described above as consisting of a hydrophobic
fabric, it will be appreciated that the windscreen may be molded
integrally across the ports of the microphone assembly cover. Such
an arrangement would simplify the manufacturing of the microphone
assembly by requiring less parts and less manufacturing steps.
Further, it would more likely provide a more effective seal between
the windscreen and the cover.
[0149] FIG. 24 shows yet another embodiment of a microphone
assembly 2000. As illustrated, microphone assembly 2000 is
positioned on the top of a rearview mirror assembly mirror housing
1630 in a manner similar to that shown in FIGS. 18-20. Similar to
that embodiment, a deflector 1670 is provided that extends from the
upper rear portion of mirror housing 1630 so as to provide a
relatively flat surface 2005 on which the microphone assembly 2000
may be mounted.
[0150] Microphone assembly 2000 includes two separate microphone
housings. A first microphone housing 2002 is positioned forward of
a second microphone housing 2004 and is positioned closer to the
face of the rearview mirror assembly and hence closer to the driver
of the vehicle. First microphone housing 2002 includes a cover 2012
having a plurality of ports 2008 through which sound may pass.
Second microphone housing 2004 likewise may include a cover 2014
having a plurality of acoustic ports 2010. Both housings preferably
include a windscreen similar to that discussed above. The
configuration of the ports on the covers and the acoustic
resistivity of the windscreens may be different for each of
housings 2002 and 2004 so as to compensate for any effects caused
by the positioning of the transducers on the rearview mirror
assembly.
[0151] Each of microphone housings 2002 and 2004 preferably include
a single transducer having its front surface facing the driver of
the vehicle. As shown in FIG. 25, the central axes of the
transducers and covers 2012 and 2014 may be aligned along a common
axis that is at an angle .theta. relative to a perpendicular
bisector (i.e., normal) to the rearview mirror surface. This is to
ensure the transducers are coaxially aligned with the driver's
mouth, since the rearview mirror surface would be at more of an
angle to allow viewing through the rear window of the vehicle. It
should be noted that the transducers need not be aligned coaxially,
but may be skewed with respect to one another. Such an embodiment
is described further below.
[0152] As also discussed further below, microphone assembly 2000 is
preferably a second order microphone assembly with the centers of
the two transducers physically separated by between about 0.75 and
1.4 inches, and preferably between about 1.0 to 1.3 inches. By
spacing the transducers 1.3 inches apart, the distance between the
transducers is approximately one-half the wavelength of sound at 5
kHz. The two transducers may be housed in the separate microphone
housings discussed above. In the presently preferred embodiment,
the transducers are identical and are spaced 1.0 inch apart. The
front transducer should preferably be as far forward as possible,
and most preferably should be about 0.25 inch from the front glass
surface of the mirror.
[0153] According to an embodiment of the present invention, the
transducer in each housing is ported so as to effectively be aimed
down the center of the vehicle rather than directly at the driver's
mouth. By aiming the transducers down the center of the vehicle,
the transducers are still able to clearly pick up the driver's
voice, but do not pick up nearly as much of the considerable noise
that originates to the side of the vehicle by the driver. In a
typical vehicle, the mirror is correctly positioned for a typical
driver when it is positioned between about 14 and 22 degrees
relative to the horizontal axis of the vehicle (i.e., a horizontal
axis of the vehicle is one that is parallel to the horizontal axes
of vehicle's axles). For purposes of the invention, an assumption
of 20 degrees is made so that the transducers are generally aligned
along a line that is 20 degrees from a line that is perpendicular
to the mirror surface in the direction away from the driver. This
results in the transducers generally being aligned with a line down
the center of the vehicle. It is also beneficial to slightly turn
the front transducer further away from the driver such that it is
no longer coaxial with the rear transducer.
[0154] To attempt to obtain a required sensitivity accuracy for the
transducers, a laser trim tab may be added to the gain stage
connected to each transducer. The transducers may then be
acoustically excited by a calibrated sound source and the output of
the transducers is monitored. The laser trim tab is then trimmed to
precisely set the gain and thereby obtain precise sensitivity
accuracy.
[0155] Because of the frequency response of components in existing
telephone networks, it may be beneficial to increase the separation
distance between the transducers to between 1.7 and 1.9 inches.
Because space may be limited on the accessory surface on which the
transducers are mounted, it may not be possible to physically
separate the transducers by such a distance. To overcome this
problem, a mechanical structure 2006 may be disposed between the
first transducer and the second transducer to increase the acoustic
path length between the first and second transducers. Mechanical
structure 2006 may have any symmetrical conical structure and is
shown in FIG. 25 as having the shape of a pyramid. As apparent from
FIG. 24, any on-axis sound passing by the first housing 2002
towards the second microphone housing 2004 must pass up and over
mechanical structure 2006. On the other hand, any sound coming
off-axis from the sides will still be received at the same time by
both microphone structures 2002 and 2004 regardless of the presence
of mechanical structure 2006. Test results have shown that a
pyramid-shaped mechanical structure 2006 having a height of 0.35
inch and side dimensions of 0.70 inch with a 45-degree incline of
the side surfaces toward the peak that the acoustic path length may
be increased by approximately 0.35 inch. Thus, greater acoustic
separation of the two transducers may be obtained without having to
physically separate the transducers by a greater distance. This
enables the structure to be mounted on relatively small
surfaces.
[0156] It should be noted that an additional common cover for the
microphone assembly 2000 shown in FIGS. 24-26 may be secured over
the illustrated structure provided that the common housing is
substantially acoustically transparent so as to not affect the
arrival times of the sound to the two transducers.
[0157] As shown in FIGS. 24 and 26, a surface of deflector 1670 may
include a structure designated as 2020 that is hereinafter referred
to as a "fine turbulence generator." Fine turbulence generator 2020
may be implemented using a fabric or other fine structure so as to
create fine turbulence between deflector 1670 and the laminar
airflow along the windshield defroster as it passes over deflector
1670. A preferred fine turbulence deflector may be implemented
using the loop portion of a hook-and-loop-type fastener such as the
VELCRO hook-and-loop fastener. Alternatively, the corresponding
surface of deflector 1670 may simply be roughened to create similar
turbulence.
[0158] While turbulence generally is undesirable due to the noise
it produces, creating very fine turbulence in the manner proposed
creates turbulence having frequency components that exceed the
audible limits of humans while reducing the turbulence of the air
passing by deflector 1670 that would produce lower frequency
components within the audible limits of humans. Because of the fine
turbulence created along the surface of deflector 1670, the laminar
airflow is deflected by the fine turbulence that is created rather
than the deflector itself. This reduces the friction of the
deflector as seen by the laminar airflow and therefore reduces the
turbulence created by the airflow that would otherwise tend to
create lower frequency noise within the audible frequencies.
[0159] Due to the large size of the mirror surface and the
proximity of the forward most transducer to the mirror surface, the
polar patterns of the two transducers may vary from one another on
a frequency dependent basis. In some applications, it may be
desirable to include second transducers in each of the two
microphone housings to alter the polar pattern over a frequency
range, and thereby compensate for this discrepancy. By utilizing
these additional transducers and utilizing additive signals to
correct the polar with regard to frequency, nearly identical and
optimum cardioid polar responses may be attained over the entire
desired pass band. According to one embodiment of the present
invention, the second transducer in the front microphone housing
may be an omni-directional transducer while the second transducer
in the rear microphone housing may be a cardioid transducer.
[0160] FIG. 27 shows a block diagram of a microphone processing
circuit 2100 that may be used with the second order microphone
assembly 2000 as depicted in FIGS. 24-26. It will be appreciated,
however, that microphone processing circuit 2100 may be used with
any second order microphone assembly regardless of whether it is
incorporated in a rearview mirror assembly, in another vehicle
accessory, or in any other audio application outside of the vehicle
environment.
[0161] Circuit 2100 includes a front transducer 2102 and a rear
transducer 2104. As discussed above, for a second order microphone
assembly, front and rear transducers are preferably disposed with
their front surfaces facing the direction of the person speaking.
The output 2104a of rear transducer 2104 is coupled to the input
2106a of a high pass filter 2106. The output of high pass filter
2106b is coupled to a first input 2108a of a summing circuit
2108.
[0162] The output 2102a of front transducer 2102 is coupled to the
input of 2110a of an all-pass phase shifter 2110. The output of
all-pass phase shifter 2110b is coupled to an inverting input 2108b
of summing circuit 2108. As discussed further below, phase shifter
2110 is provided to shift the phase of the signal from front
transducer 2102 by an amount equivalent to the phase shift inherent
in high-pass filter 2106 such that the signals from front and rear
transducers 2102 and 2104 have their phase shifted by equal amounts
prior to application to summing circuit 2108 where the signal from
front transducer 2102 is inverted and summed with the filtered
signal from rear transducer 2104 (i.e., the signals are effectively
subtracted). The output 2108c of summing circuit 2108 is coupled to
the input 2112a of a three-pole high-pass filter 2112. The output
2112b of three-pole high-pass filter 2112 may be coupled to the
input 2114a of an optional buffer circuit 2114. The output 2114b of
buffer circuit 2114 represents the output of the inventive
microphone processing circuit.
[0163] Microphone processing circuit 2100 as shown in FIG. 27,
includes a biasing circuit 2116, which produces a bias voltage
V.sub.B that is applied to each of components 2106-2114, as more
apparent from the schematic representations of each of those
components. Biasing circuit 2116 includes a pair of
series-connected resistors 2118 and 2120 coupled between a supply
voltage V.sub.S and ground. Resistors 2118 and 2120 preferably have
a resistance of 10 k/.OMEGA.. Biasing circuit 2116 further includes
a capacitor 2122 coupled between the output of biasing circuit 2116
and ground. Capacitor 2122 preferably has a capacitance of 2.2
.mu.f.
[0164] The details of components 2106-2114 are shown schematically
in FIGS. 28A-28E, and are discussed in further detail below
following a description of the general circuit operation.
[0165] To understand the performance and advantages of the
inventive microphone processing circuit 2100, it is first necessary
to understand the operation of a conventional circuit used with
second order microphone assemblies. In prior second-order
microphone processing circuits, the output of the front transducer
was simply inverted and provided to a summing circuit where the
signal was summed with the signal directly supplied from the rear
transducer. The frequency response of such a processing circuit is
shown in FIG. 29A. In FIG. 29A, plot A shows the sensitivity of the
second order microphone assembly at various frequencies with the
sound originating on-axis. Plot B shows the microphone sensitivity
at various frequencies with the sound originating 180 degrees from
the axes (i.e., from behind the microphone assembly). Plot C shows
the microphone sensitivity for various frequencies arriving at an
angle 90 degrees from the central axes of the transducers (i.e.,
directly from the side of the microphone assembly). As apparent
from FIG. 29A, such a microphone circuit is very sensitive to
higher frequencies, but is not very sensitive to lower frequencies
within the audible band for those sounds originating on-axis. To
compensate for the low frequency sensitivity, a high-pass filter
may be added at the output of the summing circuit. While such an
arrangement serves to provide a more uniform sensitivity across the
frequencies in the audible range, the introduction of the filter
renders the assembly extremely sensitive to vibration-induced
noise. More specifically, torsional vibration of the transducers is
amplified using such a configuration.
[0166] To overcome these problems, the inventive microphone
processing circuit utilizes a high-pass filter 2106 between one of
the transducers and summing circuit 2108. High-pass filter 2106
could be placed at the output of either front transducer 2102 or
rear transducer 2104. High-pass filter 2106 preferably has a
characteristic cut-off frequency at about 1 kHz. By filtering the
output of one of the transducers to reduce its bass frequency
components prior to subtraction from the other transducer output,
the bass of the resultant output is reduced by a smaller amount
than it otherwise would in the absence of filter 2106. As discussed
above, all-pass phase shifter 2110 is provided in the path of the
other transducer so as to ensure that the phase of the signals from
front and rear transducer 2102 and 2104 are shifted by the same
amount prior to reaching summing circuit 2108. FIG. 29B illustrates
the frequency response of the system when phase shifter 2110 is not
utilized. As apparent from FIG. 29B, there is a steep drop off in
response at the middle of the audible range, which results from the
phase difference of the signals that would otherwise be applied to
summing circuit 2108.
[0167] FIG. 29C shows the frequency response of the inventive
microphone processing circuit 2100 having the construction shown
generally in FIG. 27 and specifically in FIGS. 28A-28E and
described further below. As apparent from FIG. 29C, the sensitivity
of the microphone assembly to on-axis sound is relatively uniform
across the audible range. The on-axis sensitivity is referenced in
FIG. 29 as plot A. The 180-degree off-axis sound sensitivity is
designated in FIG. 29C as plot B. Plot C represents the microphone
assembly sensitivity to sound arriving off-axis at 145 degrees
while plot D represents sound originating from a point 90 degrees
off-axis. As apparent from a comparison of these plots, the second
order microphone assembly of the present invention is significantly
more sensitive to on-axis sound while is clearly less sensitive to
off-axis sound, particularly at lower frequencies. As noted above,
in an automobile environment, most the noise arrives off-axis
towards the sides of the microphone assembly. Thus, the above
described second order microphone assembly 2000 and circuitry 2100
is significantly less sensitive to noise originating from those
directions.
[0168] FIG. 28A is a schematic diagram showing the preferred
construction for high-pass filter 2106. High pass filter 2106
includes a first resistor 2124, preferably having a resistance of
8.2 k/.OMEGA., which is coupled between filter input 2106a and
supply voltage V.sub.S. A capacitor 2126, preferably having a
capacitance of 0.001 .mu.f, is coupled between input 2106a and
ground. High-pass filter 2106 also includes an operational
amplifier 2128, preferably part No. LM2904, having its
non-inverting input terminal coupled to bias voltage V.sub.B, and
its inverting input coupled to input terminal 2106a via
series-connected capacitor 2130 and resistor 2132. Capacitor 2130
preferably is a 0.01 .mu.f capacitor while resistor 2132 preferably
has a resistance of 10 k/.OMEGA.. High-pass filter 2106 also
preferably includes a feedback resistor 2134 coupled between the
inverting input and the output of amplifier 2128. Another resistor
2136 is coupled between the output of amplifier 2128 and ground.
Preferably, resistors 2134 and 2136 both have a resistance of 10
k/.OMEGA.. The output of amplifier 2128 serves as the output 2106b
of high-pass filter 2106.
[0169] FIG. 28B shows the preferred construction of all-pass phase
shifter 2110. Phase shifter 2110 includes a first resistor 2138
that is coupled between input terminal 2110a and supply voltage
V.sub.S. Resistor 2138 preferably has a resistance of 8.2
k/.OMEGA.. A capacitor 2140, preferably having a capacitance of
0.001 .mu.f, is coupled between input terminal 2110a and ground. A
capacitor 2142 and a resistor 2144 are coupled in series between
input terminal 2110a and an inverting input of an amplifier 2146.
Capacitor 2142 preferably has a capacitance of 1 .mu.f. A feedback
resistor 2148 is coupled between the inverting input and the output
of amplifier 2146. A resistor 2150 is coupled between the output of
amplifier 2146 and ground. Amplifier 2146 is preferably part No.
LM2904. Another resistor 2152 is coupled between the non-inverting
input of amplifier 2146 and biasing circuit 2116. A capacitor 2154
is coupled between the non-inverting input of amplifier 2146 and a
terminal between capacitor 2142 and resistor 2144. Capacitor 2154
preferably has a capacitance of 0.01 .mu.f. Resistors 2144, 2148,
2150, and 2152 all preferably have resistances of 10 k/.OMEGA.. The
output of amplifier 2146 serves as the output 2110b of phase
shifter 2110.
[0170] FIG. 28C shows a preferred construction for summing circuit
2108. Summing circuit 2108 includes an amplifier 2156 having its
non-inverting input coupled to biasing circuit 2116 so as to
receive a bias voltage V.sub.B. Input terminal 2108a is coupled to
the inverting input of amplifier 2156 via series-connected
capacitor 2158 and resistor 2160. Similarly, input terminal 2108b
is coupled to the inverting input of amplifier 2156 via
series-connected capacitor 2162 and resistor 2164. Capacitors 2158
and 2162 preferably have a capacitance of 1 .mu.f. A resistor 2166
is coupled between the inverting input and the output of amplifier
2156. A resistor 2168 is preferably coupled between the output of
amplifier 2156 and ground. Resistors 2160, 2164, and 2168 all
preferably have a resistance of 10 k/.OMEGA. while resistor 2166
has a resistance of 100 k/.OMEGA.. Amplifier 2156 is preferably
part No. LM2904. The output of amplifier 2156 serves as the output
2108c from summing circuit 2108.
[0171] FIG. 28D illustrates a preferred construction for three-pole
high-pass filter 2112. Bypass filter 2112 preferably includes an
amplifier 2170 and three capacitors 2172, 2174, and 2176 coupled in
series between input 2112a and the non-inverting input of amplifier
2170. Capacitors 2172, 2174, and 2176 preferably have capacitances
of 0.33 .mu.f. A resistor 2178 is coupled between ground and a
terminal between capacitors 2172 and 2174, a resistor 2180 is
coupled between the inverting input of amplifier 2170 and a
terminal between capacitors 2174 and 2176, and a resistor 2182 is
coupled between the non-inverting input of amplifier 2170 and bias
circuit 2116. A resistor 2184 is coupled between the output of
amplifier 2170 and ground. The inverting input and output of
amplifier 2170 are electrically coupled. Resistor 2178 preferably
has a resistance of 6.8 k/.OMEGA., resistor 2180 preferably has a
resistance of 1.1 k/.OMEGA., resistor 2182 preferably has a
resistance of 270 k/.OMEGA., and resistor 2182 preferably has a
resistance of 10 k/.OMEGA.. Amplifier 2170 is preferably part No.
LM2904. The output of amplifier 2170 serves as the output 2112b of
filter 2112. Having this construction, the cut-off frequency of
this high-pass filter is about 300 Hz. It should be noted that a
different cut-off frequency could be utilized in microphone
processing circuit 2100.
[0172] FIG. 28E illustrates a preferred construction for buffer
circuit 2114. Buffer circuit 2114 preferably includes an amplifier
2186 having its non-inverting input coupled to input terminal 2114a
via a capacitor 2188. A resistor 2190 is coupled between the
non-inverting input of amplifier 2186 and bias circuit 2116. The
inverting input of amplifier 2186 is coupled to ground via
series-connected resistor 2192 and capacitor 2194. A resistor 2196
is coupled between the inverting input and the output of amplifier
2186. A resistor 2198 is coupled between the output of amplifier
2186 and ground. A capacitor 2199 is coupled between the output of
amplifier 2186 and the output 2114b of buffer circuit 2114.
[0173] While the specific circuit implementation is described above
for microphone processing circuit 2100, it will be appreciated by
those skilled in the art that other configurations may be utilized
without departing from the scope of the invention.
[0174] In some applications, it may be desirable to purposely boost
the gain of the transducers in certain frequency ranges to
compensate for the effect of the vehicle on the frequency response
that is output from the microphone assembly. For example, a
microphone assembly was constructed having a generally flat
frequency response curve up to 5 kHz. However, when this microphone
assembly was placed in certain vehicles, the frequency response was
flat only to about 3.5 kHz and dropped off somewhat significantly
between 3.5 and 5 kHz. Thus, to compensate for the effect the
vehicle had on the microphone assembly output, the frequency band
between 3.5 and 5 kHz was purposely boosted to give the microphone
assembly a non-flat response curve to thereby compensate for these
effects and to provide a flat output signal up to 5 kHz from the
microphone assembly. Such a flat output up to 5 kHz is generally
desired when utilizing voice recognition processing.
[0175] FIG. 30 shows an alternative microphone processing circuit
that utilizes a digital signal processor (DSP).
[0176] As shown in FIG. 30, the microphone assembly may include one
or more transducers 2210. The microphone processing circuit of the
microphone assembly includes a DSP 2220 and may optionally include
a pre-processing circuit 2215 disposed between an input to DSP 2220
and an output of transducer(s) 2210. Alternatively, DSP 2220 could
be coupled between pre-processing circuit 2215 and transducer(s)
2210. The output of DSP 2220 may be applied to various devices such
as a voice recognition device, a recording device, or to a
transceiver of a radio or cellular telephone.
[0177] DSP 2220 may be any appropriately configured DSP, but is
preferably either of part nos. TMS320VC5.times.5409 or 5402
available from Texas Instruments. The microphone may, but need not
necessarily, include two or more transducers arranged as disclosed
above, while a corresponding pre-processing circuit such as those
disclosed above may also be used for circuit 2215. By using two
transducers with one spaced farther away from the person speaking,
the arrival time of sounds picked up by the transducers may be used
to determine the likely source of the sounds. For example, the
transducer closest to the person speaking will detect a sound
originating from that person before the furthest transducer.
Conversely, any sound that is first detected by the furthest
transducer may be identified as noise. Likewise, any sounds
arriving off-axis and received by both transducers at the same time
may also be discarded as noise.
[0178] Human vocal cords resonate and thereby create a single
frequency with overtones (also known as harmonics). All vocal cord
energy is therefore confined to the harmonics of the vocal cord
fundamental frequency. For a human male, the fundamental frequency
is typically between 35 and 120 Hz, and for a female, the
fundamental frequency is typically between 85 and 350 Hz. The DSP
filter 2220 of the present invention identifies the fundamental
frequency of the speech signals received by transducer(s) 2210 and
use the identified fundamental frequency to compute the
coefficients for an inverse comb filter that will pass only the
harmonics of the vocal cords of the person(s) whose speech signals
are received. In contrast to conventional noise filters that try to
identify the noise, the inventive filter identifies the speech. The
inventive filter may also be used to separate one talking person
from another as long as both have different fundamental
frequencies.
[0179] FIG. 31 shows a process diagram for the adaptive filter as
implemented in DSP 2220. As depicted in block 2225, the analog
audio signal from transducer(s) 2210 is converted into a digital
audio signal. A fast Fourier transform (FFT) is then performed on
the digitized audio signal as shown in block 2230. An example of an
FFT of an audio signal including a speech signal and noise is shown
in FIG. 32. Using the FFT of the digitized audio signal, the
fundamental frequency of the speech signal is determined as
depicted in block 2235. DSP 2220 identifies the fundamental
frequency by identifying frequency components in the FFT that have
amplitudes exceeding a predetermined threshold, and then
identifying the fundamental frequency as the difference in
frequency of those frequency components having an amplitude above
the predetermined threshold. As apparent from the exemplary FFT
shown in FIG. 32, the highest peaks are separated by an amount
equal to the fundamental frequency f.sub.0 and appear at
frequencies that are at multiples of the fundamental frequency.
Those peaks in the FFT correspond to the harmonic frequency
components of a person's speech.
[0180] After the fundamental frequency is determined in block 2235,
adaptive filter coefficients are generated (block 2240) and used to
configure an inverse comb filter (block 2245) that is used to
filter the digitized audio signal supplied by transducer(s) 2210.
An example of an inverse comb filter characteristic is shown in
FIG. 33 that is suitable for filtering a signal having the FFT
shown in FIG. 32. The filtered digital signal may then be converted
to an analog speech signal as depicted in block 2250. For a
discussion of how an inverse comb filter may be configured in a
DSP, see Digital Signal Processing Primer, by Ken Steiglitz, 1996,
ISBN 0-8053-1684-1.
[0181] As shown in FIG. 33, the inverse comb filter passes all
frequency components above a predetermined frequency, such as 2500
Hz. This may be desirable because certain higher frequency sounds
in human speech such as "S," "Sh," "T," and "P" sounds, may not be
at a harmonic frequency of the vocal cords. In a vehicle
environment where much of the noise is at lower frequencies,
passing all higher frequency components typically does not present
a problem. As described further below, DSP 2220 may be configured
to predict and hence separate such "S," "Sh," "T," and "P" sounds
in human speech from noise at those higher frequencies. Filtering,
such as spectral subtraction, can be employed in the region above
the inverted comb filtering frequencies to reduce noise in this
band.
[0182] By continuously monitoring the incoming audio signal for any
changes in the fundamental frequency, DSP 2220 may adjust the
filter coefficients in response to any detected change in the
fundamental frequency. The manner in which DSP 2220 adjusts filter
components may be pre-configured to prevent abrupt changes that may
occur when, for example, another occupant of the vehicle begins
speaking. The desired frequency response of the person speaking may
thus be estimated and maintained. Consistency in response is an
important factor in speech recognition. This adjustment is made by
comparing the relative intensity of the harmonics over the
reference time interval. This relationship will then be maintained.
For example, in the first few utterances, the second average
harmonic peak value may be 3 dB greater than that of the third. If
this relationship drifts, the original value will be restored. This
concept can also be applied to the relative intensity of the
sibilance utterances and the vocal cord levels. The resulting
speech output may not exactly reproduce a person's normal tonality,
but it will reproduce a consistent one. Combined with output level,
this adjustment should help vocal recognition by removing two very
important variables.
[0183] It should also be noted that DSP 2220 may configure two or
more superimposed inverse comb filters each corresponding to the
harmonics of different individuals in the vehicle. The system may
also be taught to default to the fundamental frequency most often,
or last, identified upon being activated so as to limit any delay
caused by the subsequent identification of the fundamental
frequency.
[0184] Blocks 2255 and 2260 of FIG. 31 illustrate an inventive
variable gain adjustment that may optionally be implemented in DSP
2220. The gain of the filtered digitized signal may be varied
(block 2255) prior to conversion into an analog signal. The amount
that the gain is varied is a function of the noise level detected
in the digitized audio signal received from transducer(s) 2210
corresponding to a polar pattern with a null facing the direction
of the driver, preferably a cardioid or super cardioid.
[0185] A second configuration for DSP 2220 is shown in FIG. 34.
According to the second configuration, two transducers are used
each having a polar pattern corresponding to a super-cardioid. The
first transducer 2302 is directed on axis towards the person
speaking (typically the driver in an automotive environment), while
the second transducer 2304 is positioned in the opposite direction
with a null in the polar facing the person speaking. In this
manner, while first transducer 2302 will pick-up the person's
speech as well as some noise, second transducer 2304 will not
pick-up the person's speech, but will only pick up noise including
much of the same noise picked-up by first transducer 2302. Thus,
the output signal of second transducer 2304 may be subtracted from
that of first transducer 2302 to remove unwanted noise. Second
transducer 2304 may alternatively haven an omni-directional polar
pattern.
[0186] The diagram in FIG. 34 shows that the audio signal of first
transducer 2302 is converted into a digital audio signal (block
2306) and that the audio signal of second transducer 2304 is also
converted into a digital audio signal (block 2308). The digitized
audio signals from both transducers are processed to detect the
presence of speech (block 2310) and are also both compared to one
another (block 2312). In response to the comparison of the signals
from first and second transducers 2302 and 2304, the gain/phase of
the signal from transducer 2304 is selectively adjusted (block
2314). The gain/phase adjusted signal from second transducer 2304
is inverted (block 2316) and is summed with the digitized signal
from first transducer 2302 (block 2318). The resultant summed
signal may optionally be converted into an analog signal (block
2320). Because the summed signal actually corresponds to the
subtraction of an adjusted audio signal from second transducer 2304
from that first transducer, the summed signal should represent the
speech (if present) with any noise removed. When speech is not
present, however, the summed signal should be a null. Speech may be
detected by performing a FFT on the received audio signal and
looking from a fundamental frequency in the range of that expected
for a human.
[0187] To appropriately adjust the gain/phase of the signal from
second transducer 2304, the detection of the presence of speech
(block 2310) may be used in the determination of the appropriate
gain/phase adjustment to be made. Further, nulls may be detected in
the summed signal (block 2322) for use in adjusting the gain/phase
of the signal from second transducer 2304.
[0188] As shown in FIG. 34, some phase adjustment (block 2324) may
be desired to introduce a phase delay into the audio signal from
first transducer 2302 that corresponds to that inherently
introduced during inversion (block 2316) of the audio signal from
second transducer 2304.
[0189] The system in FIG. 34 may be configured to adjust the gain
of the signal only when speech is detected to ensure that the gain
is not suddenly boosted during periods between speech and thereby
avoid boosting the noise level during those periods. This
configuration overcomes the problems typically associated with
using automatic gain control in which the gain is automatically
increased during periods between speech and thereby unnecessarily
amplifying noise.
[0190] It should be noted that both the functions outlined in FIGS.
31 and 34 may be combined in whole or in part to achieve various
significant improvements in speech processing.
[0191] The present invention also may use the time relationship
between vocal cord events and sibilance occurrences to identify the
spoken phoneme and recreate it correctly. This may add processing
delay but significantly improves vocal recognition. Knowing when
the vocal event occurred, the system can look for minor differences
relative to the preceding time interval. There are a limited number
of possibilities and due to noise, nature can be recreated more
universally than the more unique vocal cord noises. For example,
the system can determine that a "Sh" sound was uttered and recreate
a perfect "Sh" sound. Other utterances include the "S," "T," and
"P" sounds. These are all simple noise bursts of well defined
nature.
[0192] The environment around separated transducers significantly
disturbs the frequency response and polar of each transducer. For
example, a transducer located closer to the front surface of a
mirror in a rearview mirror assembly will experience a different
polar and frequency response than a transducer located farther
back. The inventive system can combine acoustic adjustments and
adaptive adjustment to compensate for these errors. The transducer
balance may be adjusted on an adaptive band by band basis to
minimize the dominant acoustic noise in each band. This assures the
greatest noise reduction possible. Such an adjustment is preferably
performed only during the intervals between speech utterances. Any
resulting reduction in speech level will be compensated
automatically. Noise reduction will be greater than any speech
level loss. This assures a maximum signal-to-noise ratio.
[0193] Typically, the only controlled analog aspect in complex
audio systems employing a DSP is gain control. In most other ways,
the microphone and its analog characteristics have been assumed to
have predetermined characteristics and the resulting DSP
application is developed around the microphone's predetermined
frequency response. The end result is a situation where the
microphone must have the same frequency response as the one upon
which the design was based in order to function correctly. This
situation prevents changing the microphone frequency response,
which potentially would provide other advantages.
[0194] A very important advantage can be achieved by reducing the
analog sensitivity in frequency bands that are dominated by noise.
If gain control is provided, the highest input signal typically
sets the gain level. If dynamic gain control is not provided, the
system gain is typically set at a fixed level corresponding to the
highest expected input signal. In a system having gain control,
when noise is dominant, the noise sets the gain level. This action
effectively prevents the gain from being set correctly for best
speech entry.
[0195] If the noise present creates signals having amplitudes
larger than that of speech signals, the possibility exists that the
noise generated signals will cause clipping in the analog stages
resulting in gross distortion aid very large spurious noise
artifacts.
[0196] The present invention addresses the above issues using two
different approaches. According to the first approach, the desired
microphone/analog response is created and an offset table from the
initial design frequency response is created. This table is used by
the DSP software to correct the digitized data creating values the
designed microphone would have yielded in the same conditions. In
other words, the DSP software need not be modified for the system
to utilize a microphone having a frequency response different from
that for which the DSP software was designed. The offset table is
provided to provide a microphone frequency response that the DSP
would expect from the design microphone despite the fact that the
microphone being used has a different frequency response. This
allows for the use of a microphone having a frequency response that
is more suitable for certain applications such as applications
where voice recognition is used. Since this approach would occur in
the first processes performed by the DSP, usually an FFT, no
concern would be present about the effect on the software that
currently limits microphone frequency response flexibility.
[0197] The first approach discussed above assumes a fixed frequency
response different from the designed-around response. A more
powerful use, one requiring appropriate DSP software, would be
adaptive. In this form, the DSP software can dynamically control
the analog frequency response. The DSP software could, for example,
determine that noise is dominant in a given frequency band and then
attenuate signals within that frequency band. The DSP software
could also determine if speech was dominant but deficient in a
particular frequency band and increase the gain for that frequency
band. Since the DSP software would know the impact of this action,
it could then compensate by post-digitization processing.
[0198] Utilizing such dynamic and adaptive control of the analog
frequency response assures the full dynamic range of the analog
portion, especially the CODEC, would be used for speech processing.
A gross difference in signal content between frequency bands could
be eliminated assuring all speech sound bands are present in the
resulting data. Since some noise, such as wind flutter, is not
easily discernable from speech, there might be some degree of
assumed noise. This would mean the bass response would generally be
more curtailed than other bands.
[0199] In general terms, the above two approaches seek to optimize
the analog frequency response while preserving the advantage of
iterative design in which the characteristics of the microphone are
too engrained to be directly changed without unforeseen
consequences.
[0200] According to another aspect of the present invention,
reliable continuity is provided through a two wire microphone
interface that removably couples a microphone assembly to an
electronic assembly. The microphone assembly includes a power
source and a two wire microphone interface. The microphone
interface includes two contacts that provide an audio signal to the
electronic assembly. A continuous direct current is provided
through the two contacts such that a low impedance path is
maintained between the microphone assembly and the electronic
assembly.
[0201] FIG. 35 depicts a simplified electrical schematic of a
microphone assembly (including a prior art microphone interface)
2400 coupled to an electronic assembly 2402 (e.g., a differential
amplifier stage). As shown in the circuit of FIG. 35, power is
provided to the microphone 2400 via a power source (VAUDIO). VAUDIO
is coupled to a first end of a resistor R5. A second end of
resistor R5 is coupled to a contact 2 of a connector J1. When
mated, contact 2 of connector J1 is coupled to a contact 4 of
connector J1 and to a first end of a resistor R6. A second end of
resistor R6 is coupled to a first end of a resistor R14. A second
end of resistor R14 is coupled to a contact 3 of connector J1.
Contact 3 of connector J1 is coupled to a contact 1 of connector
J1, which is coupled to a first end of a resistor R11. A second end
of resistor R11 is coupled to a common ground of the electronic
assembly 2402.
[0202] In brief, VAUDIO provides power to the microphone assembly
via a resistor R5. The current through resistors R5 and R6 provides
a charging current to capacitor C4, which serves to provide a
filtered microphone power supply (VMIC). A continuous wetting
current (DC) is provided by VAUDIO through resistor R5, contacts 2
and 4 of connector J1, resistors R6 and R14, contacts 3 and 1 of
connector J1 and resistor R11. Transistor Q1, which is coupled to
the first end of resistor R6 and the second end of resistor R14,
represents the load presented by a microphone preamplifier.
[0203] Turning to FIG. 36, a simplified electrical schematic of a
microphone assembly 2500 (including a microphone interface,
according to an embodiment of the present invention) coupled to an
electronic assembly 2502 (e.g., a differential amplifier stage) is
shown. VAUDIO is coupled to a first end of a resistor R5. A second
end of resistor R5 is coupled to a first end of a resistor R6. A
second end of resistor R6 is coupled to a contact 2 of a connector
J1. When mated, contact 2 of connector J1 is coupled to a contact 4
of connector J1 and a first end of a resistor R12. A second end of
resistor R12 is coupled to a first end of a resistor R8. A second
end of resistor R8 is coupled to a first end of a resistor R13. A
second end of resistor R13 is coupled to a contact 3 of connector
J1, which is coupled to contact 1 of connector J1. Contact 1 of
connector J1 is coupled to a first end of a resistor R11. A second
end of resistor R11 is coupled to a common ground of the electronic
assembly 2502.
[0204] As shown in FIG. 36, while an auxiliary power supply (V1)
provides power to the microphone assembly 2500 (or at least a
portion of microphone assembly 2500), the wetting current (DC) is
supplied by the electronic assembly 2502 power source VAUDIO. The
wetting current (DC) is supplied from VAUDIO through resistors R5
and R6, contacts 2 and 4 of connector J1, resistors R12, R8, R13
and resistor R11. The microphone interface, according to the
present invention, provides a wetting current for more
sophisticated microphone assemblies, such as those that incorporate
digital signal processors (DSPs), which receive power from an
auxiliary power source. The present invention allows connectors to
be used that have non-precious metal contacts, which reduces the
cost of the interface while at the same time providing a reliable
connection between the microphone assembly 2500 and the electronic
assembly 2502. The possible selection of values for resistors R5,
R6, R8, R11, R12 and R13 can widely vary provided that the gain and
bandwidth of the microphone assembly and any associated amplifiers
are not adversely affected. If desired, one of resistors R5 or R6
can be replaced with a short. Also, resistors R11, R12 and R13 can
be replaced with shorts, if desired. The value for resistors R8 and
R5 or R6 are then selected to provide an appropriate amount of
wetting current. For example, if VAUDIO is twelve volts and a one
milliampere (mA) wetting current is desired; if a 2 k/.OMEGA.
resistor is selected for resistor R5 and resistors R6, R11, R12 and
R13 are shorts, then a 10 k/.OMEGA. resistor is selected for
resistor R8. One of ordinary skill in the art will appreciate that
resistors can be more generally an impedance (e.g., R8 can be a
choke or active circuit). The component values indicated in FIG. 36
provide generally acceptable performance for the microphone
assembly utilized.
[0205] FIG. 37 depicts yet another embodiment of the present
invention where the wetting current is supplied from the auxiliary
power supply (V1). The wetting current (DC) is supplied from power
supply V1 through resistors R5 and R12, contacts 4 and 2 of a
connector J1, a resistor R8, contacts 1 and 3 of connector J1 and a
resistor R11. If desired, resistors R11, R12 and R13 can be
replaced with shorts. The value for resistors R5 and R8 are then
selected to provide an appropriate amount of wetting current. The
embodiment of FIG. 37 is particularly useful, from the view point
of the manufacturer of microphone assembly 2600, in that the only
component that a manufacturer of electronic assembly 2602 need
provide is resistor R8, across contacts 1 and 2 of connector
J1.
[0206] FIG. 38 depicts yet another embodiment of the present
invention wherein the input to the electronic assembly 2702,
provided from microphone assembly 2700, is balanced. The wetting
current (DC) is supplied from power supply (V1) through a resistor
R15, a resistor R16, contacts 4 and 2 of connector J1, a resistor
R8, contacts 1 and 3 of connector J1 and a resistor R20. If
desired, resistors R16, R17 and R20 can be replaced with shorts.
The value for resistors R8 and R15 are then selected to provide an
appropriate amount of wetting current. The wetting current (DC) can
be supplied from a voltage supply, a resistor, a constant current
source, inductor or other power source connected to one of the
microphone assembly leads. Providing that the microphone has a DC
path for it to complete the wetting current circuit, the source of
the current is immaterial.
[0207] As shown in FIG. 38, the audio is AC coupled from the
microphone assembly output stage to the electronic assembly 2702.
The present invention can be extended to multiple connectors that
may be included within a microphone assembly or an electronic
assembly. According to the present invention, all connectors have a
DC current flowing through them to maintain a wetting circuit.
Thus, oxidation of the contacts will not disadvantageously affect
the circuits utilizing embodiments of the present invention.
Additionally, the DC voltage of the microphone input can be used to
verify interface continuity for built in test capability.
[0208] FIGS. 39A-39D show an alternative embodiment of the present
invention in which deflector 1670 includes a cloth deflector
portion 3000. Cloth deflector 3000 advantageously deflects airflow
from the defroster away from the microphone assembly while allowing
sound reflecting off the windshield to penetrate the cloth and
reach the microphone assembly. Cloth deflector portion 3000 enables
deflector 1670 to be made more compact. By making deflector 1670
more compact, it is less likely to strike the windshield and limit
upward movement of the rearview mirror assembly towards the
windshield. If the deflector were simply made more compact without
using cloth deflector portion 3000, it would be much less
effective. While cloth deflector portion 3000 extends upward
farther than deflector 1670 would otherwise extend, cloth portion
3000 is deformable and does not limit upward movement of the mirror
assembly and preferably contacts the windshield to further prevent
airflow from reaching the microphone assembly.
[0209] Cloth deflector portion 3000 is preferably made of a
polyester material having a weave that is open enough to allow
sound to pass through without also allowing significant airflow
through the cloth. The cloth is preferably the same material that
is used for the windscreens built into the microphone housings. A
preferred cloth material has 120 .mu.m mesh holes, a 49 cm mesh
count, a thread diameter of 80 .mu.m, and a 35% open area. The
cloth deflector portion 3000 may be attached to the rear of
deflector 1670 by any suitable means such as an adhesive or the
like.
[0210] Cloth deflector portion 3000 preferably extends behind the
microphone assembly to a height higher than the microphone
assembly. By so configuring the cloth deflector portion, the cloth
deflector is better able to deflect the airflow from the defroster
past the microphone assembly. FIG. 40 is a graph showing the
microphone output in dBV versus frequency measured with a vehicle
defroster turned on full speed for three different configurations.
The first configuration had front and rear microphone
transducers-mounted on the top of a mirror housing with no air
deflector. Plot A in FIG. 40 represents the direct output of the
front microphone transducer while plot B represents the direct
output of the rear microphone transducer. Plot C represents the
output of a microphone assembly having an air deflector similar to
that shown in FIGS. 18-20. Plot D represents the output of a
microphone assembly having an air deflector with a cloth deflecting
portion similar to that shown in FIGS. 39A-39D. As apparent from a
comparison of plots C and D shown in FIG. 40, adding a cloth
deflecting portion reduces the noise from the defroster
approximately 15 dBV from a deflector without such a cloth portion.
As apparent from a comparison of plots A, B, and D, the deflector
including a cloth deflecting portion reduces the defroster noise
approximately 30 dBV relative to a mirror assembly having no
deflector.
[0211] When a DSP is utilized to process the microphone assembly
output signals, it is desirable to provide the DSP with the outputs
from two laterally spaced-apart microphone transducers. One example
is a microphone assembly utilizing two laterally offset transducers
as shown in FIG. 11 and described above. By providing the DSP with
two such output signals rather than adding or subtracting the
signals from one another first before providing the resultant
signal to the DSP, the DSP may adaptively utilize the information
from the separate signals. For example, by laterally spacing
microphone transducers and providing the separate output signals to
a DSP, the DSP may monitor the noise levels on both microphone
transducer output lines and select the output of one transducer
over the other when excessive noise is produced on the other
transducer. It has been discovered that wind noise produced by the
defroster and wind arriving from the vehicle windows or moon roof
is often quite gusty such that, if the transducers are spaced
adequately far apart, the wind noise may temporarily affect one of
the two transducers without affecting the other. As will be
discussed further below, it is advantageous to angle the central
axes of the two laterally separated transducers so as to provide
different directional characteristics for each of the transducers
such that the DSP may then utilize this additional directional
information to reduce the level of noise. As also described below,
different directional characteristics may be achieved by modifying
the configuration of the housing ports and windscreen(s).
[0212] FIGS. 41A-41D show another embodiment of the present
invention. According to this embodiment, two separate microphone
assemblies 3502 and 3504 are provided on a flat upper surface
portion 2005 of mirror housing 1630 that is circumscribed by
deflector 1670. Flat surface portion 2005 is preferably angled
slightly downward toward the rear of housing 1630 to provide for
additional clearance from the windshield. As best shown in FIG.
41D, microphone assemblies 3502 and 3504 are laterally offset from
one another relative to the driver, preferably by at least two
inches, and have their respective transducer central axes A and B
provided at an angle relative to a normal N of the surface of the
mirror. By angling the central axes A and B of the transducers
within microphone assemblies 3502 and 3504, the directional
characteristics of the microphones are further modified (in
addition to their lateral spacing) so as to provide a DSP circuit
with yet additional information from which to process voice signals
and eliminate noise. By so locating the two microphone assemblies
3502 and 3504, they tend to receive different airflow impulse. By
angling the transducers away from one another so that their central
axes A and B are not parallel to one another and are angled with
respect to normal N, it is more likely that only one of the two
microphone assemblies will be severely impacted by a wind gust at
the same moment in time. Preferably, wind deflector 1670 is
utilized to deflect the laminar airflow coming from the rear of the
mirror housing over the microphone assemblies. Any side deviation
will deflect the flow from one transducer as it drives toward the
second. The end result is one transducer is left free from wind
excitation. Since this effect causes great differences with
relatively close spacing, the desirable acoustic properties
associated with a fairly close spacing are preserved. The end
result is all of the benefits from the use of an air deflector are
obtained when dealing with airflow arriving from a central
defroster vent. This embodiment assures that in the case of a
deflection of the airflow, only one transducer is likely to be
impacted. The other transducer is, therefore, free from gross
airflow noise.
[0213] As described below it is possible, and sometimes preferable,
to include both transducers for the microphone assemblies 3502 and
3504 within the same windscreen and enclosure. In some
circumstances, it may be preferable to use two acoustically
separated windscreens. The use of two separate windscreens assures
that the transducers will be reacting only to local wind impact.
This further assures that native airflow differences will be
retained after the application of conventional airflow defense.
[0214] As shown in FIGS. 41A-41D, a separator 3500 may be provided
between microphone assemblies 3502 and 3504. Separator 3500
provides a physical side airflow deflector. In this manner, the
transducer provided on the leeward side of separator 3500 is
totally free from airflow impact. Separator 3500 increases the
difference between the two transducer(s) reaction to airflow
arriving from the side rather than from below.
[0215] By rotating the transducers relative to the mirror and/or to
each other, airflow difference is further increased. In addition,
the resulting change in aiming angle creates the opportunity to
achieve a degree of acoustic noise reduction through transducer
selection. If airflow is not the dominant noise, a significant
difference in acoustic noise resulting from the different null
locations can be used by the DSP for noise reduction through the
selection of one of the two transducers. For example, when the left
transducer has a null shifted right and the right transducer has a
null shifted left, and when there is more noise on the left
transducer, the DSP software would select the right transducer and,
with no additional processing, achieve noise reduction with no
impact on the signal quality.
[0216] The polar differences between transducers 3502 and 3504 may
be exploited with essentially no on-axis difference. During periods
where the noise in both signals is of the same relative magnitude,
a comparison of the spectrums will reflect the relative polar
difference at the angle on entry. From this difference and pattern
matching to the location that would yield the difference, the
location of the sound can be determined. Once the location
difference pattern is established, a spectral band not fitting the
pattern can be safely removed from the signal. The fundamental
advantage of all of the above actions pertaining to the embodiment
shown in FIGS. 41A-41D is that the resulting audio signal is
undistorted and has a consistent frequency response. This is in
contrast to conventional DSP processing where the process leaves
artifacts and is inherently a high distortion process.
[0217] Also, by laterally spacing two transducers and providing
them on a mirror, the time of arrival can additionally be used to
determine the location of a sound burst. Any burst not arriving
with the time difference associated with the driver is not
passed.
[0218] A preferred construction of a microphone assembly is
described below with respect to FIGS. 42-49, in which the ports,
windscreen, and/or the first and second transducers are configured
such that the null of a first polar sensitivity pattern associated
with the first transducer is aimed at the driver of the vehicle,
and the null of a second polar sensitivity pattern associated with
the second transducer is aimed at the front passenger area of the
vehicle. In general, the first and second transducers are spaced
closely together in a relatively small and narrow microphone
housing having a windscreen with very high acoustical resistivity
(i.e., about 8 to 9 acoustic ohms/cm.sub.2) disposed across the
ports of the housing. This approach seeks to maximize the
correlation of noise in the two transducers and de-correlation of
the speech signals. DSP processes can then subtract one from the
other, greatly reducing the noise and enhancing the signal content.
This design is very effective against wind noise and acoustic noise
in highly reverberant conditions. It can also differentiate between
passenger and driver side noise or speech.
[0219] By rotating the polar sensitivity patterns that would
otherwise be exhibited by the two transducers such that they have
their nulls aimed at either the driver or the front passenger, the
forward lobes of the polar patterns partially overlap which
improves noise correlation. Close spacing and possibly a common
frontal feed structure assure wind excitation will be highly
correlated as well. Since noise from the center and front of the
vehicle and air flow noise are usually the dominant noises,
addressing these effectively is very significant. Aiming the null
of one transducer away from the cab may seem to be counter
intuitive as it decreases the driver content, but in this
construction, one transducer remains almost as sensitive to the
driver and the very low driver signal content in the signal from
the other transducer assures the driver signal will emerge from the
subtraction-like process in the DSP. Very significant is the fact
that the noise in both transducers is nearly identical. The null
steering described above may be achieved frequency band by
frequency band so polar complexity due to the mirror and other
factors can be compensated as well. Once accomplished for a
condition, no further processing would be needed and no distortion
would be caused.
[0220] The inventive microphone construction described generally
above and in detailed embodiments below, achieve performance levels
only previously achieved by systems using transducer arrays that
consume significantly larger spaces and require positioning in
multiple locations on a mirror with resulting cabling and other
secondary cost aspects. The inventive microphone construction is
preferably located on the top of a rearview mirror assembly in an
area proximate a deflector as discussed above. Nevertheless, the
inventive microphone may be mounted at other locations on a mirror
assembly, including on the mirror assembly mounting structure, as
well as in any other vehicle accessory such as a headliner, sun
visor, overhead console, A-pillar, or a console extending between
the headliner and a mirror assembly.
[0221] Using the above inventive microphone construction, the
associated DSP software may process the two transducer signals by
adjusting long-term subtraction during non-speech times to the
lowest possible value. This may be in the form of sensitivity
changes by frequency band such that non-speech times for that band
were minimal. This assures that noise is at the lowest U value
after subtraction. Also, during driver speech, the output is
minimized from the virtual driver microphone created by the
inventive microphone construction. During front passenger speech,
the output from the virtual passenger microphone is minimized. A
starting point from both of these minimizations can be in the form
of a calibration using a sound source in an actual vehicle. This
can be real time or in the form of stored vehicle specific values.
In other words, the DSP software is given coefficients for
computing these two special locations either via calibration of the
current system or from data obtained from a test system.
[0222] An additional advantage of this construction is its inherent
noise cancellation. Echo tends to enter both transducers at the
same phase and strength so it will cancel like any of the other
noises. This is also true of road noise coming thee the rear center
of the vehicle.
[0223] Four different microphone assembly configurations that
achieve these benefits are discussed below.
[0224] FIG. 42 shows a first construction of a microphone assembly
3600 employing the above features. As shown, a first transducer
3602 and a second transducer 3604 are mounted with their front
surfaces 3606 and 3608, respectively, facing one another. First and
second transducers 3602 and 3604 are further aligned with their
central axes being co-linear. Both transducers are housed in a
common microphone housing 3610, which is shown in outline in FIG.
42 and shown perspectively in FIGS. 43A and 43B. First transducer
3602 generally faces the front passenger and second transducer 3604
generally faces the driver.
[0225] Microphone housing 3610 includes numerous ports.
Specifically, housing 3610 includes four upper/side ports
3612a-3612d resembling elongated slots that extend sideways across
the top 3614 and sides 3616a and 3616b of housing 3610. Four ports
3618a-3618d are provided in the front surface 3620 (i.e., the side
of the housing facing the rear of the vehicle) of housing 3610. In
the rear surface 3622 of housing 3610 are provided two ports 3624a
and 3624b, which are spaced apart from one another by a distance
exceeding at least one to two times the size of the port openings.
Two additional rear ports 3626a and 3626b may be provided between
ports 3624a and 3624b, although, for the reasons stated below,
ports 3626a and 3626b are preferably plugged or not present or
otherwise open.
[0226] A windscreen material (not shown) is preferably sealed
across each of the open ports of by housing 3610. This windscreen
preferably has an acoustic resistivity of between about 8 to 9
acoustic ohms per square centimeter. This greatly reduces wind flow
noise, while permitting null steering to aim the nulls at the
driver and the front passenger seat.
[0227] FIG. 44 shows the two polar sensitivity patterns for the two
transducers 3602 and 3604 of the microphone assembly 3600 shown in
FIGS. 42, 43A and 43B. As will be apparent from FIG. 44, the nulls
of the two patterns are aimed at either the driver (at about 60
degrees) or the front passenger (at about 300 degrees). This is
achieved by blocking or eliminating ports 3626 and is aided by
using a high resistivity windscreen and a relatively small and
narrow acoustic chamber in housing 3610.
[0228] A second configuration for achieving similar advantages
includes first and second transducers 3602 and 3604 aligned in the
same manner shown in FIG. 42. The microphone assembly 3650
according to this embodiment differs, however, in that the
microphone housing 3660 includes different ports that are plugged
or eliminated. Specifically, in this configuration, which is shown
in FIGS. 45A and 45B, front ports 3618a and 3618b are plugged or
otherwise eliminated, rear port 3626a is-open, and rear ports 3626b
and 3624b are plugged or otherwise eliminated. Such porting creates
a non-symmetric port configuration that accounts for rotation of
the rearview mirror assembly (when the microphone assembly is
mounted on the rearview mirror). FIG. 46 shows the two polar
patterns associated with the two transducers 3602 and 3604 when
used in the housing 3660 shown in FIGS. 45A and 45B. One benefit of
plugging or eliminating some of the ports, particularly in the rear
of the housing, is that this further blocks the direct air flow on
the transducers thereby lowering airflow noise.
[0229] FIG. 47 shows a third microphone assembly 3700 that achieves
null steering and its inherent benefits. This third construction
differs in that the microphone housing 3710 (FIGS. 48A and 48B)
does not have any of the aforementioned ports plugged or
eliminated. Instead, the nulls of the polar patterns are aimed by
rotating second transducer 3604 thirty degrees so that its front
surface is aimed more directly at the driver. The transducers
should be closely spaced together to avoid phase differences, and
should thus preferably be spaced with their diaphragms about one
half inch apart.
[0230] FIG. 49 shows a fourth microphone assembly 3750 that
achieves null steering and its inherent benefits. This fourth
construction utilizes the same housing 3710 (shown in FIGS. 48A and
48B) as the third construction, but differs from the third
construction in that the first and second transducers 3602 and 3604
are both rotated such that the rear surface of the first transducer
3602 more directly faces the driver and the rear surface of the
second transducer 3604 more directly faces the front passenger. The
each transducer is preferably rotated about 25 degrees with an
included angle between the front surfaces of the transducers at
about 50 degrees, however good performance is possible with each
transducer rotated anywhere from zero to 60 degrees. The preferred
rotation is dependent on the vehicle and the assumed rotation of
the mirror (when disposed thereon). The rotation angle is most
preferably two times the assumed mirror rotation angle. A typical
range of assumed mirror rotation angles is between 10 and 25
degrees thus yielding a typical preferred transducer rotation angle
of 20 to 50 degrees. In this construction, the driver speech enters
the null angle of one transducer and had very little driver speech
content, while the driver speech enters the other transducer at 90
degrees to its null and has content roughly 6 dB below the on-axis,
but still far greater than the null of the other. By aiming the
frontal lobes of the transducers towards the same point on the
windshield, equal amounts of noise are present so that when the two
signals from the two transducers are subtracted, the noise
cancels.
[0231] Another embodiment of the present invention is shown in
FIGS. 50A-50E. In general, DSP processes relating to microphone
arrays, beam forming, and polar steering exploit predictable phase
differences between the signals obtained from transducers located
at different locations. This, in turn, requires transducers to be
spaced close enough to present phase difference inversion for the
highest frequency addressed by the process. The embodiment
discussed below uses true time of arrival and as such can use far
greater spacing or one large spacing over the entire speech
bandwidth. The present embodiment uses the difference in polar
response between the two transducers as a location determining
mechanism. Unlike second order concepts, use of the difference in
polar response is also independent of spacing. This concept can be
implemented to separate sub-bands on the basis of origin and
relative magnitude. Similar to the embodiment disclosed above with
reference to FIGS. 41A-41B, the present embodiment uses the concept
of gating. The conditions determine whether a signal or signal
component are passed. This is in contrast to techniques that filter
by adding or subtracting to form the passed signal. The advantage
is that there is less distortion and fewer limitations on the
design of the system. In broad terms, this embodiment of the
invention extracts wanted sound signals from high levels of ambient
noise.
[0232] This embodiment effectively creates two electronic ears that
will supply signals that are free of non-acoustic noise and rich in
data supporting advanced DSP processes. Specifically, these
artificial ears are free of airflow and vibration noise. The degree
of airflow resistance being such that flow noise is insignificant
relative to the threshold of concern. Therefore, the present
embodiment has no detrimental effect on resulting DSP operations.
The freedom from non-acoustic noise and the presence of very
significant and consistent position is then used to define a series
of processes capable of extracting very natural sound and spectral
content speech from vehicle conditions severe enough that speech
quality is typically degraded to the point of poor vocal
recognition performance. This embodiment works particularly well
when provided on a rearview mirror of a vehicle insofar as the
mirror is effectively positioned in "free space" and positioned
such that the maximum angular separation exists between sound
source locations. The preferred form uses other aspects of the
mirror location such as the presence of the windshield to predict
noise arrival angles and perfect the artificial ears ability to
operate effectively in this environment.
[0233] Typically, microphones in automotive applications produce
very high outputs as the result of the air flowing past them. In
contrast, ears have virtually no airflow sensitivity. Since airflow
noise has none of the relationships expected in acoustic noise, it
interferes with noise reduction processes. Artificial ears are
achieved by laterally separating the microphone assemblies 3802 and
3804 at opposite ends of a mirror housing 1630 (preferably spacing
the assemblies at least about 5 cm apart, more preferably about 18
cm) and by hyper-extending the "D" of the transducers 3820 to at
least about 3.5 mm, more preferably to at least about 10 mm. This
creates a very high acoustic sensitivity of one component of the
audio sensing that a microphone utilizes. A secondary rear cavity
(3826, FIG. 50E) of greater volume is created with an acoustic
resistor 3828 placed at a rear port 3808 with the cavity between
resistor 3828 and the rear of a transducer 3820. This cavity 3826
increases the sound sensing mechanism to restore the relationship
needed to achieve the desired polar properties. A high acoustic
resistance cover 3822 similar to 3828 is placed over external
forward port 3806 to severely damp the ports. This very high
damping of both ports lowers the acoustic sensitivity and the
airflow noise. Since the dominant vibration to microphone output
conversion comes from the vibration of the microphone against
static air, vibration noise is also reduced.
[0234] The end result is a normal acoustic sensitivity with
profoundly lower airflow and vibration noise. Since this is a
fundamental improvement in signal to airflow noise, it applies to
all airflow coming from any direction. The resulting long "D" of
the microphone assembly is positioned along the rear surface of the
mirror housing along a diagonal with the lower portion angled
inward toward the middle and the upper portion angled outward. This
results in a high degree of noise rejection for sounds coming from
below the mirror along the windshield boundary and a great deal of
difference for sounds coming at an angle to the common axis between
the two due to the angular positioning.
[0235] The use of the hyper-long "D" improves greatly the
directional properties for the majority of the passed band. Higher
frequencies are not necessarily helped. This deficit is addressed
by adding directional means such as a partially horned (or flared)
opening 3810, 3816 toward the forward port 3806, 3812. While
frequency response may be negatively impacted, this aspect maybe
corrected by electronic equalization ideally done prior to
digitization.
[0236] By providing similar constructions on opposite sides of the
rearview mirror housing 1630, two signals may be obtained that are
free from airflow noise and that reject the dominant spatial noise
location, each with a very high degree of directionality and each
aimed to provide a great degree of spectral difference related to
source angular position. Since microphone assemblies 3802 and 3804
are widely spaced, there is also a significant arrival time
difference. These "artificial ears" produce all the data types and
freedom from unwanted airflow and mechanical noise needed for the
companion DSP algorithms. The DSP algorithms may thus exploit the
additional data and enjoy the freedom from non-acoustic noise
content.
[0237] The preferred default for most applications is to have the
DSP provide no signal until speech is detected. Thus, the preferred
process is based on not passing a signal unless the speech
detection criteria are met as opposed to always passing the signal
and trying to lower the noise content. While this may provide
processing delays, compensation can be accomplished by providing a
slight delay in the delivered signal to allow processing and yet
not use the first utterance of a spoken word. The process begins
with the determination in each ear channel that a change in the
input has occurred consistent with a speech utterance. This is a
well-established DSP process.
[0238] The difference in the case of the present invention is that
this action is done in two channels by only passing speech-like
events. The present invention avoids times when speech content is
so low that it is virtually useless. The threshold may be set
higher for more robust vocal recognition and better speech quality
or may be set lower for higher noise to speech situations. The time
of arrival may then be utilized to begin the process of processing
only that speech from the desired spatial location (i.e., the
location in which the driver or other passengers are located).
Incorrect arrival time difference will narrow the possibly
conflicting noises to those arriving from a line of source
locations around the central axes of the two ports 3806 and 3812.
Then, by applying the DSP's stored knowledge of the desired user,
and of human speech in general, the user's fundamental frequencies
may be determined to create a comb-pass filter. The result is that
only those bands likely to contain speech are present. This is most
effective in the bands dominated by vocal cord sounds.
[0239] At this point, any bands are passed that are likely to
contain speech and only those sounds from the correct location and
only those sounds that vary like speech are passed. The relative
spectral content may then be used to further add location
separation. For every spatial location, one can map the relative
frequency responses for that entry angle for both signals. One
would only need to address those regions where speech bands are
present. By comparing the difference in spectral content of the two
signals from microphone assemblies 3802 and 3804, the DSP can
determine if the current dominant signals are coming from the
focused location. Even more useful, the DSP can determine if a time
varying band in the passed bands originates from other than the
focused location. This is achieved by comparing the relative
magnitudes to the response maps. For example, if the difference
should be +3 dB left versus right, and the difference is -2 dB, the
DSP will know that this particular band did not originate at the
focused spatial location and can be removed. At this point, only
speech sounds from the desired location have been passed.
[0240] At this time, the DSP's knowledge of the target user may be
used to reconstruct missing speech bands. Specifically, there will
be bands where there is important speech content, but the speech
content is not large enough to be significant and will be lost in
the filtering process. Humans know what a speaking person sounds
like from less noisy times and apply that knowledge during very
high noise conditions to extrapolate the speech bands. The DSP may
use the same form of processing. Specifically, over time, the DSP
may generate a harmonic amplitude map for the range of observed
fundamental frequencies. If the fundamental frequency is known, it
may be used as the map reference and extract the relative
magnitudes of the harmonics. Since every human has a consistent
harmonic map, as the result of fixed head cavities, the DSP can
apply the known harmonic amplitudes to estimate the missing ones.
For example, human speech usually loses its high frequency content
in very high noise environments. In lesser noise, where some of the
high bands are not lost, knowledge may be gained of this speech and
used to fill in the missing bands in the higher noise
environments.
[0241] The sequence of filters and the number of filters used can
vary depending on need, benefit, or cost. The key being to exploit
the rich data derived from the artificial ears and the knowledge of
the speaking human to yield speech free from the detrimental
effects of high noise. With reference to FIGS. 50A-50E, it is noted
that the two microphone assemblies 3802 and 3804 are integrated
into the rear of the mirror housing 1630 and are disposed such that
the central axes of the transducers provided in these assemblies
are at an angle with respect to one another and with respect to a
normal to the mirror surface. Further, the transducer central axes
are aimed at an angle upward relative to the position of the
driver. This allows the microphone assemblies to be integrated more
to the rear of the mirror assembly and somewhat obscured from the
view of the driver or other passengers.
[0242] While the above embodiment addresses the problems in the
automotive environment on a broadband basis, the transducers used
may be omni-directional and the DSP could utilize time of arrival
for the lower frequency bands while using the directional
characteristics provided by the horn at the forward port for the
higher frequency bands.
[0243] When a microphone assembly 3900 is remotely located in the
vehicle from an associated DSP circuit 3912 (FIG. 51), induced
noise is typically present on the electrical conductor 3908
extending from the microphone assembly to the DSP circuit. To
eliminate this noise, a reference line 3901 is also run from the
location of the microphone assembly to the DSP circuit. An
impedance matching circuit 3903 is provided at the microphone end
of the reference line 3901 to match the impedance of the microphone
transducer 3902. Because only induced noise is present on this
reference line, the induced noise may be detected and then
subtracted from the signal delivered from the microphone assembly.
When more than one signal from one or more microphone assemblies
are to be delivered to a DSP circuit, the number of lines that must
be run through the vehicle are correspondingly multiplied. For
example, for a system utilizing two microphone transducers with two
corresponding output signals to be delivered to the DSP circuit, at
least one, if not two, reference lines may be required. The
addition of all these electrical conductors extending through the
vehicle adds significantly to the cost of such a system.
Accordingly, the need exists for a system that would allow for more
than one transducer to be utilized while minimizing the number of
electrical conductor lines that need to be run to a remote DSP
circuit.
[0244] To eliminate the need for the reference lines in the above
system, a circuit such as that shown in FIG. 52 may be utilized.
Specifically, with two transducers 3902 and 3904 provided in a
mirror assembly 3906, the respective output lines 3908 and 3910 are
provided to a DSP circuit 3912 that is remotely located from the
rearview mirror assembly or other vehicle accessory in which the
microphone transducers are mounted. A phase inverter 3914 is
provided in the path of one of lines 3908 and 3910 in order to
invert the phase of the acoustic signal sensed by one of the two
transducers. DSP circuit 3912 will know in advance that the audio
signal from this particular transducer is inverted and process it
accordingly. The noise that is induced on the two lines, however,
between the mirror assembly and the DSP circuit will not be
inverted. Thus, the DSP circuit may differentiate the audio signals
from the noise that is common on both lines 3908 and 3910. Since
there is little time of arrival or phase difference between the two
fairly closely spaced microphone transducers, there will be very
little, if any, driver speech content lost by the cancellation
process. The only acoustic content that might be lost is noise or
other sound arriving such that significant phase differences in the
outputs occur. The DSP may alternatively re-invert the phase of the
second signal and then differentiate the audio signals from the
line-induced noise based on common signals that are out of phase
with one another.
[0245] The microphone assembly described above can be incorporated
anywhere in the interior of a vehicle. For example, the microphone
assemblies can be located within the interior trim of a vehicle, an
overhead console, a visor, a rearview mirror assembly, the housing
of an electronic rear vision display, or within a mini-overhead
console provided near the rearview mirror mounting structure on the
windshield. In a preferred embodiment, the microphone assembly is
incorporated within or on an automotive rearview mirror assembly.
If desired, the contacts of the connector that couples the
microphone assembly to the electronic assembly can be plated with a
precious metal (e.g., gold or silver) to facilitate improved
continuity.
[0246] Thus, it can be seen that an improved microphone assembly
for vehicles is disclosed. It is envisioned that the microphone
assembly may be applied to a wide variety of performance
applications, in that the microphone assembly can include a single
transducer or multiple transducers. By using multiple transducers,
significantly improved performance is achieved. Use of one
transducer, having a single diaphragm or multiple diaphragms
suitably ported to achieve a desired directional pattern, offers a
lower cost microphone that can be used in the same mount and
housing as the multiple transducer microphone assembly, in
applications where the higher performance is not required.
[0247] The rearview mirror assembly 4001 according to an
alternative embodiment is shown in FIGS. 53A-53E. The mirror
assembly 4001 includes a mirror 4008 mounted in an elongated mirror
housing 4006 pivotably carried on mirror support 4004. The mirror
4008 may be any conventional interior rearview mirror, such as a
prismatic mirror of the type used with a mirror housing manually
adjustable for daytime and nighttime operation, or a multiple
element mirror effecting automatic reflectivity adjustment, such as
an electrooptic or electrochromic mirror. The elongated mirror
housing 106 may be of any conventional manufacture such as
integrally molded plastic.
[0248] As will be explained in more detail below, two microphone
assemblies 4020a and 4020b are provided along the back surface 4007
of mirror housing 4006 (i.e., that surface facing forward of the
vehicle). As apparent from FIG. 53A, microphone assembles 4020a and
4020b are not visible from the front of the mirror assembly and
hence are generally not visible to the vehicle occupants.
[0249] In general, DSP processes relating to microphone arrays,
beam forming, and polar steering exploit predictable phase
differences between the signals obtained from transducers located
at different locations. This, in turn, requires transducers to be
spaced close enough to present phase difference inversion for the
highest frequency addressed by the process. The first embodiment
discussed below uses true time of arrival and as such can use far
greater spacing or one large spacing over the entire speech
bandwidth. The present invention uses the difference in polar
response between the two transducers as a location determining
mechanism. Unlike second order concepts, use of the difference in
polar response is also independent of spacing. This concept can be
implemented to separate sub-bands on the basis of origin and
relative magnitude. The present embodiment uses the concept of
gating. The conditions determine whether a signal or signal
component are passed. This is in contrast to techniques that filter
by adding or subtracting to form the passed signal. The advantage
is that there is less distortion and fewer limitations on the
design of the system. In broad terms, this embodiment of the
invention extracts wanted sound signals from high levels of ambient
noise.
[0250] This alternative embodiment effectively creates two
electronic ears that will supply signals that are free of
non-acoustic noise and rich in data supporting advanced DSP
processes. Specifically, these artificial ears are free of airflow
and vibration noise. The degree of airflow resistance being such
that flow noise is insignificant relative to the threshold of
concern. Therefore, the present embodiment has no detrimental
effect on resulting DSP operations. The freedom from non-acoustic
noise and the presence of very significant and consistent position
is then used to define a series of processes capable of extracting
very natural sound and spectral content speech from vehicle
conditions severe enough that speech quality is typically degraded
to the point of poor vocal recognition performance. This embodiment
works particularly well when provided on a rearview mirror of a
vehicle insofar as the mirror is effectively positioned in "free
space" and positioned such that the maximum angular separation
exists between sound source locations. The preferred form uses
other aspects of the mirror location such as the presence of the
windshield to predict noise arrival angles and perfect the
artificial ears' ability to operate effectively in this
environment.
[0251] Typically, microphones in automotive applications produce
very high outputs as the result of the air flowing past them. In
contrast, ears have virtually no airflow sensitivity. Since airflow
noise has none of the relationships expected in acoustic noise, it
interferes with noise reduction processes. Artificial ears are
achieved by laterally separating the microphone assemblies 4020a
and 4020b at opposite ends of mirror housing 4006 (preferably
spacing the assemblies at least about 5 cm apart, more preferably
about 18 cm) and by hyper-extending the "D" of the transducers
4025a and 4025b of respective assemblies 4020a and 4020b to at
least about 8 mm, more preferably to at least about 15 mm. This
creates a very high acoustic sensitivity of one component of the
audio sensing that a microphone utilizes. A secondary rear cavity
(4026, FIG. 53E) of greater volume is created with an acoustic
resistor 4028a placed at a rear port 4024a with the cavity between
resistor 4028a and the rear of a transducer 4025a. This cavity
4026a increases the sound sensing mechanism to restore the
relationship needed to achieve the desired polar properties.
[0252] A high acoustic resistance cover 4030a similar to 4028a is
placed over external forward port 4022a to severely damp the ports.
This very high damping of both ports lowers the acoustic
sensitivity and the airflow noise. Since the dominant vibration to
microphone output conversion comes from the vibration of the
microphone against static air, vibration noise is also reduced. The
end result is a normal acoustic sensitivity with profoundly lower
airflow and vibration noise. Since this is a fundamental
improvement in signal to airflow noise, it applies to all airflow
coming from any direction. The resulting long "D" of the microphone
assembly is positioned along the rear surface of the mirror housing
along a diagonal with the lower portion angled inward toward the
middle and the upper portion angled outward. This results in a high
degree of noise rejection for sounds coming from below the mirror
along the windshield boundary and a great deal of difference for
sounds coming at an angle to the common axis between the two due to
the angular positioning.
[0253] The use of the hyper-long "D" improves greatly the
directional properties for the majority of the passed band. Higher
frequencies are not necessarily helped. This deficit is addressed
by adding directional means such as a partially horned (or flared)
opening 4032a, 4032b toward the forward port 4022a, 4022b (FIG.
53B-53D). While frequency response may be negatively impacted, this
aspect may be corrected by electronic equalization ideally done
prior to digitization.
[0254] By providing similar constructions on opposite sides of the
rearview mirror housing 4006, two signals may be obtained that are
free from airflow noise and that reject the dominant spatial noise
location, each with a very high degree of directionality and each
aimed to provide a great degree of spectral difference related to
source angular position. Since microphone assemblies 4020a and
4020b are widely spaced, there is also a significant arrival time
difference. These "artificial ears" produce all the data types and
freedom from unwanted airflow and mechanical noise needed for the
companion DSP algorithms. The DSP algorithms may thus exploit the
additional data and enjoy the freedom from non-acoustic noise
content.
[0255] The preferred default for most applications is to have the
DSP provide no signal until speech is detected. Thus, the preferred
process is based on not passing a signal unless the speech
detection criteria are met as opposed to always passing the signal
and trying to lower the noise content. While this may provide
processing delays, compensation can be accomplished by providing a
slight delay in the delivered signal to allow processing and yet
not use the first utterance of a spoken word. The process begins
with the determination in each ear channel that a change in the
input has occurred consistent with a speech utterance. This is a
well-established DSP process. The difference in the case of the
present invention is that this action is done in two channels by
only passing speech-like events. The present invention avoids times
when speech content is so low that it is virtually useless. The
threshold may be set higher for more robust vocal recognition and
better speech quality or may be set lower for higher noise to
speech situations. The time of arrival may then be utilized to
begin the process of processing only that speech from the desired
spatial location (i.e., the location in which the driver or other
passengers are located). Incorrect arrival time difference will
narrow the possibly conflicting noises to those arriving source
locations around the line connecting the two transducer's center
lines. Then, by applying the DSP's stored knowledge of the desired
user, and of human speech in general, the user's fundamental
frequencies may be determined to create a comb pass filter. The
result is that only those bands likely to contain speech are
present. This is most effective in the bands dominated by vocal
cord sounds.
[0256] At this point, any bands are passed that are likely to
contain speech and only those sounds from the correct location and
only those sounds that vary like speech are passed. The relative
spectral content may then be used to further add location
separation. For every spatial location, one can map the relative
frequency responses for that entry angle for both signals. One
would only need to address those regions where speech bands are
present. By comparing the difference in spectral content of the two
signals from microphone assemblies 4020a and 4020b, the DSP can
determine if the current dominant signals are coming from the
focused location. Even more useful, the DSP can determine if a time
varying band in the passed bands originates from other than the
focused location. This is achieved by comparing the relative
magnitudes to the response maps. For example, if the difference
should be +3 dB left versus right, and the difference is -2 dB, the
DSP will know that this particular band did not originate at the
focused spatial location and can be removed. At this point, only
speech sounds from the desired location have been passed. At this
time, the DSP's knowledge of the target user may be used to
reconstruct missing speech bands. Specifically, there will be bands
where there is important speech content, but the speech content is
not large enough to be significant and will be lost in the
filtering process. Humans know what a speaking person sounds like
from less noisy times and apply that knowledge during very high
noise conditions to extrapolate the speech bands. The DSP may use
the same form of processing. Specifically, over time, the DSP may
generate a harmonic amplitude map for the range of observed
fundamental frequencies. If the fundamental frequency is known, it
may be used as the map reference and extract the relative
magnitudes of the harmonics. Since every human has a consistent
harmonic map, as the result of fixed head cavities, the DSP can
apply the known harmonic amplitudes to estimate the missing ones.
For example, human speech usually loses its high frequency content
in very high noise environments. In lesser noise, where some of the
high bands are not lost, knowledge may be gained of this speech and
used to fill in the missing bands in the higher noise
environments.
[0257] The sequence of filters and the number of filters used can
vary depending on need, benefit, or cost. The key being to exploit
the rich data derived from the artificial ears and the knowledge of
the speaking human to yield speech free from the detrimental
effects of high noise. With reference to FIGS. 53A-53E, it is noted
that the two microphone assemblies 4020a and 4020b are integrated
into the rear of the mirror housing 4006 and are disposed such that
the central axes of the transducers provided in these assemblies
are at an angle with respect to one another and with respect to a
normal to the mirror surface. Further, the transducer central axes
are aimed at an angle upward relative to the position of the
driver. This allows the microphone assemblies to be integrated more
to the rear of the mirror assembly and somewhat obscured from the
view of the driver or other passengers.
[0258] While the above embodiment addresses the problems in the
automotive environment on a broadband basis, the transducers used
may be omni-directional and the DSP could utilize time of arrival
for the lower frequency bands while using the directional
characteristics provided by the horn at the forward port for the
higher frequency bands.
[0259] A preferred second embodiment of an interior rearview mirror
assembly 4101 of the present invention is shown in FIGS. 54A-54D.
The front view of interior rearview mirror assembly 4101 is not
shown insofar as its appearance would be similar to the interior
rearview mirror assembly 4001 shown in FIG. 53A. As shown in FIGS.
54A-54D and as described below, the microphone assemblies 4120a and
4120b are also mounted on the back surface 4107 of the mirror
housing 4106 and are not visible from the front of the mirror
assembly.
[0260] The microphone assemblies 4120a and 4120b are preferably
mounted on the mirror assembly and may be substantially identical.
Only one of the two microphone assemblies is shown and described in
detail. Microphone assembly 4120a includes a microphone housing
4115, a transducer 4125, and a circuit board 4126. The microphone
housing 4115 (FIGS. 55-57) is generally rectangular, although the
housing could have a generally square foot print, an elongated
elliptical or rectangular foot print, or any other shape desired by
the microphone designer. The microphone housing 4115 includes front
ports 4116 that face upwards and rear ports 4118 that downward. The
ports 4116 and 4118 provide sound passages through the microphone
housing. The ports 4116, 4118 can have any suitable opening shape
or size. In the embodiment shown in FIGS. 54A-62, microphone
housing 4115 includes four front ports 4116a-4116d provided in the
front surface (i.e., the side of the housing facing upward) of
microphone housing 4115, and four rear ports 4118a-4118d in the
rear surface (i.e., the side of the housing facing downward) of
microphone housing 4115. The front and rear ports are similar in
shape and position and are preferably symmetrical.
[0261] The microphone housing 4115 also includes resilient mounting
tabs 4140 for insertion into openings (not shown) in the back
surface of mirror housing 4106 to thereby secure microphone
assembly 4120a to mirror housing 4106. For example, the tabs can be
generally L-shaped in profile for insertion into the mirror housing
4106. Alternately, the tabs 4140 can be elongate snap connectors
that slide into an opening (not shown) in the back surface of the
mirror housing and snap into engagement with the inside surface of
the mirror housing 4106 after full insertion. The microphone
housing 4115 can be integrally molded plastic, stamped metal, or of
any other suitable manufacture.
[0262] The transducers 4125 used in the microphone assemblies 4120a
and 4120b are preferably substantially identical. The transducers
4125 can be any suitable, conventional transducers, such as
electret, piezoelectric, or condenser transducers. The transducers
may be, for example, electret transducers such as those
commercially available from Matsushita of America (doing business
as Panasonic), and may advantageously be unidirectional
transducers. If electret transducers are employed, the transducers
can be suitably conditioned to better maintain transducer
performance over the life of the microphone assemblies. For
example, the diaphragms of the transducers 4125 can be baked prior
to assembly into the transducers.
[0263] The circuit board 4126 has a conductive layer on one of its
surfaces that is etched and electrically connected to the leads of
transducer 4125. The transducer leads may be connected to a
pre-processing circuit that may be mounted to the conductive layer
of circuit board 4126. Although the pre-preprocessing circuit can
be mounted on the circuit board 4126 in the microphone housing, it
will be recognized that the such a circuit as well as other
circuits such as a digital signal processor (DSP) can alternatively
be mounted on a printed circuit board 4127 (FIG. 54D) in the mirror
housing 4106, and further that in the case of an electrooptic
mirror, such as an electrochromic mirror 4108, the circuits can be
mounted on a common circuit board with the mirror electrical
components, or the circuits and the mirror electrical components
can be mounted on separate circuit boards within the mirror housing
4106. Further still, such processing circuits may be located
elsewhere in the vehicle, such as in the mirror assembly mount, an
overhead console, an on-window console, an A-pillar, or in other
locations. Examples of such processing and pre-processing circuits
are disclosed in commonly assigned U.S. Pat. No. 6,882,734.
[0264] The electrical connection of the transducer leads and the
components of an pre-processing or other processing circuit, are
preferably by electrical traces in the conductive layer of the
circuit board, formed by conventional means such as etching, and
vias extending through the dielectric substrate of the printed
circuit board. The circuit board may include holes for receipt of
posts on microphone housing 4115. Such posts may be heat-staked to
the circuit board substrate after the posts are inserted through
the holes therein to secure the connection of the circuit board
4126 to the microphone housing 4115 and insure that the microphone
assembly provides acoustically isolated sound channels between the
transducer 4125 and the ports 4116 and 4118, as described in
greater detail herein below.
[0265] To assemble the microphone assembly 4120a, the transducer
4125 is first mounted on the circuit board 4126. As will be
described in detail below, an acoustic dam 4130 (FIGS. 58-62) is
preferably inserted between the circuit board 4126 and microphone
housing 4115. The transducer 4125, circuit board 4126, is then
secured to the microphone housing 4115 with the acoustic dam 4130
therebetween.
[0266] Microphone transducer 4125 is preferably mounted sideways
through a hole 4134 formed in printed circuit board 4126. A portion
of transducer 4125 would thus extend below the bottom surface of
circuit board 4126 and a portion would also extend above a top
surface of printed circuit board 4126. Mounting the transducer in
this orientation and position relative to the circuit board
provides several advantages. First, the electrical contacts on the
transducers may be directly soldered to traces on the printed
circuit board. This avoids the need for manually connecting wires
to the transducer contacts and subsequently manually connecting
those wires to the circuit board. Thus, the transducer may be
mounted to the circuit board using conventional circuit board
populating devices.
[0267] Another advantage of mounting the transducers such that they
extend above and below the surfaces of the printed circuit board is
that one side of the circuit board may include a conductive layer
serving as a ground plane. Such a ground plane may shield the
transducers from electromagnetic interference (EMI) that may be
produced by other components within the rearview mirror assembly or
in other components within the vehicle. Such EMI can introduce
significant noise into the signal delivered by the transducers. In
a preferred embodiment, each transducer is mounted in a circuit
board having a conductive ground plane facing the acoustically
active portion of the transducer while the circuit components are
mounted to the opposite side.
[0268] Microphone subassembly 4020 further includes a windscreen
4042, which protects the transducer and circuit board from the
external environment. Windscreen 4042 is preferably made of a
hydrophobic heat-sensitive adhesive-coated fabric and is adhesively
attached to the underside and inner surfaces microphone housing
4115 across ports 4116 and 4118. Microphone housing 4115 is
preferably tightly bonded about circuit board 4126 to provide a
water-impervious enclosure for transducer 4125.
[0269] While it has been typical in conventional microphones to
minimize the acoustic resistivity of a windscreen by increasing the
porosity of the windscreen, the microphone assembly of the present
invention advantageously utilizes a windscreen with a higher
acoustic resistivity by decreasing the porosity of windscreen and
yet obtaining not only better water-resistant properties, but to
also improved the acoustic characteristics for the microphone
assembly. The use of a high resistively windscreen is particularly
advantageous when the microphone assembly is mounted on a rearview
mirror assembly since significant noise may be introduced from the
windshield defroster. Specifically, the acoustic resistivity of
windscreen 4142 may be increased to at least about 1 acoustic
.OMEGA./cm.sub.2 and preferably has an acoustic resistivity of at
least about 2 acoustic .OMEGA./cm.sub.2, and more preferably has an
acoustic resistivity of at least about 8 to 9 acoustic
.OMEGA./cm.sub.2. Further, as described below, the acoustic
resistivity of windscreen 4142 may be varied to also vary the
directionality and polarity of the microphone assembly.
[0270] With the microphone transducers 4125 of the two microphone
assemblies 4120a and 4120b sealed in separate housings and having
their own windscreens, the ports and acoustic resistivity of the
windscreens may be different for the different microphone
assemblies transducers so as to compensate for any effects
experienced by the transducers as a result of the positioning of
the transducers on the vehicle accessory. For example, when one
microphone assembly (i.e., 4120a) is to be positioned closer to the
windshield as a result of typical tilting of mirror housing 4106,
its polar pattern may be slightly different from that of the other
microphone assembly 4120b. Thus, by selecting an appropriate
microphone housing design/port configuration, and windscreen
resistivity, the effects of the differences resulting from the
positioning of the transducers of the two assemblies may be
compensated such that the transducers exhibit substantially similar
polar patterns and other characteristics. While the windscreen has
been described above as consisting of a hydrophobic fabric, it will
be appreciated that the windscreen may be molded integrally across
the ports of the microphone housing. Such an arrangement would
simplify the manufacturing of the microphone assembly by requiring
less parts and less manufacturing steps. Further, it would more
likely provide a more effective seal between the windscreen and the
microphone housing. To attempt to obtain a required sensitivity
accuracy for the transducers, a laser trim tab may be added to the
gain stage connected to each transducer. The transducers may then
be acoustically excited by a calibrated sound source and the output
of the transducers is monitored. The laser trim tab is then trimmed
to precisely set the gain and thereby obtain precise sensitivity
accuracy.
[0271] A problem with mounting microphone subassemblies to the top
or back of the mirror housing results from the fact that the
microphone assemblies are closer to the windshield. When the
windshield defroster is activated, a sheet of air travels upward
along the windshield. Thus, when the microphone subassemblies are
placed on the back or top of the mirror housing, it is exposed to
more airflow as the air from the defroster passes between the
mirror housing and the window past the microphone subassembly. This
airflow creates turbulence as it passes over the microphone
subassembly, which creates a significant amount of noise.
[0272] To solve this problem when the microphone assembly is
mounted to the top of a rearview mirror housing, commonly-assigned
U.S. Pat. No. 6,882,734 and PCT Application Publication No. WO
01/37519 A2 disclose the use of a deflector that extends upward
from the rear of the mirror housing so as to smoothly deflect the
airflow from the defroster over and/or beside microphone
subassembly so that it does not impact the transducers or create
any turbulence as it passes over and around the microphone
assembly. Because the airflow primarily would enter the rear of the
microphone subassembly, the deflectors are designed to redirect the
air with minimal impact on the frequency response of the microphone
subassembly. This is important for high intelligibility in the
motor vehicle environment. With no direct air impact and the
avoidance of turbulence near the microphone subassembly, the
microphone assembly may advantageously be mounted on the top of the
mirror housing can offer superior resistance to airflow-generated
noise.
[0273] As described above, the microphone assemblies used in the
second embodiment have rear ports 4118 that open downward. Normally
this would pose a serious problem since the defroster airflow would
directly strike these rear ports and thus generate significant
noise. As described above and in detail below, the use of a very
high acoustic resistivity windscreen 4142 significantly reduces the
noise level caused by such airflow. Nevertheless, it is also
advantageous to configure the back surface 4107 of mirror housing
4106 such that integral deflectors 4146a and 4146b are provided
proximate the lower (rear) surface of the microphone assemblies
4120a and 4120b so as to reduce or prevent such airflow from
directly striking the microphone assemblies. Deflectors 4146a and
4146b may be provided by mounting the microphone assemblies 4120a
and 4120b in recessed portions 4148a and 4148b of back surface 4107
of mirror housing 4106. Such recesses would each include a wall
4150a and 4150b that has a height that is generally equal to or
greater than the height of microphone housing 4115. Wall 4150a and
4150b may be flat, tapered and/or contoured around the lower wall
and one end of microphone housing 4115. Preferably, recesses 4148a
and 4148b would not have any walls opposite walls 4150a and 4150b
such that front ports 4116 of microphone housing 4115 open upwards
and are unobstructed by any other structure on the mirror
housing.
[0274] As shown in the drawings, the combination of the mounting of
the microphone assemblies 4120a and 4120b on the back surface 4107
of mirror housing 4106 and the mounting of the transducers 4125 in
microphone housings 4115 in the orientation shown, results in the
central axis of the transducers 4125 extending generally
vertically. Normal tilting of the mirror housing 4106 may result in
the central axis tilted nearly parallel to the windshield 105 of
the vehicle 100. As described below, such positioning and
orientation of the transducers 4125 results in several
advantages.
[0275] One advantage is that the transducers may be mounted in a
through-hole fashion as discussed above, while the circuit board
4126 may be mounted substantially parallel to a surface of mirror
4108, most likely the rear surface of mirror 4108. If the mirror
4108 is an electro-optic mirror, a mother board 4127 is typically
provided in the mirror housing 4106 parallel to the rear surface of
mirror 4108 (see FIG. 54D). Thus, circuit board 4126 would be
substantially parallel to mother board 4127 and could be readily
electrically coupled to mother board 4127 using conventional
connector plugs 4128. Additionally, by having both circuit boards
in parallel with one another and in parallel with the rear surface
of mirror 4108, less space need be provided in mirror housing 4106
to accommodate these components.
[0276] By being located on the back 4107 of the mirror housing 4106
and in the basic plane of mirror 4108, the virtual axis of
microphone assembly 4120a is aimed upward. This in turn means the
rear aligns with the windshield boundary, since noise is greatest
along this boundary this alignment offers the greatest average
noise rejection. In other words this aiming condition lowers the
noise from common sources including defroster fan and duct noise,
road noise, rain noise, and wind-on-windshield noise.
[0277] Another benefit of upward aiming and being just behind the
forward surface of mirror housing 4115 is freedom from the comb
effect. All high frequency sounds from the vehicle cabin enter the
microphone by reflecting off the windshield and/or headliner at
which microphone assemblies 4120a and 4120b are aimed (See FIG.
54C). Direct high frequency sound is stopped by the front surface
of the mirror. Since comb effect occurs when two identical signals
mix with different arrival times, the removal of the direct sound
lowers the high frequency signals to only one. Comb effect is a
very disruptive condition as it removes almost completely regions
of the band. These missing regions contain important sonic data
that when lost impairs voice recognition and it makes human speech
sound hollow and reverberant.
[0278] Another benefit of mounting the microphone assemblies 4120a
and 4120b to the rear surface and the ends of mirror housing 4106
is that the mirror physically blocks high frequency sounds coming
from directions in which the sound path would pass through the
mirror 4108 to reach the microphone assembly 4120a, 4120b. Placing
the microphone assemblies 4120a, 4120b into notched recesses 4148a,
4148b at each end the mirror housing 4106 imparts a great deal of
increased directional ability in the higher frequency portion of
the band roughly 2500 Hz and above.
[0279] Prior art array microphones are based on the use of the same
basic directional aspect. This is often a requirement such that
when time of arrival is adjusted the speech signals will add.
Differences in aiming angle would interfere with this addition for
sounds coming off axis which the array aiming ability requires. In
prior art arrays, the microphone transducers must be placed close
enough together to achieve time alignment by simple maximization of
signal. This prevents wide spacing relative to the wavelength of
the highest frequency sound in the pass band. Conversely, a wide
enough separation is often required to have a difference of
meaningful size when the wavelength decreases. The combination of
these two effects forces the use of more the two transducers to get
effective array microphone operation trough the entire audio
band.
[0280] The inventive second embodiment manifests directional
attributes that are nearly identical through the mid and lower
frequencies. Since the microphone assemblies are also widely
separated they function effectively in the mid and lower
frequencies. Since the microphone assemblies are too widely
separated for conventional array functionality in the higher
frequencies, it may be desirable to add an additional means to
achieve beam steering.
[0281] The mirror-added directional ability provides this second
beam forming means. Since there is very significant difference in
the high frequency output for a given point in space, amplitude
comparison can be used to augment time of arrival extending beam
forming beyond that supported by the spacing. In effect, the two
microphone transducers yield the functionality of four.
[0282] In terms of spacing dimensions, advantages arise as the
spacing is increased from 3 inches, the greatest spacing possible
for a top frequency of roughly 5 kHz, with increasing advantage
until the spacing increases beyond the point of high frequency
directional onset, 2.5 kHz or 6 inches. The onset of high frequency
directional function is progressive so the actual range of maximum
separation can vary from 5 to 7 inches.
[0283] Another advantageous aspect is the styling freedom offered
by recessed rear mounting. The microphone is not visible from the
cab and lies on a large surface supporting a large microphone with
no additional size or protrusions. Further by avoiding the central
area of the mirror housing, the microphone assemblies do not
interfere with mounting or wiring. Since the center of the mirror
housing must be strong to resist vibration avoidance of the center
also preserves the ability to effectively use reinforcing ribs and
other strength enhancing details.
[0284] To gain the full benefits from mounting the microphone
assemblies on the rear surface of the mirror housing, acoustic dam
4130 may be employed. This is because the microphone assemblies are
preferably recessed into the mirror housing to protect the rear
ports 4118 from direct air impingement from defroster airflow. This
situation in turn decreases the effective "D" by adding delay to
the sound arriving at the rear port. Sound traveling from the rear
to the front takes some additional time to reach the rear port. The
difference between the arrival at the front or rear forms the phase
difference that produces the null. This difference is reduced
altering the resulting null angle. In other words, the microphone
should be made more directional to have the correct manifested
directional properties when recessed into the rear of the mirror
housing.
[0285] All prior art assumes the transducer portion is free within
the containment shell. In other words, sound waves are passing by
the transducer and it responds to them as passing waves. In the
case of "D" extension a greater portion of the available pressure
difference from these passing waves is yielded. This is used to
compensate for reduced wave intensity due to the impact of the port
resistance of the outer shell.
[0286] The acoustic dam 4130 does not function like a "D" extender
but rather forms additional pressure difference effectively
funneling acoustic energy into the transducer. This is accomplished
by dividing the interior space into acoustic zones. It is the
difference in external "D" between these regions that causes the
increased pressure difference.
[0287] For example, in the second embodiment, dam 4130 forms
defined regions of the outer grill that impact specific zones since
transducer 4125 passes through a hole in dam 4130, one side is the
virtual front of the transducer and the other the back. The center
of the port area feeding each zone acts like the port of a
microphone of that size. In other words, if these ports on the
average are 1 inch apart then the microphone has an effective
external "D" of 1 inch. The virtual aiming direction is also
determined by this center of area location. In other words, if the
microphone is aligned straight forward but the two virtual ports
are rotated 45 degrees, then the aiming point for the actual
microphone will be rotated 45 degrees.
[0288] In the second embodiment shown in FIGS. 58-62, dam 4130 does
not completely separate the zones it forms when dividing the
acoustic chamber defined by the housing 4115 and circuit board
4126. This allows the pressure to equalize between the zones.
Flowing air creates different pressures in each zone by connecting
these zones this pressure difference is reduced. This connection
does not impact acoustic pressure differences because they are the
result of a consistent external pressure difference and because
flowing air noise is a near DC phenomena. The length of the dam
4130 in relation to the length of the open acoustic chamber area
determines the weighting factor of the external ports 4116 and
4118. Those ports near or over the open regions 4132 have little
impact on the virtual "D" and conversely those farthest from the
openings 4132 have the greatest impact. As a result dam width can
be used to tune the design to optimize the desired directional
aspects and the flowing air rejection. The best ratios are from 50%
dam to 90% dam.
[0289] The dam extends and acoustically seals the typically narrow
thickness dimension of the microphone housing 4115. Any gap close
to the center of the dammed zones allows the pressure difference to
cross equalize thereby lowering the difference the transducer
perceives. Thus, as shown in FIGS. 58-60 and 62, groove forming
members 4150 may be provided on the top of circuit board 4126 so as
to receive and hold acoustic dam 4130 and thereby ensure an
acoustic seal between the dam and the circuit board. Similarly, a
groove 4152 may be provided on the top inner surface of microphone
housing 4115 to receive, hold and create an acoustic seal with dam
4130. To enhance the acoustic seal of dam 4130 about the periphery
of transducer 4125, an epoxy 4154 may be applied therebetween.
[0290] To demonstrate the effectiveness of acoustic dam 4130 in
combination with a very high acoustic resistivity windscreen 4142,
a three different prototype rearview mirror assemblies were
constructed having microphone assemblies similar to the second
embodiment described above and shown in FIGS. 54A-62. The first
prototype used a windscreen with a low acoustic resistivity fabric.
A polar plot was obtained for this first prototype at 1000 Hz. A
copy of the polar plot is shown in FIG. 63. Note the directional
sensitivity of this polar pattern is generally what is desired so
as to have greatest sensitivity upward (i.e., at 0 degrees) and the
lowest sensitivity downward towards the defroster. Unfortunately,
despite the low sensitivity downward, this first prototype remains
relatively sensitive to noise caused by the laminar airflow from
the defroster that travels up the windshield of the vehicle due in
part to the fact that the rear ports of the microphone assembly
open towards this airflow. A plot of the sensitivity of the first
prototype to such noise over a frequency band of 20 Hz to 1000 Hz
is shown as plot A in FIG. 64.
[0291] To reduce the sensitivity of the microphone assembly to the
defroster airflow, a windscreen having a very high acoustic
resistivity was used in the second prototype. As shown in plot B of
FIG. 64, the use of the very high acoustic resistivity windscreen
significantly reduced the sensitivity of the microphone to the
defroster airflow noise. However, very high acoustic resistivity
windscreen adversely made the microphone sensitivity much less
directional as shown in the polar plot of FIG. 65.
[0292] A third prototype was then constructed and tested whereby
acoustic dam 4130 was added to the prototype having the very high
acoustic resistivity windscreen. Surprisingly, as shown in the
polar plot of FIG. 66, the directional sensitivity of the
microphone assembly was recovered despite the use of the very high
acoustic resistivity windscreen. Therefore, the benefit of noise
rejection provided by the windscreen may be exploited without any
loss in directionality.
[0293] There are several useful variations on the basic dam. These
include forming more than two zones supporting more than one
transducer in a single outer housing. Since these zones are
acoustically as separate as the outer ports, a single housing can
hold multiple transducers to gain the advantage of a large
nitration volume and yet have each transducer act as if it were in
its own separate housing as far as acoustic directional properties
are concerned. It will be appreciated that acoustic dam 4130 could
be integral extension of circuit board or microphone housing rather
than a separate element. The acoustic dam design frees the designer
from the trade-offs of the prior art. Large housings can be used
yet act as several smaller ones. Transducers can be aimed
internally different from the aiming direction of the external
microphone without using ducts that impair higher frequency
performance.
[0294] One aspect derived from this design is the ability to create
highly directional microphones with directional attributes that do
not vary with frequency to the degree prior art microphones do. In
a typical microphone design, in order to increase the directional
aspect from omni-directional through all possibilities to
bi-directional, the transducer's internal damping must be lowered.
The assignees prior "D" extender designs modestly improved this
relationship by adding additional directional pressures. The new
acoustic dam allows very directional microphones with very high
damping factors. In other words, the acoustic resistance is so high
that it swamps out the other variables that cause directional
parameters to change with frequency.
[0295] To demonstrate the ability of the inventive structure to
maintain directional parameters over frequency bands of interest,
the third prototype rearview mirror described above (i.e., having
the very high acoustic resistance windscreen and the acoustic dam)
was placed in a test chamber with the mirror glass face up and in a
horizontal plane. The mirror assembly as rotated about a vertical
axis extending perpendicularly through the middle of the mirror
glass and polar plots were obtained at various frequencies for the
driver-side microphone assembly. These polar plots are shown in
FIGS. 67-69 in which the 0 degree axis corresponds to the top of
the mirror assembly and the 180 degree axis corresponds to the
bottom of the mirror assembly. FIG. 67 shows a polar plot at 250
Hz, whereas FIG. 68 shows various plots between 300 Hz and 2 kHz
and FIG. 69 shows various plots taken between 3 kHz and 6 kHz. In
viewing these plots it will be noted that the plots do not vary
significantly from frequency to frequency and that the greatest
sensitivity is at about 30 degrees. This ideal insofar as the
mirror assembly is typically rotated and tilted relative to the
driver so as to have maximum sensitivity to receive sound waves
from the driver that reflect off of the windshield and/or headliner
of the vehicle.
[0296] To test the directionality in a different plane, the same
mirror assembly was placed in the test chamber on end such that the
axis about which the mirror assembly is rotated being parallel to
the axis of the mirror extended along its longest dimension. In
these resulting polar plots, the zero degree axis corresponds to an
axis that is perpendicular to, and extends in front of, the mirror
glass. FIG. 70 shows various polar plots between 300 Hz and 1 kHz,
whereas FIG. 71 shows various plots between 3 kHz and 6 kHz and
FIG. 18 shows various plots taken between 6.5 kHz and 8 kHz. In
these plots, it will again be noted that there is little various in
directional sensitivity as a function of frequency. Also, the
maximum sensitivity for frequencies between 3 kHz and 8 kHz occurs
at about 120 degrees, which would be ideally corresponding to an
upwards direction when the mirror assembly is tilted downward by
the driver in a normal viewing position.
[0297] Finally, by allowing port areas to set aiming direction and
effective "D", the dam design supports the use of any housing
styling (even non-symmetrical designs) since port area can be
symmetrical and the acoustic dam can form regions of the correct
volume. In this case, the advantage takes the form of freedom of
physical design, and thus housings that are larger and have more
complex exterior shapes can be used.
[0298] As suggested above, the above described acoustic dam (or at
least a modified version thereof) may be implemented in a
microphone assembly having a microphone housing in which two or
more transducers are mounted. An example of such a microphone
assembly 4220 is shown in FIGS. 73 and 74. This particular
embodiment of the inventive microphone assembly is intended for
mounting to the top surface of the mirror housing in a manner
similar to that disclosed in commonly-assigned U.S. Pat. No.
6,882,734. It will be appreciated, however, that the depicted
microphone assembly 4220 could be mounted elsewhere.
[0299] Microphone housing 4215 preferably includes front ports 4216
and rear ports 4218, and may include top ports 4217. In the example
shown in FIGS. 73 and 74, the two centermost rear ports 4218 are
closed as are two of top ports 4217 (the closed ports are shown in
dashed lines). As shown in FIG. 73, three zones 4240a, 4240b, and
4240c are formed within microphone housing 4215 by an acoustic dam
4230. First zone 4240a is common to the fronts of the transducers
4225a and 4225b. Second zone 4240b is formed to the rear of
transducer 4225a and third zone 4240c is formed to the rear of
transducer 4225b. These zones steer the aiming direction of the
first transducer 4225a to better aim at the driver yet keep the
transducers physically facing in opposite directions to gain the
advantages in air flow and vibration cancellation. Thus, more than
one transducer may be provided in a housing and more than two zones
may be formed to thereby provide the freedom to aim the sensitivity
of the microphone assembly based on the outer port locations rather
than physical orientation of the transducers.
[0300] The acoustic dam 4230 forces a definition of the ports 4216,
4217, and 4218 that will contribute to the front and those that
will add up to be the rear signal. There will be a center of area
that will act as the virtual location for the front and another for
the rear. A line through these two centers will form the aiming
axis independent of the transducer orientation. In this third
depicted embodiment, the rear zones are formed such that the
driver-facing microphone transducer 4225a will aim more into the
cab and the rear of the passenger-facing transducer 4225b aims more
away from the cab. This is accomplished by changing to which side
of the acoustic dam 4230 the rear ports 4218 connect. In this third
embodiment, acoustic dam 4230 extends from the top of the circuit
board to the top inner surface of the microphone housing 4215 and
extends tightly around all of the upper peripheral edge of
transducer 4225a and tightly around half of the upper peripheral
edge of transducer 4225b. This provides for a tight acoustic seal
through the acoustic chamber defined by housing 4215 with the
exception of the openings at one side of transducer 4225b and at
the ends of dam 4230.
[0301] The inventive microphone construction is preferably located
on the back of a rearview mirror assembly housing. Nevertheless,
the certain aspects of the inventive microphone may be implemented
in microphones mounted at other locations on a mirror assembly,
including on the mirror assembly mounting structure, the top,
bottom, or sides of the mirror housing, as well as in any other
vehicle accessory such as a headliner, sun visor, overhead console,
A-pillar, or a console extending between the headliner and a mirror
assembly. For example, the above-described acoustic dam may be
employed in various microphone assemblies whether used in vehicle
applications or any other non-vehicle applications. Additionally,
the mounting of two microphone assemblies in a vehicle such that
the microphones are spaced apart with an acoustic barrier
therebetween like the portion of the rear surface of the mirror
housing. For example, if an overhead console is provided, the
microphone assemblies could be recessed into the console and spaced
apart in a manner similar to the disclosed mirror assembly
implementation. Further, if the overhead console protrudes downward
from the headliner, the microphone assemblies could be mounted on
either side of the console with or without recesses.
[0302] The above description is considered that of the preferred
embodiments only. Modifications of the invention will occur to
those skilled in the art and to those who male or use the
invention. Therefore, it is understood that the embodiments shown
in the drawings and described above are merely for illustrative
purposes and not intended to limit the scope of the invention,
which is defined by the following claims as interpreted according
to the principles of patent law, including the doctrine of
equivalents.
* * * * *