US005982906A
`5,982,906
`(114) Patent Number:
`United States Patent 55
`Ono
`[45] Date of Patent:
`Nov. 9, 1999
`
`
`[54] NOISE SUPPRESSING TRANSMITTER AND
`NOISE SUPPRESSING METHOD
`
`Primary Examiner—Vivian Chang
`Attorney, Agent, or Firm—Laff, Whitesel & Saret, Ltd; J.
`Warren Whitesel
`
`[75]
`
`Inventor: Yoshihiro Ono, Tokyo, Japan
`
`[57]
`
`ABSTRACT
`
`[73] Assignee: NEC Corporation, Tokyo, Japan
`
`[21] Appl. No.: 08/974,659
`
`Signals outputted from microphones 111 and 112 are input-
`ted throughfilters 118 and 119 to noise suppressing sections
`A and B. One having smaller poweris selected from output
`signals $107a and S107b by switching a selection switch
`Nov. 19, 1997
`Filed:
`[22]
`SW in a signal power comparing section 117, to which the
`Foreign Application Priority Data
`[30]
`output signals S$107a and S107b have been inputted, and a
`transmission signal S108 level-adjusted by a level adjuster is
`Nov. 22, 1996
`[JP]
`Japan Season eee eeeeeeaeneeee see seeeee eens 8-312479
`—-—-©utputted. In the noise suppressing section A (similarly in
`[SL]
`Tints C15 ccccccsscssssssssnstvesessnstvetnesnee HO4B 15/00
`[52] U.S. CU. cecsscsssssssesnssesnetesnstee 381/94.2; 381/94.1_the section B), microphoneoutputsignals $102 and S103are
`[58] Field of Searchccc 381/66, 94.1, 94.2,
`Processed by an adder 1134 and a subtracter 14a, a signal
`381/93, 71.1, 94.3, 83: 379/410 406
`$105a outputted from the subtracter 14a is inputted to a
`,
`,
`oo
`,
`subtracter 116a@ through a filter 115¢@ having a transmission
`References Cited
`function Fa(z) decided by the positions of the microphones
`111 and 112 and a noise suppression signal $107a obtained
`by subtracting the signal $105a from a signal S104out-
` Putted from the adder 1132 is outputted.
`
`[56]
`
`U.S. PATENT DOCUMENTS
`4/1996 Hirano vcecccsccssssssscssssseseereereertene 38166
`5,513,265
`
`7/1997 Ikeda..........
`. 381/94.1
`5,644,641
`5,694,474 12/1997 Ngo et ab. wien 381/66
`
`22 Claims, 8 Drawing Sheets
`
`y A NOISE SUPPRESSING SECTION
`
` Wa(z)sio1a
`
`SIGNAL
`POWER
`COMPARING
`SECTION
`
`“B NOISE SUPPRESSING SECTION
`
`1
`
`APPLE 1014
`
`APPLE 1014
`
`1
`
`

`

`TWNODIS
`
`qAMOd
`
`ONWdWOOD
`
`NOLLOAS
`
`
`
`ONISSSUddNSASIONV
`
`ISid
`
`
`
`NOLLOASSNISSAYddNS
`
`U.S. Patent
`
`Nov.9, 1999
`
`Sheet 1 of 8
`
`5,982,906
`
`80LS
`
`ASIONg\ NOILOAS
`
`2
`
`

`

`Nov. 9, 1999
`
`FIG.2 W(z)
`
`$107
`
`—N— w
`
`n
`
`U.S. Patent
`
`Sheet 2 of 8
`
`5,982,906
`
`3
`
`

`

`U.S. Patent
`
`Nov. 9, 1999
`
`Sheet 3 of 8
`
`5,982,906
`
`FIG. 3
`
`
`
`112:MICROPHONE(2)
`
`114:MICROPHONE(1)
`
`4
`
`

`

`U.S. Patent
`
`Nov. 9, 1999
`
`Sheet 4 of 8
`
`5,982,906
`
`FIG. 4
`
`
`
`|
`/
`!
`|
`SL
`
`8
`
`5
`
`

`

`U.S. Patent
`
`Nov. 9, 1999
`
`Sheet 5 of 8
`
`5,982,906
`
`FIG. 5
`
`SIGNAL
`POWER
`COMPARING
`
`A MOISE SUPPRESSING SECTION
`
`SECTION
`
`D NOISE SUPPRESSING SECTION
`
`6
`
`

`

`U.S. Patent
`
`Nov. 9, 1999
`
`Sheet 6 of 8
`
`5,982,906
`
`FIG.6
`
`W1(z)
`
`7
`
`

`

`U.S. Patent
`
`Nov. 9, 1999
`
`Sheet 7 of 8
`
`5,982,906
`
`FIG. 7
`
`112:MICROPHONE (2) 111:MICROPHONE(1)
`
`211:MICROPHONE(3)
`
`8
`
`

`

`U.S. Patent
`
`Nov. 9, 1999
`
`Sheet 8 of 8
`
`5,982,906
`
`FIG.8
`
`112
`
`d3
`
`d2
`
`<———__ wiz)
`
`FIG.9
`
`PRIOR ART
`
`9
`
`

`

`5,982,906
`
`1
`NOISE SUPPRESSING TRANSMITTER AND
`NOISE SUPPRESSING METHOD
`
`BACKGROUND OF THE INVENTION
`
`invention relates to a noise suppressing
`The present
`transmitter for outputting transmission signals after ambient
`noises are suppressed, which is used for a television tele-
`phone system or a portable telephoneset.
`Conventionally, as a hands-free device for a television
`telephone system, a noise suppressing transmitter for out-
`putting transmission signals after ambient noises have been
`suppressed has been used. This noise suppressing transmit-
`ter has also been used for a portable telephone set. As an
`example of such a noise suppressing transmitter, one dis-
`closed in JP-A-110349/1987 is known.
`
`10
`
`15
`
`25
`
`2
`making thin of a portable telephone set or a cordless
`telephoneset, it is difficult to dispose the two microphones
`1 and 2 in the direction of a straight line from a transmission
`voice source by keeping a proper distance because of
`structural reasons.
`As described above, in the noise suppressing transmitter
`of the conventional example, distortion occurs in the output
`transmission signals. In addition, the two microphones can-
`not be disposed by keeping a proper distance in the direction
`of a straight line with respect to a transmission direction in
`the miniaturized and thin device. Consequently, it is impos-
`sible to obtain high-quality speech voices in which noises
`are surely suppressed. In other words, the noise suppressing
`transmitter of the conventional example is disadvantageous
`in that it cannot be installed easily in the miniaturized and
`thin device.
`The present invention was made in order to solve the
`FIG. 9 is a circuit diagram showing main constituting
`above-described problems-inherent in the prior art. A pur-
`portions of a conventional noise suppressing transmitter. In
`pose of the invention to provide a noise suppressing
`the conventional example of FIG. 9, two microphones1 and
`transmitter, which can prevent the occurrence of distortion
`20
`
`2 are disposed onastraight line identical to the transmitting in output transmission signals, provide high-quality speech
`voices with noises surely suppressed and beeasily installed
`direction of a person M.A transmission (output) signal S11
`in a miniaturized and thin device.
`of the microphone 1 is inputted to a subtracter 3, and a
`transmission (output) signal S12 of the microphone 2 is
`phase-inverted and then inputted to the subtracter 3. The
`subtracter 3 extracts a transmission signal and outputs a
`transmission signal $13 obtained by suppressing ambient
`noises.
`
`SUMMARYOF THE INVENTION
`
`invention is made to solve the above-
`The present
`mentioned problems.
`The objective of the present invention is achieved by a
`noise suppressing transmitter for outputting voice signals by
`suppressing ambient noises, comprising: first and second
`microphones;first and second noise suppressing means for
`suppressing noises by performing addition and subtraction
`for output signals from said first and second microphones;
`power comparing means for comparing power between
`output signals from said first and second noise suppressing
`means; and selecting meansfor selecting, amongsaid output
`signals of said first and second noise suppressing means, an
`output signal having smaller power based on a signal from
`said power comparing means, and outputting said output
`signal.
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`This and other objects, features and advantages of the
`present invention will become more apparent upon a reading
`of the following detailed description and drawings,
`in
`which:
`
`FIG. 1 is a Block diagram showing a constitution of an
`embodimentof a noise suppressing transmitter of the present
`invention;
`FIG. 2 is a view illustrating a principle of noise suppres-
`sion of the first embodiment;
`FIG. 3 is a front view showing a constitution of a portable
`telephone set having a hand set disposed therein in thefirst
`embodiment, the hand set using two microphones;
`FIG. 4 is a view showing relationships between micro-
`phone positions and incoming voices in the first embodi-
`ment;
`FIG. 5 is a block diagram showing a constitution of a
`second embodiment;
`FIG. 6 is a view showing disposition of two microphones
`in the second embodiment;
`FIG. 7 is a front view showing a constitution of a portable
`telephone set having a hand set disposed therein in the
`second embodiment, the hand set using three microphones;
`FIG. 8 is a view showing position and distance relation-
`ships among the three microphonesandrelationships thereof
`with noise incoming directions in the second embodiment;
`and
`
`The following problemscan be pointed out for the above-
`noted conventional noise suppressing transmitter.
`(1) Distortion occurs in the output transmission signal S13
`because of addition of the two signals S11 and S12. For
`performing such addition based on digital signal processing,
`the following expression (1) is established, in which the
`transmission signal $11 of the microphone 1 is represented
`by S,(z) and the transmission signal $12 of the microphone
`S is represented by S,(z).
`Expression 1
`
`30
`
`35
`
`So(2)=51@)D@)
`
`@)
`
`40
`
`represents a transmission
`in the expression (1)
`D(z)
`function between the two microphones 1 and 2. Based on a
`difference between the transmission signals $11 and $12 of
`the two microphones1 and 2, S(z) of the output transmission
`signal S13 is represented by the following expression (3).
`
`[Expression 3]
`S(z) = S1(z) — S2@)
`= S1 (21 — DZ)
`
`3)
`
`From the expression (3), the output transmission signal
`S13 takes a value obtained by multiplying a signal at the
`time of collecting sounds only by one of the microphones 1
`and 2 by the transmission function of 1-D(z). A distance
`between the two microphones 1 and 2 takes an interval
`equivalent to one sampling cycle of digital signals. For
`example, if a sampling frequency is 8 kHz, the distance is
`4.35 cm based on the following expression (4), and the
`output transmission signal $13 is changed to a sound quality
`after having been passed through a high-pass filter (HPF).
`Expression 4
`
`45
`
`50
`
`55
`
`60
`
`1-D(2)=1-24
`
`(4)
`
`65
`
`(2) For the microphone of an interphone or the like
`hand-set or fixed in the wall following miniaturization and
`
`10
`
`10
`
`

`

`3
`FIG. 9 is a circuit diagram showing a main section
`constitution of a conventional noise suppressing transmitter.
`
`DESCRIPTION OF THE EMBODIMENTS
`
`4
`P(z) outputted from the adder 113 and a signal S105, Q(z)
`as a subtraction result outputted from the subtracter 114 are
`represented by the following expressions (7) and (8).
`
`5,982,906
`
`10
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`Next, the preferred embodiments of the noise suppressing
`transmitter of the present invention will be described in
`detail with reference to the accompanying drawings.
`FIG. 1 is a block diagram showinga constitutionofa first
`embodiment of the noise suppressing transmitter of the
`present invention.
`In FIG. 1, a transmitted voice $100 and ambient noises
`$101a and $1015 are inputted to microphones 111 and 112.
`Signals outputted from these microphones 111 and 112 are
`inputted to filters 118 and 119 for performing delaying
`respectively. Microphone output signals $102 and S103
`delayed by the filters 118 and 119 are inputted to noise
`suppressing sections A and B.
`Signals $107@ and $1076 outputted from the noise sup-
`pressing sections A and B are inputted to a signal power
`comparing section 117.
`In the signal power comparing
`section 117, a signal having smaller poweris selected from
`the output signals S$107a@ and S107b supplied to a selection
`switch SW and put througha level adjuster 120, and thereby
`a transmission signal $108 matching to a % level is output-
`ted.
`
`Next, the constitutions of the noise suppressing sections A
`and B will be described.
`
`The two noise suppressing sections are similarly con-
`structed. The delayed microphone output signals $102 and
`$103 are respectively inputted to an adder 113a@ (113B) and
`a subtracter 114a (1145)in the noise suppressing sections A
`and B. Asignal $105a (S105b) outputted from the subtracter
`114a (1145) is inputted to a filter 115@ (1155) having a
`transmission function Fa(z) decided by the positions of the
`microphones 111 and 112. In a subtracter 116a (116d), the
`output signal S106a (S106b) of the filter 115¢@ (115b)is
`subtracted from the output signal $104a (S104b) of the
`adder 113a@ (1135) and the result of this subtraction is
`outputted as a noise suppressed signal S107a (S1075).
`Next, the principle of noise suppression will be described
`by using numerical expressions.
`FIG. 2 is a view illustrating the principle of noise sup-
`pression of the first embodiment shown in FIG. 1.
`In FIG. 2, a transmitted voice signal and a noise signal are
`denoted respectively by codes S(z) and W(z). The two
`microphones are disposed at equal distances from a trans-
`mitted sound source, and it can be understood that trans-
`mitted voices reaching the microphones have identical
`phases. On the other hand, if noise signals come from a
`direction nearly vertical to the transmitted voices, phase
`differences occur among the noise signals reaching the
`microphones. This relationship is represented by a transmis-
`sion function D(z). Signals $102, A(z) and S103, Biz)
`outputted from the microphones can be represented by the
`following expressions (5) and (6).
`Expression 5
`
`[Expression 7]
`P(z) = A(z) + BY)
`= 2S(z) + W(z)(1 + D(z)
`
`[Expression 8]
`Q(z) = Biz) - A®)
`= Wiz)(1 — Di))
`
`1)
`
`(8)
`
`From these expressions (7) and (8), the following expres-
`sion (9) is obtained.
`
`[Expression 9]
`
`(9)
`
`S(z) = V2 [Piz)- WZ). + Di]
`
`= 1/2 [P@) - Q@(U + Diz) /d - D@)I
`
`Herein, since a transmission function for Q(z) corre-
`sponds to the filter 115, a transmission function F(z) is
`represented by the following expression (10).
`Expression 10
`
`F@)-0+D@))/A-D@)
`
`(10)
`
`From the foregoing description, it can be understood that
`a transmitted voice can be extracted by suppressing noise
`signals. For a noise signal W(z),
`that is, a noise signal
`entering from an opposite side with the incoming direction
`of a transmitted sound source as a symmetrical axis, noise
`suppression is performed by the constitution shown in FIG.
`1. Accordingly, by adjusting the transmission function F(z)
`of the filter 115, noises coming from optional directions can
`be suppressed. Next, the application of the microphones
`shown in FIGS. 1 and 2 to a portable telephone set having
`a hand set will be described.
`FIG. 3 is a front view showing the constitution of a
`portable telephone set,
`in which a hand set using two
`microphones is disposed, and FIG. 4 is a view showing a
`relationship between the positions of the microphones and
`incoming voices.
`In FIGS. 3 and 4, since noises come in from sufficiently
`far places for an interval between two microphones(1) 111
`and (2) 112, the incoming directions thereof can be under-
`stood to be identical to each other in the handset.
`Therefore, a noise signal inputted to the microphone (1)
`111 is delayed by an amount equivalent to a distance (d)
`compared with a noise signal inputted to the microphone (2)
`112. Assuming that a sampling frequency for signal pro-
`cessing is 8 kHz, by selecting a microphone position so as
`to satisfy d=42.5 mm, a delay is equal to a delay of one
`sampling cycle. This relationship is represented by the
`expression (11).
`Expression 11
`
`A(z)=S(2)+W(2)D(z)
`
`Expression 6
`
`B(2)=8(2)+W(2)
`
`(5)
`
`Db(z)=z+
`
`(11)
`
`Accordingly, a transmission function Fb(z) of the filter
`115b is represented by the following expression (12).
`
`(6)
`
`65
`
`Fb(2)=(142-2)/(1-24)
`
`(12)
`
`these two
`By performing addition or subtraction for
`output signals S102, A(z) and S103, B(z), a signal A104,
`
`Herein, since the transmission function Fb(z) may be
`unstable depending on inputs, the term z+ of the denomi-
`
`11
`
`11
`
`

`

`5,982,906
`
`5
`nator must be multiplied by a coefficient of O<a<1 in order
`to compensate for its instability. This relationship is repre-
`sented by the following expression (13).
`Expression 13
`
`Fb(z)=(1+2-2)/(1-az")
`
`(13)
`
`However, (a) must be approximated to 1 as much as
`possible.
`Similarly, a transmission function Fa(z) ofthe filter 115a
`is represented by the following expression (14).
`Expression 14
`
`Fb(z)=Fa(z)=(1+24)/(1-az+)
`
`(14)
`
`As a result, noises can be suppressed even if the noises
`come from any directions inclined by 0 degrees with respect
`to the incoming direction of transmitted voices. Further, if
`the transmission functions of the filters 115@ and 115b are
`
`not identical to each other, microphones having noise sup-
`pression characteristics asymmetrical between left and right
`sides when seen from the transmitted voice incoming direc-
`tion can be realized.
`
`10
`
`15
`
`20
`
`Expression 14
`
`Fb(z)=Fa(z)=(1+24)/(1-az+)
`
`(14)
`
`25
`
`As a result, noises can be suppressed even if the noises
`come from any directions inclined by 0 degrees with respect
`to the incoming direction of transmitted voices. Further, if
`the transmission functions of the filters 115@ and 115b are
`not identical to each other, microphones having noise sup-
`pression characteristics asymmetrical between left and right
`sides when seen from the transmitted voice incoming direc-
`tion can be realized.
`Next, the second embodimentwill be described.
`FIG. 5 is a block diagram showing the constitution of the
`second embodiment, and FIG. 6 is a view showing the
`disposition of the two microphones of the second embodi-
`ment.
`
`In FIGS. 5 and 6, it is assumed that a sound source of a
`transmitted voice S(z) is set at equal distances from respec-
`tive microphones 111, 112 and 211. A noise W1(z) comes
`from the direction of an extendedstraightline for connecting
`the microphones 111 and 112. A noise W2(z) comes from the
`direction of an extendedstraight line for connecting a center
`point between the microphones 111 and 112 and the micro-
`phone 211.
`In FIG. 5, processing from the two microphones 111 and
`112 to the level adjuster 120 is similar to that of the first
`embodiment, and thus overlapped descriptions will be omit-
`ted. However, in the signal power comparing section 117, an
`output signal S203 obtained by selecting one of the output
`signals 105a@ and 105b by a selection switch SWais further
`outputted.
`In this case, if the level of an output signal S107a is
`smaller than that of an output signal $107b, the selection
`switch SWaselects the output signal $105a, and this output
`signal $105a is then sent out as an output signal S203. If the
`level of the output signal S107b is smaller, the selection
`switch SWaselects the output signal $105b, and this output
`signal $1055 is then sent out as an output signal S203.
`Asignal outputted from the microphone 211is inputted to
`a filter 218. A delayed microphone output signal S202 is
`inputted to noise suppressing sections C and D. Signals
`S$207a and S207b outputted from the noise suppressing
`sections C and D are inputted to a power comparing section
`217 and a selection switch SWb. The selection switch SWb
`
`selects an output signal having a smaller level from the
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`12
`
`6
`output signals S207a and S207b based on a signal outputted
`from the signal power comparing section 217, and a level
`adjuster 220 adjusts this signal to a % level and then outputs
`the signal as a transmission signal S208.
`Next, the constitutions of the noise suppression sections C
`and D will be described.
`
`The constitutions of these two noise suppressing sections
`C and D are similar to each other. The delayed microphone
`output signal S202 is inputted to an adder 213a (subtracter
`2145). A transmission signal from the level adjuster 120 is
`inputted to a subtracter 214a (adder 2135).
`A signal S205a (S205b) outputted from the subtracter
`214a (2145) is inputted to a filter 215a@ (215b) having a
`transmission function Ga(z) decided based on the positions
`of the microphones. In a subtracter 216a@ (2165), a signal
`$206a (S206b) outputted from the filter 215a (2155) is
`subtracted from a signal S204a (S2045) outputted from the
`adder 213a (213d), and the result of this subtraction is
`inputted to an adder 219a (2195).
`A signal $203 outputted from the selection switch SWa is
`subjected to filtering by a filter 220a (2205), and then
`inputted to the adder 219a@ (219d). A signal outputted from
`the adder 219a (2195)is outputted as an output signal 2074
`(207b) of the noise suppressing section C (D). Next, the
`principle of the noise suppression of the second embodiment
`will be described based on numerical expressions.
`In FIG. 6, it is assumed that a transmitted voice signal is
`denoted by S(z), noise signals are denoted by W1(z) and
`W2(z) and transmission functions of acoustic passing depen-
`dent on the disposition of the three microphones 111, 112
`and 211 are represented by D1(z), D1(z) and D1(z) respec-
`tively. It is also assumedthat noise signals come in vertically
`with respect to transmitted voices. Processing from the two
`microphones111 and 112 to the outputting of a transmission
`signal $108 is similar to that of the first embodiment, and
`thus overlapped descriptions will be omitted. A transmission
`signal S108 is represented by the following expression (15).
`Expression 15
`
`S(@)+W2()
`
`(15)
`
`A signal $202, Aa(z) outputted from the microphone 211
`and a transmission signal S108, Ba(z) are respectively
`represented by the following expressions (16) and (17).
`Expression 16
`
`Aa(2)=S(2)+W2@)D22)+W2)D3(2)
`
`Expression 17
`
`Ba(z)=S(2)+W,(2)
`
`(16)
`
`(47)
`
`The output signal S202, Aa(z) and the transmission signal
`$108, Ba(z) are inputted to the two noise suppressing
`sections C and D. The noise suppressing sections C and D
`perform similar operations. The output signal 204a (2045)
`obtained by addition and the output signal S205a (2055)
`obtained by subtraction are represented by the following
`expressions (18) and (19).
`
`[Expression 18]
`Pa(z) = Aa(z) + Ba(z)
`= 2S(z) + Wo(z)(L + Do(z)) + Wi (Z)Ds(z)
`
`(18)
`
`12
`
`

`

`5,982,906
`
`7
`-continued
`
`[Expression 19]
`Q@) = BR) - A®)
`= W2(z)(1 — D(z) — Wy (z)D3 (2)
`
`(19)
`
`From these expressions (18) and (19),
`expression (21) is obtained.
`
`the following
`
`[Expression 21]
`S(@) = 1/2 [Pa(z) - 1. + Do(z))/ (1 — Da(@))Qa@ -
`
`21)
`
`2/U. - Da(z))D3@)/U - Di (OW)]
`
`Herein, since transmission functions for Q(z) and Qa(z)
`correspond to the filters 215a@ and 220a,
`the following
`expressions (23) and (24) are obtained.
`Expression 23
`
`Ga(z)=(1+D2@))/(1-D2))
`
`Expression 24
`
`Ha(z)=-2/(1-D2(2))D3(2)/1-D2@))
`
`(23)
`
`(24)
`
`Similar processing is performed for the noise suppressing
`section D.
`The noise suppressing section C processes a noise W2(z)
`shown in FIG. 6 so as to suppress the same, and the noise
`suppressing section D processes a noise W2a(z) shown in
`FIG. 6 so as to suppress the same. Accordingly, the output
`signals $207a and S207b of the two noise suppressing
`sections C and D are compared with each other for power,
`and the signal having smaller power is outputted via the
`selection switch SWb and the level adjuster 220. From the
`foregoing description, it can be understood that a transmitted
`voice after having suppressed noise signals can be extracted
`without
`the occurrence of distortion.
`In the above
`
`the incoming direction of noise signals was
`description,
`assumed to be vertical
`to the transmitted voice, but by
`adjusting the transmitting functionsof the filters 215a, 215b,
`220a and 2205,noises of given directions can be suppressed.
`Next, the use of the microphones shown in FIGS. 5 and
`6 for the hand set of a portable telephone set will be
`described.
`
`FIG. 7 is a front view showing the constitution of a
`portable telephone set,
`in which a hand set using three
`microphonesis disposed, and FIG. 8 is a view showing the
`positional and distance relations between the three micro-
`phonesandthe relations thereof with noise incoming direc-
`tions.
`In FIGS. 7 and 8, a transmitted sound sourceis placed at
`equal distances from the microphones 111, 112 and 211, and
`its voice and its incoming direction are assumed to be in a
`direction penetrating the front of a paper surface to the back.
`Since two noises W1(z) and W2(z) comein from sufficiently
`far places for an interval between the two microphones,the
`incomingdirections thereof can be consideredto be identical
`to each other in the handset.
`The noise signal W1(z) inputted to the microphone 111 is
`delayed by a distance dl compared with a noise signal
`inputted to the microphone 112. The noise signal W1(z)
`inputted to the microphone 211 is delayed by a distance d3
`compared with a noise signal inputted to the microphone
`112. Further, the noise signal W2(z) inputted to the micro-
`phone 211 is delayed by a distance d2 compared with noise
`signals inputted to the microphones 111 and 112.
`
`10
`
`15
`
`20
`
`30
`
`35
`
`40
`
`50
`
`55
`
`60
`
`65
`
`13
`
`8
`Herein,if a sampling frequency for signal processing is 16
`kHz,by selecting the positions of the microphones 111, 112
`and 211 so asto satisfy dl=d2=42.5 mm and d3=21.25 mm,
`delays by distances d1 and d2 occur for two sampling cycles,
`and a delay by a distance d3 occurs for one sampling cycle.
`Theserelationships are represented by the following expres-
`sions (25) and (26).
`
`[Expression 25]
`Di@=D,="%°
`[0064]
`
`[Expression 26]
`D3) =z
`
`(25)
`
`(26)
`
`Accordingly, transmission functions of the filters 1155,
`215a and 220a are represented respectively by the following
`expressions (27), (28) and (29).
`Expression 27
`
`Fb(2)=(142-?)/(1-Z-2)
`
`Expression 28
`
`Ga(2)=(14Z-?\/(1-Z7)
`
`Expression 29
`
`Ha(2)=-2/(1-2-?)x“(1-2°2)
`
`(27)
`
`(28)
`
`(29)
`
`Then, the term z~* of the denominatoris multiplied by a
`coefficient of O<a<1 in order to compensate for the insta-
`bility of each filter. This relationship is represented by each
`of the following expressions (30), (31) and (32).
`Expression 30
`
`Fb(2)=(142-°)/(1-az-?)
`
`Expression 31
`
`Ga(2)=(142)/(1-az"?)
`
`Expression 32
`
`Ha(2)=-2/(1-az~?)z4y(1-az)
`
`(30)
`
`(31)
`
`(32)
`
`However, (a) must be approximated to 1 as much as
`possible.
`Transmission functionsofthe filters 115a, 215b and 220b
`are similarly decided. Specifically, these transmission func-
`tions are represented by the following expressions (33), (34)
`and (35).
`Expression 33
`
`Fa(z)=(142-?)/(1-27)
`
`Expression 34
`
`Ga(2)=(142-)/(1-2)
`
`Expression 35
`
`Hb(2)=22°7/(1-2-)zYd?)
`
`(33)
`
`(34)
`
`(35)
`
`As a result, even if noises come in from any directions
`vertical to the incoming direction of transmitted voices, the
`noises can be suppressed. If the transmission functionsof the
`filters 115a@ and 115),the filters 215a@ and 215b andthe filters
`220a and 2205 are notidentical to each other, microphones
`having noise suppression characteristics can be realized, the
`
`13
`
`

`

`5,982,906
`
`9
`in
`noise suppressing characteristics being asymmetrical
`upper and lowerand left and right sides when seen from the
`incoming direction of transmitted voices.
`Since a transmitted sound source direction can be option-
`ally set based on Ea(z), Eb(z) and Ez(z) immediately after
`the inputs are made to the microphones,if the microphones
`cannot be disposed at equal distances from the original
`sound source direction, the microphones can beinstalled at
`equal distances by adjusting the delaying amounts thereof.
`As apparent from the foregoing, according to the noise
`suppressing transmitter of the present invention, the two or
`three microphones are disposed on the straight line (on the
`same plane) orthogonal to the transmitted sound source,
`addition and subtraction are performed for voice signals
`outputted from the microphones, a signal obtained by fil-
`tering the subtracted signal based on the transmission func-
`tion decided by the disposition of the microphonesis sub-
`tracted from a signal obtained by addition, and after
`suppressing noises, an output signal is sent out.
`As a result, no distortion occurs in output transmission
`signal and a high-quality speech voice having noises sup-
`pressed can be obtained. In addition, since the two or three
`microphonesare disposed on the same place as the casing,
`disposing sections are not so thick. In other words,installing
`of the two or three microphones in a miniaturized and thin
`device can be facilitated.
`The entire disclosure of Japanese Patent Application No.
`8-312479 filed on Nov. 22, 1996 including specification,
`claims, drawing and summary are incorporated herein by
`reference in its entirety.
`Whatis claimedis:
`1. A noise suppressing transmitter for outputting voice
`signals by suppressing ambient noises, comprising:
`first and second microphones;
`first and second noise suppressing meansfor suppressing
`noises by performing addition and subtraction for out-
`put signals from said first and second microphones;
`power comparing means for comparing power between
`output signals from said first and second noise sup-
`pressing means; and
`selecting meansfor selecting, among said output signals
`of said first and second noise suppressing means, an
`output signal having smaller power based on a signal
`from said power comparing means,and outputting said
`output signal.
`2. The noise suppressing transmitter of claim 1, wherein
`said first and second microphones are disposed on straight
`lines orthogonal to transmitted voice sources.
`3. The noise suppressing transmitter of claim 1, wherein
`said first noise suppressing means includes an adder for
`adding together said output signals from said first and
`second microphones, a first subtracter for subtracting said
`output signal of said second microphone from said output
`signalof said first microphone,a filter for filtering an output
`signal from said first subtracter based on a transmission
`function determined by disposition of said first and second
`microphones and a second subtracter for subtracting an
`output signal of said filter from an output signal of said
`adder, and said second noise suppressing means includes an
`adder for adding together said output signals from saidfirst
`and second microphones, a first subtracter for subtracting
`said output signal of said first microphone from said output
`signal of said second microphone, a filter for filtering an
`output signal from said subtracter based on a transmission
`function determined by disposition of said first and second
`microphones and a second subtracter for subtracting an
`output signal of said filter from an output signal from said
`adder.
`
`10
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`10
`4. The noise suppressing transmitter of claim 1, further
`comprising level adjusting means for adjusting a level of an
`output signal from said selecting means and outputting a
`level-adjusted signal.
`5. The noise suppressing transmitter of claim 1, wherein
`said first and second microphones are disposed on a plane
`identical to a plane for a casing of a microphone of a device
`for performing voice transmission.
`6. The noise suppressing transmitter of claim 4, wherein
`said device for performing voice transmissionis at least one
`of a hands-free device of a television telephone system and
`a portable telephoneset.
`7. A noise suppressing transmitter for outputting voice
`signals by suppressing noises, comprising:
`first, second and third microphones;
`first and second noise suppressing means for suppressing
`noises by performing addition and subtraction for out-
`put signals of said first and second microphones;
`first power comparing means for comparing power
`between output signals from said first and second noise
`suppressing means;
`first selecting means for selecting, among said output
`signals of said first and second noise suppressing
`means, an output signal having smaller power based on
`a signal from said power comparing means, and out-
`putting said output signal;
`third and fourth noise suppressing means for suppressing
`noises by performing addition and subtraction among
`an output signal from said third microphone, said
`output signal of said first selecting means and a signal
`obtained by subtraction in at least one of said first and
`second noise suppressing means;
`second power comparing means for comparing power
`between output signals from said third and fourth noise
`suppressing means; and
`second selecting means for selecting, among said output
`signals of said third and fourth noise suppressing
`means, an output signal having smaller power based on
`a signal from said second power comparing means, and
`outputting said output signal.
`8. The noise suppressing transmitter of claim 7, wherein
`said first, second and third microphones are disposed on
`straight lines orthogonal to transmitted voice sources.
`9. The noise suppressing transmitter of claim 7, further
`comprising level adjusting meansfor adjusting a level of a
`signal from oneof said first and second selecting means and
`outputting a level-adjusted signal.
`10. The noise suppressing transmitter of claim 7, wherein
`said first, second and third microphones are disposed on a
`plane identical to a plane for a casing of a microphone of a
`device for performing voice transmission.
`11. The noise suppressing transmitter of claim 10,
`wherein said device for performing voice transmission is at
`least one of a hands-free device of a television telephone
`system and a portable telephoneset.
`12. A noise suppressing transmitter for outputting voice
`signals by suppressing ambient noises, comprising:
`first and second microphones disposed on straight lines
`orthogonal to transmitted voice sources;
`first and second noise suppressing means;
`power comparing means for comparing power between
`output signals from said first and second noise sup-
`pressing means;
`selecting means for selecting, among said output signals
`of said first and second noise suppressing means, an
`
`14
`
`14
`
`

`

`12
`of said first and second microphones and producinga first
`noise suppressed signal by subtracting said filtered signal
`from a signal obtained by said adding, and said second noise
`suppressed signal producing step includes adding together
`said output signals from said first and second microphones,
`subtracting said output signal of said first microphone from
`said output