JP2007235943A - Hands-free system for speech signal acquisition - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a hands-free system for speech signal acquisition in a vehicular cabin. <P>SOLUTION: The present invention is directed to a hands-free system for speech signal acquisition in a vehicular cabin comprising at least one first directional microphone (1) being oriented in a first direction (2), at least one second directional microphone (3) being oriented in a second direction different from the first direction, and a noise reduction device (4) configured to cancel a noise part in a signal from the at least one first directional microphone using a signal from the at least one second directional microphone. <P>COPYRIGHT: (C)2007,JPO&amp;INPIT

Description

  The present invention relates to a hands-free system for audio signal acquisition in a vehicle cabin that exhibits improved signal quality.

  Hands-free systems are used for various purposes in vehicles, particularly in car cabins. For example, a hands-free system can be used to enable hands-free telephone communications. As another example, control of equipment such as a navigation system or an air conditioning system may depend on a hands-free system. For this purpose, the acquired signal is processed using a speech recognition device that determines as a control signal input by the user.

  In the most common hands-free system, at least one microphone is provided to pick up the audio signal from the user. These microphones are usually fixedly mounted inside the passenger compartment of the vehicle. The position of the microphone is selected based on various criteria. On the other hand, microphone positioning and orientation with respect to the expected speaker must be optimized. On the other hand, there are usually structural and / or aesthetic or design limitations that limit the number of possible positions.

  The hands-free system provides high comfort for the user because no headset or handset is required. However, since the sound level of the voice signal decreases with the distance from the speaker's mouth to the microphone, the quality of the signal acquired by the microphone (especially the signal-to-noise ratio) is very poor. In view of this, there is a need to improve the signal quality of the acquired signal and improve / enhance the signal-to-noise ratio.

  One possibility to obtain noise suppression in the signal is to use a microphone array containing several microphones in combination with spatial filtering. In this case, an omni-directional microphone is usually used in the microphone array. However, because of the microphone alignment within the array and subsequent spatial filtering alignment, the desired orientation of the array can be achieved with the result of noise suppression. In particular, the directionality of the system is controlled to be the most sensitive in the expected speaker direction. Noise signals from other directions (eg, other talking people or loudspeakers) are suppressed. A possible implementation of such a microphone array system is described in M.C. Brandstein, D.C. Ward: Microphone arrays: signal processing techniques and applications, Springer Verlag, Berlin (Germany), 2001.

In all of these microphone array systems, the array microphones are used to pick up both audio and noise signals. Separation of the noise signal from the audio signal is accomplished by processing the entire signal acquired by the microphone array system. However, since the signal to be processed always contains both a speech part and a noise part, the processing of the signal is very complex and the resulting signal still contains a noise part.
M.M. Brandstein, D.C. Ward: Microphone arrays: signal processing techniques and applications, Springer Verlag, Berlin (Germany), 2001

  Thus, it is a problem underlying the present invention to provide a simplified hands-free system that results in a further improved signal-to-noise ratio in the acquired signal. This problem is solved by a hands-free system according to claim 1.

  Accordingly, a hands-free system for obtaining an audio signal in a passenger cabin of a vehicle, wherein the at least one first directional microphone is oriented in a first direction and a second direction different from the first direction At least one second directional microphone directed to and a noise reduction device, wherein a noise portion in the signal from the at least one directional microphone is represented by a signal from the at least one second directional microphone. There is provided a hands-free system comprising a noise reduction device configured to cancel using.

  With such a hands-free system, a signal having an audio part and a noise part is acquired on the one hand using at least one first directional microphone, and on the other hand using at least one second directional microphone ( It is possible to acquire a noise signal (at least one second directional microphone is pointing in a different direction than the at least one first microphone). Thus, a signal picked up by at least one second microphone that contains (almost) only noise can be used to remove a noise portion from the signal picked up by at least one first microphone.

  In this context, it is pointed out that the term “noise” indicates an undesired signal, whereas the term “speech” is used to indicate any type of desired signal. In other words, “noise” may include signals that also have audio elements emanating from, for example, a loudspeaker or another speaker in a vehicle cabin. However, these signals are not desired for specific purposes.

  The use of a directional microphone has the advantage that the first microphone can be arranged to face the expected speaker position, so that the second microphone originated from the speaker because of the different facing position Only signals that are not signals are acquired, or such signals are acquired with priority. Therefore, this hands-free system makes it possible to acquire a signal having a desired audio part without an additional complicated filtering method.

  The angle between the pointing direction of the first directional microphone and the pointing direction of the second directional microphone is between 20 ° and 180 °, preferably between 90 ° and 180 °, most preferably Is between 130 ° and 180 °.

  The noise reduction device may be configured to subtract a signal based on the signal from the at least one second directional microphone from the signal from the at least one first directional microphone.

  In this way, the noise part is calculated from the signal acquired by the at least one first directional microphone in a convenient and simple manner. In particular, the subtraction, ie the signal based on the signal from the at least one second directional microphone, may be based on or equal to the sum or average of the signals acquired by all the second directional microphones.

  The noise reduction device may be further configured to cancel a noise portion in the signal from the at least one second directional microphone using the signal from the at least one first directional microphone.

  In this way, the hands-free system can acquire enhanced audio signals or desired signals from two different directions. In particular, in a vehicle cabin, at least one first microphone and at least one second microphone are arranged such that at least one first microphone faces the driver and at least one second microphone is on the front seat. It may be arranged to face the passenger. Then, when the driver is speaking, the first microphone is responsible for picking up the audio signal and the second microphone obtains the noise reference signal. When the front seat passenger is speaking, the roles of the first and second microphones are switched.

  In other words, the noise reduction device is based on a noise reference signal acquired by at least one microphone (i.e., the second microphone or the first microphone, respectively) of the first microphone or the second microphone. It is configured symmetrically so as to cancel the noise portion in the signal acquired by either.

  In the above-described embodiment, the hands-free system is configured in accordance with predetermined criteria in a noise portion that is in a signal from the at least one first directional microphone and / or in a signal from the at least one second microphone. It is possible to further comprise a control means for controlling signal processing of the noise reduction device so that a certain noise portion is canceled.

  With such a control means, the noise reduction system is used in a convenient manner with the first and second microphones to acquire the audio signal from one of the two pointing directions of the microphone and point to the other. It is possible to suppress the noise portion of the microphone using the existing microphone.

  Several criteria can be used by this control means. For example, whenever an audio signal is detected in a signal from a first microphone, a noise portion in the signal from at least one first microphone causes the signal from the second microphone to be a noise reference signal. The noise reduction device is controlled so as to be cancelled. In this way, priority is given to speakers to which at least one first microphone is facing.

  Alternatively, the control means can be configured such that the noise part in the signal from the microphone from which the audio signal was first detected is canceled. In this case, each time one speaker stops speaking, both the first microphone and the second microphone are ready to acquire the audio signal.

  In the previously described hands-free system, at least one first directional microphone and / or at least one second directional microphone can be provided as a microphone array having at least two directional microphones, and a noise reduction system May comprise a beamformer for processing the signal from each microphone array.

  In particular, the use of a microphone array and corresponding beamformer to pick up the desired signal further improves the signal-to-noise ratio in the signal containing the desired signal. This will enhance the resulting signal after cancellation of the noise portion. If at least one second microphone is also used to pick up the desired signal, such an advantageous enhancement of the resulting signal is the same for the second microphone as well as the microphone array and the corresponding beamforth. Obtained when the mer is used.

  If both the first microphone and the second microphone are provided as a microphone array having at least two directional microphones, a beamformer may be provided for processing signals from both microphone arrays. Alternatively, a beamformer may be provided for each microphone array.

  According to a further aspect of the above hands-free system, the number of first microphones may be equivalent to the number of second microphones. Such a symmetric case makes it possible to pick up the desired audio signal with similar signal quality using both the first and second microphones.

  The at least one first microphone and the at least one second microphone may be provided in pairs such that the first microphone and the second microphone are each disposed on a common support frame.

  With such a pair arrangement, a dual microphone is obtained that includes a first directional microphone and a second directional microphone, and each microphone faces a different direction. This type of dual microphone can be provided in a very compact form. The compact form has the advantage that it does not require much space, especially when mounting a microphone or microphone array in a vehicle cabin. Each first microphone and second microphone may be further provided in a common housing.

  In the above hands-free system, each first microphone may have a first output, each second microphone may have a second output, and a noise reduction device may be connected to each microphone output. One input unit, a second input unit for each second microphone output unit, a cancellation means, a first signal path connecting at least one first input unit and the cancellation means, and at least A second signal path connecting one second input unit and the cancellation means, wherein the cancellation means uses the signal received via the second signal path to Each of the first microphone outputs is connected to a first input of the noise reduction device, and may be configured to cancel the noise portion from the signal received via the path, 2 microphone output unit is connected to the second input of the noise reduction device.

  In this way, a particularly advantageous configuration of the noise reduction device and a connection via corresponding signal paths to at least one first microphone and second microphone are obtained.

  In particular, the cancellation means may be configured to subtract a signal based on the signal received via the second signal path from the signal received via the first signal path.

  In the above hands-free system, an adaptive filter, particularly based on the LMS, NLMS or RLS algorithm, can be provided between the cancellation means on the second signal path and each second input.

  Such an adaptive filter leads to further enhancement of the audio signal. The adaptation of the adaptive filter can be based on the output signal of the cancellation means. Each adaptive filter may be adaptively optimized to evaluate residual noise in the signal received by the cancellation means via the first signal path. Thus, the adaptive filter is adapted to cancel such residual noise as much as possible.

  In the above hands-free system, at least one first directional microphone is provided as a microphone array having at least two directional microphones, a beam former is provided on the first signal path, and a noise reduction device. For each of the first inputs, the first signal path may comprise a sub-path connecting the first input and the beamformer.

  As a result of such an arrangement having a microphone array and a beamformer, the signal acquired by the first microphone is further enhanced, especially with respect to the signal to noise ratio.

  The beamformer can be a fixed beamformer. As an example, the beamformer can be a delay-and-sum beamformer in which signals from different microphones are delayed so that signals entering the microphone from a preferred direction can be summed in phase. In this way, the delayed microphone signals are added. According to another option, a filter-and-sum type (beamformer “superdirective beamformer”) may be used. Furthermore, instead of a fixed beamformer, an adaptive beamformer can also be used.

  The hand-free system may further include a third signal path connecting each first input unit and the cancellation means, wherein a blocking matrix and a subsequent adaptive filter are provided in the direction of signal flow. A third signal path for each first input section, a subpath connecting the first input section and the blocking matrix, and a subpath connecting the blocking matrix and the adaptive filter Is provided.

  In this way an adaptable structure is obtained. This structure is sometimes referred to as a “generalized side-lobe canceller (GSC). In this way, signal quality can be further improved.

  The blocking matrix functions to block portions of the signal so that a noise reference signal is obtained. The blocking matrix can be realized in various ways. For example, the blocking matrix may have a form in which subtraction of adjacent input channels is performed in pairs. Another possibility is that L.M. Griffiths, Ch. Jim: “An alternative approach to linearly constrained adaptive beamforming”, IEEE Trans. On Antennas and Propagation, Vol. 30, no. 1, pp. 27-34 (January 1982).

  The hand-free system includes a subtracting unit between each second input unit of the noise reduction device on the second signal path and the cancellation unit, a first signal path and each subtracting unit. And a fourth signal path having a sub-bus to be connected.

  Such a configuration provides a reduction in the desired signal portion in the signal received from at least one second directional microphone. As a result, these desired signal portions are not removed from the signal received from the at least one directional microphone.

  The adaptation of the adaptive filter on the subpath may be based on the output of the corresponding cancellation means on the second signal path. With these additional adaptive filters, it is possible to compensate for the desired signal portion that leaks to the noise reference signal.

  The above hands free system may further comprise an adaptive controller having a speech detector for controlling the adaptive filter according to speech detection. In particular, the hands-free system may comprise an adaptive controller for adapting all adaptive filters according to speech detection.

  In particular, the adaptive controller may be configured to initiate adaptation of the adaptive filter on the second signal path if no speech is detected on the first signal path. The adaptive controller adapts the adaptive filter on the signal path connecting the first signal path and the subtracting means on the second signal path when the desired speech is detected in the signal on the first signal path. It can be configured to start. Such an adaptive controller further prevents cancellation of the desired signal.

  In the above hands-free system, each first microphone output unit may be connected to a second input unit of the noise reduction device, and each second microphone output unit is connected to a first input unit of the noise reduction device. And the hands-free system may perform signal processing of the noise reduction device from a noise portion and / or from at least one second microphone in a signal from at least one first microphone according to predetermined criteria. It may further comprise control means for controlling the noise part in the signal to be canceled by the cancellation means.

  Such a configuration leads to a symmetrical arrangement. This arrangement makes it possible to cancel a noise part in the signal received from at least one second directional microphone using the signal received from at least one first directional microphone. Have advantages.

  In particular, the cancellation means may be further configured to subtract a signal based on the signal received by the first signal path from the signal received by the second signal path. Furthermore, a signal path having an adaptive filter connecting the first input and the cancellation means and / or a beamformer connecting the second input and the cancellation means may be provided. . These signal paths may be the same as the first signal path and the second signal path described above. The adaptive filter and beamformer can be configured similarly to those on the first signal path and the second signal path.

  It is possible to process signals acquired by other microphones using the adaptive filters and beamformers described above. For this purpose, the second input can be connected to a beamformer on the first signal path, and the first input can be connected to an adaptive filter on the second signal path. In this case, the control means is configured to control the signal flow in the noise reduction device accordingly. Thus, depending on which microphone (first microphone or second microphone) is used to pick up the desired signal, the beam former and cancellation means in the noise reduction device are used in an appropriate manner.

  According to another aspect, the noise reduction device can be a first noise reduction device for providing a first enhanced output signal, and the hands-free system is from at least a first second directional microphone. A second noise reduction device configured to provide a second enhanced output signal by using the signal to cancel a noise portion in the signal from the at least one second directional microphone; Can be prepared.

  This makes it possible to independently enhance the signals from the first microphone and the second microphone. In particular, two augmented output signals (corresponding to a first directional microphone and a second directional microphone, respectively) can be provided.

  In such a case, the hands-free system may further comprise a signal mixer for providing a combined enhanced output signal by combining the first enhanced output signal and the second enhanced output signal.

  The signal mixer is configured such that the mixing weight is controlled in a time varying manner depending on speech activity detected for the at least one first directional microphone and / or the at least one second directional microphone. Can be done. Alternatively, the signal mixer may be configured such that one of the mixing weights is set to 0 according to a predetermined criterion. As an example, if speech activity is detected for one of the first microphone or the second microphone, the mixing weight for the other of the first microphone or the second microphone is set to zero. In this way, the speaker who starts speaking first is given priority.

  In the hand-free system, the second noise reduction device processes a signal from at least one first directional microphone, and the first noise reduction device processes a signal from at least one second directional microphone. Processing at least in part the same way as and processing a signal from at least one second directional microphone, and a first noise reduction device processing a signal from at least one first directional microphone. It may be configured to process in at least partly the same way as processing.

  A configuration that is at least partially symmetrical is thus obtained. In particular, the second noise reduction device includes at least a part of the same components as the first noise reduction device. For example, the second noise reduction device includes a first input unit for each second microphone output unit, a second input unit for each first microphone output unit, a cancellation means, and at least one first A first signal path for connecting the input unit and the cancellation means, and a second signal path for connecting at least one second input unit and the cancellation means. The signal received via the two signal paths can be used to cancel the noise portion from the signal received via the first signal path, each first microphone output section being configured to reduce noise. Connected to the second input of the device, each second microphone output is connected to the first input of the noise reduction device. As a further example, the second noise reduction device may comprise a beamformer and / or cancellation means as in the case of the first noise reduction device.

  The hands-free system may further include an acoustic echo canceller (AEC).

  Such acoustic echo cancellers can use the loudspeaker excitation signal as a reference for canceling further noise portions in the signal received from the at least one first directional microphone. In this way, acoustic feedback from the loudspeaker to the microphone is also compensated. Preferably, if more than one loudspeaker is present, the acoustic echo canceller may comprise one filter for each loudspeaker.

  The present invention comprises a first directional microphone oriented in a first direction and a second directional microphone oriented in a second direction different from the first direction, the first microphone and The use of dual microphones in which a second microphone is arranged on a common support frame, in particular to form a microphone array for a hands-free system for acquiring audio signals in a vehicle cabin as described above; Provide use.

  In addition, a vehicle cabin can be provided with the above hand-free system.

  In the following, further features and advantages of the invention will be described with reference to the drawings.

Furthermore, the present invention provides the following means.
・ (Item 1)
A hands-free system for acquiring audio signals in a vehicle cabin, the system comprising:
At least one first directional microphone (1) oriented in a first direction (2);
At least one second directional microphone (3) oriented in a second direction different from the first direction;
A noise reduction device (4), wherein a noise part in the signal from the at least one first directional microphone is canceled using the signal from the at least one second directional microphone. A hands-free system comprising a configured noise reduction device.
・ (Item 2)
Item 1. The noise reduction device is configured to subtract a signal based on a signal from the at least one second directional microphone from a signal from the at least one first directional microphone. The described hands-free system.
・ (Item 3)
The noise reduction device is further configured to cancel a noise portion in the signal from the at least one second directional microphone using the signal from the at least one first directional microphone. Item 3. The hands-free system according to item 1 or 2.
・ (Item 4)
According to a predetermined criterion, the noise is such that a noise part in the signal from the at least one first directional microphone and / or a noise part in the signal from the at least one second microphone are canceled. Item 4. The hand-free system according to Item 3, further comprising control means for controlling signal processing of the reduction device.
・ (Item 5)
The at least one first directional microphone and / or the at least one second directional microphone are provided as a microphone array having at least two directional microphones;
The noise reduction system comprises a beamformer (8) for processing the signal from each microphone array.
Item 5. The hands-free system according to any one of items 1 to 4.
・ (Item 6)
6. The hands-free system according to any one of items 1 to 5, wherein the number of first microphones is equal to the number of second microphones.
・ (Item 7)
The item, wherein the at least one first microphone and the at least one second microphone are provided in pairs such that the first microphone and the second microphone are each disposed on a common support frame. The hands-free system according to any one of 1 to 6.
・ (Item 8)
Each of the first microphones has a first output;
Each of the second microphones has a second output;
The noise reduction device
A first input (6) for each first microphone output and / or second microphone output;
A second input (9) for each second microphone output and / or first microphone output;
Cancellation means (5);
A first signal path (7) connecting the at least one first input (6) and the cancellation means (5);
A second signal path (10) connecting the at least one second input (9) and the cancellation means (5);
The cancellation means (5) uses the signal received via the second signal path (10) to remove the noise portion from the signal received via the first signal path (7). Is configured to cancel,
Each first microphone output is connected to a first input (6) of the noise reduction device (4), and each second microphone output is connected to a second input of the noise reduction device. The hands-free system according to any one of items 1 to 7, connected to (9).
・ (Item 9)
Item 9. The item 8, wherein the cancellation means is configured to subtract a signal based on a signal received via the second signal path from a signal received via the first signal path. Hands-free system.
(Item 10)
In item 8 or 9, an adaptive filter (11), in particular based on the LMS, NLMS or RLS algorithm, is provided between the cancellation means on the second signal path and each second input. The described hands-free system.
・ (Item 11)
The at least one first directional microphone is provided as a microphone array having at least two directional microphones;
A beamformer (8) is provided on the first signal path (7), and for each first input of the noise reduction device, the first signal path (7) is connected to the first signal path (7). The hand-free system according to any one of items 8 to 10, further comprising a sub-path (7 ') for connecting the input unit (6) and the beam former (8).
(Item 12)
Item 12. The hands-free system according to Item 11, wherein the beamformer is a fixed beamformer.
(Item 13)
A third signal path (13) for connecting each first input unit to the cancellation means; and in the direction of signal flow, a blocking matrix (14) and a subsequent adaptive filter (15); Is provided in the third signal path (13), and for each first input, the third signal path connects the first input and the blocking matrix, and 13. The hands-free system according to item 11 or 12, comprising a blocking matrix and a sub-path connecting the adaptive filter.
(Item 14)
Subtracting means (16) between the second input section of the noise reduction device on the second signal path and the cancellation means, and the first signal path and each subtracting means. 14. The hands-free system according to any one of items 8 to 13, further comprising a fourth signal path (17) having a sub-bus to be connected.
(Item 15)
15. A hands-free system according to any one of items 8-14, further comprising an adaptive controller having a voice detector for controlling the adaptive filter according to the detection of the voice.
(Item 16)
The first microphone output units are connected to the second input unit (6) of the noise reduction device (4), and the second microphone output units are connected to the first input unit of the noise reduction device. And the signal processing of the noise reduction device connected to (9) and in accordance with a predetermined criterion a noise part in the signal from the at least one first microphone and / or the at least one second microphone Item 16. The hand-free system according to any one of Items 8 to 15, further comprising control means for controlling so that a noise portion in a signal from is canceled by the cancellation means.
・ (Item 17)
The noise reduction device is a first noise reduction device for providing a first enhanced output signal; and
A signal from at least one second directional microphone is used to provide a second enhanced output signal by canceling a noise portion in the signal from the at least one second directional microphone. Further comprising a second noise reduction device configured;
The hands-free system according to any one of items 1 to 16.
(Item 18)
18. The hands-free system of item 17, further comprising a signal mixer for providing a combined enhanced output signal by combining the first enhanced output signal and the second enhanced output signal.
(Item 19)
The second noise reduction device processes a signal from the at least one first directional microphone, and the first noise reduction device processes a signal from the at least one second directional microphone. Processing at least partially in the same manner and processing the signal from the at least one second directional microphone, and the first noise reduction device processing the signal from the at least one first directional microphone. 19. A hands-free system according to item 17 or 18, configured to process in at least partly the same way as to process.
(Item 20)
The hand-free system according to any one of items 1 to 19, further comprising an acoustic echo canceller (19).
・ (Item 21)
A first directional microphone (1) oriented in a first direction and a second directional microphone (3) oriented in a second direction different from the first direction, 21. Use of a dual microphone in which one microphone and the second microphone are arranged on a common support frame, particularly for obtaining an audio signal in a vehicle cabin according to any one of items 1-20 Use to form a microphone array for a hands-free system.

(Summary)
The present invention is a hands-free system for acquiring an audio signal in a passenger cabin of a vehicle, comprising at least one first directional microphone (1) oriented in a first direction (2), At least one second directional microphone (3) oriented in a second direction different from the direction and a noise reduction device (4) in a signal from the at least one first directional microphone The present invention relates to a hands-free system comprising a noise reduction device configured to cancel a noise portion using a signal from at least one second directional microphone.

  FIG. 1 illustrates the configuration of a hands-free system for audio signal acquisition and noise suppression. The example shown in this figure comprises two microphone arrays. The first microphone array comprises two first directional microphones 1 oriented in the first direction indicated by arrow 2. The second microphone array includes two second directional microphones 3 that are oriented in a second direction different from the first direction.

In the illustrated embodiment, the first microphone array points to the expected speaker direction. The outputs of these first microphones are used to generate the primary signal x _P (k).

The hands-free system further includes a noise reduction device 4. The cancellation means 5 forms part of the noise reduction device 4. The cancellation means 5 serves to cancel the noise part from the primary signal x _P. The noise reduction device 4 includes a first input unit 6 for each first microphone 1.

  The first signal path 7 connects the first input unit 6 and the cancellation means 5. A beam former 8 is provided on the first signal path 7. In this example, the beam former 8 is a fixed beam former (FBF). As an example, the beamformer may be a delay-and-sum type or a filter-and-sum type (“superdirective”) beamformer. However, adaptive beamformers can be used as well. The first signal path 7 includes a sub path 7 ′ that connects the first input unit 6 of the noise reduction device 4 and the beam former 8.

The first microphone 1 is used to pick up the desired audio signal represented by x _{S, 1} (k) and x _{S, 2} (k), whereas the second microphone 3 is x _{N , 1} (k) and x _{N, 2} (k) are used to pick up only the noise signals (or at least preferentially). The noise reduction device 4 has an input unit 9 for receiving a signal from the second microphone 3.

The second signal path 10 connects the second input unit 9 and the cancellation means 5. On the second signal path 10, an adaptive filter 11 having transfer functions h _{N, 1} (k) and h _{N, 2} (k) is present between both the second input 9 and the cancellation means 5. Provided.

Adaptive filter 11 is used to filter the noise reference signal, these filters is adapted to cancel maximize the residual noise portion of the primary signal x _P. There are various adaptation methods for updating the adaptive filter 11. As an example, an LMS, NLM or RLS algorithm may be used.

The main advantage of the hands-free system shown in FIG. 1 is that a first microphone array with a directional microphone 1 is used to pick up a desired audio signal that is deteriorated by noise components, whereas the directional microphone 3 The second microphone array with is used to pick up only noise. After the speech signals x _{S, 1} (k) and x _{S, 2} (k) from the first microphone are processed by the beamformer 8, the primary signal x _P (k) is sent to the cancellation means 5. In the cancellation means 5, the adaptively filtered noise reference signals x _{N, 1} (k) and x _{N, 2} (k) are subtracted from the primary signal.

  In the example shown in FIG. 1, each microphone array includes two directional microphones. However, according to an alternative, only one directional microphone can be provided for each direction. In this case, a beamformer is not necessary. As another alternative, each microphone array may comprise two or more directional microphones.

  Furthermore, the number of microphones in the first microphone array and the second microphone array need not be the same. In particular, if the first directional microphone is intended to acquire a desired audio signal, while the second directional microphone is only intended to acquire a noise signal, a microphone array having three or more microphones is first. Fewer microphones, in particular only one directional microphone, can be provided as at least one second directional microphone.

  In the example shown in FIG. 1, different microphones are provided as independent microphones. As a result, if such a hands-free system is to be installed in a vehicle cabin (eg, an overhead console), each microphone must be installed independently. An alternative to this is shown in FIG.

  In this figure, a dual microphone is schematically shown. Such a dual microphone includes a first directional microphone 1 and a second directional microphone 3, and both microphones indicate different directions. Their pointing directions are indicated by arrows 2. The angle α between the pointing directions is preferably between 90 ° and 180 °, in particular between 130 ° and 180 °.

  In this dual microphone, two directional microphones are mounted on a common substrate (not shown). Furthermore, the illustrated example also includes a common housing 12. In this way, a very compact arrangement of two directional microphones is obtained.

  As shown in FIG. 3, two of these dual microphones can be arranged to form two microphone arrays. The first microphone 1 of both dual microphones together forms a first microphone array, while the second microphone 3 of both dual microphones forms a second microphone array. Although FIG. 3 illustrates only two dual microphones, other numbers of dual microphones can be used as well to obtain a microphone array, each with three or more directional microphones. The point should be understood.

  The distance d between two adjacent dual microphones depends on the type of beamformer used, the distance between the microphone array and the expected mouth position of the speaker (eg driver), for mounting the microphone array It is selected according to various parameters such as the interval restriction. According to a general example, dual microphones can be placed at a spacing of d = 7 cm.

  The angle α between the pointing directions can be selected depending on the intended use of the microphone array. For example, if both microphone arrays are used to pick up sound in the car, the directional microphone is directed between the speaker and the front seat passenger.

  FIG. 4 shows an example of a hands-free system that, when used, reduces the problem of leaking a desired audio signal into the noise reference microphone 3 (due to acoustic wave reflection or non-ideal directional characteristics of the microphone). Show. In this figure, the same reference numerals as those in FIG. 1 denote the same elements as those in FIG.

In this example, a third signal path 13 that connects the first input unit 6 and the cancellation means 5 is provided. On this third signal path 13, a blocking matrix (BM 14) and a subsequent adaptive filter 15 having a transfer function h _GSC (k) in the direction of signal flow are provided. The blocking matrix 14 is configured to block desired audio portions in the audio microphone signals x _{S, 1} (k) and x _{S, 2} (k). The resulting signal x _B (k) is used as the noise reference signal.

Thus, there are a total of three noise reference signals with two microphone signals x _{N, 1} (k) and x _{N, 2} (k). Such a structure comprising a blocking matrix followed by an adaptive filter is called a “generalized sidelobe canceller” (GSC). The cancellation means 5, three noise reference signals are subtracted from the primary signal x _P.

  When the audio signal leaks to the noise reference signal, not only the noise part of the reference signal but also a part of the noise part of the noise reference signal correlates with the primary signal. In this case, the adaptive filter shown in FIG. 1 also has a risk of canceling a desired audio signal portion.

To overcome this problem, the noise reduction device shown in FIG. 5 can be provided. In this example, an additional subtraction means 16 is provided on the second signal path 10 between each second input section and the cancellation means 5. Furthermore, a fourth signal path 17 that connects the first signal path 7 and the subtracting means 16 is provided. On each fourth signal path 17, an adaptive filter 18 having transfer functions h _{N, 1} (k) and h _{N, 2} (k) is provided. These additional adaptive filters 18 are used to clean the noise reference signal from the speech portion. The resulting signals x _{R, 1} (k) and x _{R, 2} (k) are used as inputs to the adaptive filter 11.

  Preferably, the filter 18 for reducing the audio part should be updated when the desired speaker is active. On the other hand, to avoid further signal cancellation, the noise filter 11 should not be updated when the desired speaker is active. In view of this, an adaptive controller (not shown) may be provided with a sound detector for providing the desired speaker sound and freezing or updating the adaptive filter, respectively. The audio filter 18 is then adapted only during the desired speaker speech activity (“voice adaptation”), whereas the noise filter 11 is adapted during the noise period (“noise adaptation”).

A further example of a hands-free system is shown in FIG. In addition to the embodiment shown in FIG. 1, the example of FIG. 6 further includes an acoustic echo canceller (AEC) 19. The acoustic echo canceller 19 is a further adaptive filter that uses the loudspeaker signal x _L (k) as an additional noise reference signal. In this way, additional noise portion from the initial signal x _P in cancellation means 5 may be subtracted.

  It should be pointed out that the initial stage signal and / or the noise reference signal used in the example can be weighted according to the filtering along each signal path.

  In the case of multichannel playback via several loudspeakers, an extra filter 19 is preferred for each playback channel in the echo canceller. For example, two filters are advantageous for stereo playback.

  In the above example, the first microphone is always used to provide the primary signal, while the second microphone describes the structure of the noise reduction device so that it is responsible for producing the noise reference signal. However, since the microphone is provided in a symmetric manner, the noise reduction device can be configured accordingly. In other words, signal processing can be applied to signals from either the first microphone array or the second microphone array. In the case of a hands-free system installed in a vehicle cabin, this makes it possible to utilize the system, for example, to enhance audio signals from both drivers and front seat passengers.

  When the noise reduction device is configured by such a method, the same beam former and adaptive filter can be used for any microphone array. In this case, control means may be provided to control the signal flow from each microphone array to the appropriate element for the particular case of the noise reduction device. For example, when the driver is speaking, the structure shown in the figure exists, but when the front seat passenger is speaking, the microphone array and the corresponding connection to the filter in the noise reduction device are switched. In other words, in the latter case, the microphone pointing to the driver is used to pick up the noise reference signal and the microphone pointing to the front seat passenger is used to pick up the voice.

  Control of the signal flow of the microphone signal to each filter of the noise reduction device may depend on a predetermined criterion. For example, control may depend on the speaking activity of both speakers. That is, the microphone array that picks up the voice is first switched to the beamformer and this configuration can be maintained until the first voice pause occurs. Alternatively, this signal flow configuration can be maintained unchanged (predetermined time or until a manual reset is performed).

  Furthermore, the hand-free system shown in the figure and described above can be provided with a second noise reduction configured similarly to the first noise reduction device. In the noise reduction device, a signal from the second directional microphone of the second noise reduction device is sent to the first input unit, and a signal from the first directional microphone is sent to the second input unit. There is a difference. In particular, when the components and signal paths in the second noise reduction device are the same as those in the first noise reduction device, the first microphone signal and the second microphone signal are in the case of the first noise reduction device. It is processed in the opposite way compared to.

  Preferably, such a hands-free system also includes a signal mixer for combining the output signals of the first noise reduction device and the second noise reduction device to obtain a combined enhanced output signal. The mixing weight, i.e. the weight by which the signal is multiplied before addition, can be controlled in a time-varying manner depending on the speech activity detected for the first and second microphones. As an alternative example, the weights can be controlled such that by initially setting the corresponding mixing weights to zero, the speaker who first started speaking is applied continuously and other speakers are blocked.

  Further modifications and variations of the present invention will be apparent to those skilled in the art upon viewing this description. Accordingly, this description is to be construed as illustrative and for the purpose of teaching those skilled in the art the general manner of carrying out the invention. It should be understood that the forms of the invention formed and described herein are to be construed as presently preferred embodiments.

FIG. 1 shows a block diagram of an example of a hands-free system for acquiring audio signals with noise reduction. FIG. 2 illustrates a dual microphone configuration. FIG. 3 illustrates an array of two dual microphones. FIG. 4 shows a block diagram of a hands-free system with enhanced noise reduction. FIG. 5 shows a block diagram of a hands-free system with an additional adaptive filter. FIG. 6 shows a hands-free system having an acoustic echo canceller.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 1st directional microphone 2 1st direction 3 2nd directional microphone 4 Noise reduction apparatus 5 Cancellation means 6 1st input part 7 1st signal path 7 'Subpath 8 Beamformer 9 2nd Input unit 10 Second signal path 13 Third signal path 14 Blocking matrices 15 and 18 Adaptive filter 16 Subtraction means 17 Fourth signal path 19 Acoustic echo canceller

Claims (21)

A hands-free system for acquiring audio signals in a vehicle cabin, the system comprising:
At least one first directional microphone (1) oriented in a first direction (2);
At least one second directional microphone (3) oriented in a second direction different from the first direction;
A noise reduction device (4), wherein a noise part in the signal from the at least one first directional microphone is canceled using the signal from the at least one second directional microphone. A hands-free system comprising a configured noise reduction device.   The noise reduction device is configured to subtract a signal based on a signal from the at least one second directional microphone from a signal from the at least one first directional microphone. Hands free system as described in.   The noise reduction device is further configured to cancel a noise portion in a signal from the at least one second directional microphone using a signal from the at least one first directional microphone. The hands-free system according to claim 1 or 2.   In accordance with a predetermined criterion, the noise is canceled such that a noise part in the signal from the at least one first directional microphone and / or a noise part in the signal from the at least one second microphone are canceled. The hand-free system according to claim 3, further comprising control means for controlling signal processing of the reduction device. The at least one first directional microphone and / or the at least one second directional microphone are provided as a microphone array having at least two directional microphones;
The noise reduction system comprises a beamformer (8) for processing the signal from each microphone array.
The hands-free system according to any one of claims 1 to 4.   The hands-free system according to claim 1, wherein the number of first microphones is equal to the number of second microphones.   The at least one first microphone and the at least one second microphone are provided in pairs such that the first microphone and the second microphone are each disposed on a common support frame. Item 7. The hands-free system according to any one of items 1 to 6. Each of the first microphones has a first output;
Each second microphone has a second output;
The noise reduction device is
A first input (6) for each first microphone output and / or second microphone output;
A second input (9) for each second microphone output and / or first microphone output;
Cancellation means (5);
A first signal path (7) connecting the at least one first input (6) and the cancellation means (5);
A second signal path (10) connecting the at least one second input (9) and the cancellation means (5);
The cancellation means (5) uses the signal received via the second signal path (10) to remove the noise part from the signal received via the first signal path (7). Is configured to cancel,
Each first microphone output is connected to a first input (6) of the noise reduction device (4), and each second microphone output is connected to a second input of the noise reduction device. The hand-free system according to any one of claims 1 to 7, which is connected to (9).   9. The canceling means is configured to subtract a signal based on a signal received via the second signal path from a signal received via the first signal path. The described hands-free system.   10. An adaptive filter (11), in particular based on an LMS, NLMS or RLS algorithm, is provided between the cancellation means on the second signal path and each second input. Hands free system as described in. The at least one first directional microphone is provided as a microphone array having at least two directional microphones;
A beamformer (8) is provided on the first signal path (7), and for each first input of the noise reduction device, the first signal path (7) is connected to the first signal path (7). The hand-free system according to any one of claims 8 to 10, further comprising a sub-path (7 ') for connecting the input unit (6) of the laser beam and the beam former (8).   The hand-free system according to claim 11, wherein the beam former is a fixed beam former.   A third signal path (13) for connecting each first input unit and the cancellation means, further comprising: a blocking matrix (14) in the direction of signal flow; and an adaptive filter (15) that follows. Is provided in the third signal path (13), and for each first input, the third signal path connects the first input and the blocking matrix, and The hand-free system according to claim 11 or 12, comprising a sub-path connecting a blocking matrix and the adaptive filter.   Subtracting means (16) between the second input part of the noise reduction device on the second signal path and the cancellation means, the first signal path and each subtracting means, 14. The hands-free system according to any one of claims 8 to 13, further comprising a fourth signal path (17) having a sub-bus to be connected.   15. A hands-free system according to any one of claims 8 to 14, further comprising an adaptive controller having a speech detector for controlling the adaptive filter according to speech detection.   Each first microphone output is connected to a second input (6) of the noise reduction device (4), and each second microphone output is a first input of the noise reduction device. And the signal processing of the noise reduction device connected to (9) and according to a predetermined criterion, a noise part in the signal from the at least one first microphone and / or the at least one second microphone The hand-free system according to any one of claims 8 to 15, further comprising control means for controlling so that a noise portion in a signal from is canceled by the cancellation means. The noise reduction device is a first noise reduction device for providing a first enhanced output signal; and
A signal from at least one second directional microphone is used to provide a second enhanced output signal by canceling a noise portion in the signal from said at least one second directional microphone. Further comprising a second noise reduction device configured;
The hands-free system according to any one of claims 1 to 16.   The hands-free system of claim 17, further comprising a signal mixer for providing a combined enhanced output signal by combining the first enhanced output signal and the second enhanced output signal.   The second noise reduction device processes a signal from the at least one first directional microphone, and the first noise reduction device processes a signal from the at least one second directional microphone; Processing at least partially in the same manner and processing the signal from the at least one second directional microphone, and the first noise reduction device processing the signal from the at least one first directional microphone. 19. A hands-free system according to claim 17 or 18, configured to process in at least partly the same way as to process.   The hands-free system according to any one of the preceding claims, further comprising an acoustic echo canceller (19).   A first directional microphone (1) oriented in a first direction and a second directional microphone (3) oriented in a second direction different from the first direction, 21. Use of a dual microphone in which one microphone and the second microphone are arranged on a common support frame, in particular in a vehicle cabin according to any one of claims 1-20. Use to form a microphone array for a hands-free system for acquisition.

Images