DE102018102821B4 - A DEVICE FOR DETECTING AND PROCESSING AN ACOUSTIC INPUT SIGNAL - Google Patents

Abstract

An electronic device (102), comprising: a non-linear array of microphones (101, 121) including at least three microphones; a direction detection element (430) configured to detect a direction of a first acoustic signal (A) generated by the microphones (101, 121) in the non-linear array of microphones is received from an external acoustic source (150), to be determined by using a controlling software program run by a processor (118) in the electronic device (102) is performed;a first signal processor (405) configured to remove a first portion of an acoustic signal (313) generated in a second acoustic source (155) from a composite acoustic signal (301) by using the controlling software program, wherein the first signal processor (405) comprises: a microphone selector (440) configured to include a first microphone and a second microphone from the non-linear array select ng from microphones (101, 121) based on the determined direction received from the direction detection element (430);a signal delay element (441) configured to receive the composite acoustic signal (301) from the second microphone and generate a delayed audible signal after a predetermined delay time has elapsed;a signal inverting element (442) configured to receive the delayed audible signal and to generate an inverted audible signal based on the delayed audible signal; anda first signal combining element (414) configured to combine the inverted acoustic signal and the composite acoustic signal received from the first microphone and to generate a first combined acoustic signal;a second signal processor (407) configured removing a second portion of the acoustic signal (313) generated in the second acoustic source (155) from the composite acoustic signal (310) by use of the controlling software program, the second signal processor comprising:a signal delay element (431A-431C) configured to receive the composite acoustic signal from at least two of the microphones (101, 121) in the non-linear array of microphones and a second delayed acoustic signal for each of the at least two microphones (101, 121) based on the determined direction received from the direction detection element (430). n; anda second signal combining element (432) configured to combine each of the second delayed acoustic signals generated by the signal delay element (431A-431C) and to generate a second combined acoustic signal; anda third signal combining element (443) configured to combine the first combined acoustic signal and the second combined acoustic signal to generate a third combined acoustic signal.

Description

BACKGROUND OF THE INVENTION

field of invention

Embodiments of the present disclosure relate to a method and apparatus for processing an acoustic signal to generate a processed acoustic signal having an improved signal-to-noise ratio.

Description of the prior art

The popularity of, and confidence in, electronic devices has increased significantly over the past decade. The popularity of electronic devices such as smart phones, touch-sensitive devices, PDAs, portable computers, portable music players has increased in the last decade. Video telephony and video conferencing facilities have also grown in popularity in recent years, in large part due to the proliferation of high speed Internet and price reductions for facilities that enable it. As the number of electronic devices and the confidence in these electronic devices increases, there is a need for these devices to receive and process an acoustic input signal received from a user so that the acoustic input can be used to cause a desired task is executed.

There has been a desire for years to build systems that can detect, process, and/or transmit various types of acoustic input signals originating from a human being. Although this goal has begun to be realized in recent years, currently available systems are unable to produce results suitable for accurately capturing these received acoustic inputs in environments where noise is common or not are well controllable. In most conventional microphone-containing devices capable of recognizing and/or processing various types of acoustic input signals, it is often difficult for the electronics that process the acoustic input (e.g., speech recognition hardware) to produce the desired human To clearly separate speech from unwanted noise or background noise. This inability to separate acoustic input signals from ambient noise within the environment is primarily due to the difficulties involved in extracting and recognizing the individual sounds that make up human speech. These difficulties are even greater in noisy environments. Put simply, language can be thought of as a sequence of sounds derived from basic sounds called "phonemes" made by a human. One or more phonemes represent a word or phrase. Thus, in order to obtain speech recognition, the extraction of the specific phonemes contained in the received speech is required, which is often extremely difficult in noisy environments.

Furthermore, conventional voice or speech recognition hardware is typically limited to detecting speech within the low end of the speech frequency range, approximately between about 100 Hertz (Hz) and about 3000 Hz due to limitations on the device's sampling frequency and the geometry of the microphone arrays. Thus, a large amount of useful data is lost by these conventional methods because they are unable to detect speech over the full range of speech, extending from 100 Hz to about 8000 Hz, and thus the information contained in lost at the higher end of the speech range, which is between 3000 Hz and 8000 Hz.

With the increase in popularity of speech recognition systems, many users employ them in a variety of environments. These diverse devices are commonly used in a wide variety of moderately noisy to extremely noisy environments, such as an office, conference room, airport, or restaurant. There are already several conventional methods for performing noise reduction, however, many conventional methods can be classified as types of filtering. In the prior art, speech and noise are detected in the same input channel because they are in the same frequency band and have similar correlation properties. Consequently, the filtering inevitably has an effect on both the speech signal and the background noise signal. Distinguishing between voice signal and background noise signal is a challenging task. Speech components received by conventional electronic devices may be perceived as noise components and may be suppressed or filtered along with the noise components. Although speech recognition technology continues to advance, a clear separation of the voiced component from noise components of an audio signal, or in other words, a high signal-to-noise ratio (SNR) is required for acceptable levels of accuracy in the voice identification required, or in some cases for the delivery and recovery of the received audio signal at a remote location.

Further, with the increase in the number of electronic devices and the confidence in these electronic devices, there is a growing desire for the electronic devices to be independent of conventional power sources in the form of outlets so that these independent electronic devices are portable. However, power supply in portable electronic devices is usually limited to a finite energy storage capacity provided by a battery. The rate at which the device consumes energy determines the amount of time the device will operate before the battery needs to be recharged or replaced. Therefore, it is desirable to find ways to reduce the power consumption of the portable device's electronic components, such as speech recognition elements, in order to improve the battery life of the portable electronic devices.

US 2007/0025562 A1 describes methods and devices for targeted sound detection. A microphone array has two or more microphones. Each microphone is paired with a variety of filters. The filters are configured to filter input signals that correspond to the noise picked up by the microphones, thereby producing a filtered output signal. One or more sets of filter parameters for the plurality of filters are pre-calibrated to determine one or more corresponding pre-calibrated listening zones. Each set of filter parameters is selected to capture portions of the input signals that correspond to noise within a specific listening zone and to filter out noise originating outside of the specific listening zone.

A particular pre-calibrated listening zone is selected at runtime by applying to the plurality of filters a set of filter coefficients corresponding to the particular pre-calibrated listening zone. As a result, the microphone array can detect noise originating within the specific listening area and filter out noise originating outside of the specific listening area.

Therefore, there is a need for an electronic device that solves the problems described above. There is also a need for an electronic device capable of efficiently filtering unwanted noise from an input acoustic signal received from an acoustic source.

OVERVIEW OF THE INVENTION

The above needs are met by the electronic device according to claim 1. Further advantageous embodiments are described in the dependent patent claims.

character list

• 1A ist eine schematische Ansicht, die ein Beispiel eines Akustiksignal-Verarbeitungssystems gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung darstellt.
• 1B ist eine Draufsicht des in 1A gezeigten Akustiksignal-Verarbeitungssystems, ohne dass ein zentrales Mikrofon gezeigt ist, gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung.
• 1C ist eine schematische Blockansicht von Bauteilkomponenten, die in dem in 1A gezeigten Akustiksignal-Verarbeitungssystem vorhanden sind, gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung.
• 2A ist eine Ansicht, die eine Draufsicht eines Akustiksignal-Verarbeitungssystems und einer akustischen Quelle gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung zeigt.
• 2B ist ein Graph, der die Zeitverzögerung darstellt, die sich für jedes der in 2A dargestellten Mikrofone ergibt, wenn ein akustisches Signal aus der akustischen Quelle jedem der Mikrofone zugeführt wird.
• 3A ist ein Graph, der ein zusammengesetztes akustisches Signal zeigt, das von einem Mikrofon in dem Akustiksignal-Verarbeitungssystem gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung empfangen werden kann.
• 3B enthält Graphen, die die diversen Komponenten darstellen, die in den zusammengesetzten akustischen Signal vorhanden sind, das in 3A dargestellt ist.
• 4A ist eine schematische Ansicht, die ein Beispiel einer Akustiksignal-Verarbeitungseinrichtung zeigt, die in dem Akustiksignal-Verarbeitungssystem gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung vorhanden ist.
• 4B ist eine schematische Ansicht, die Bereiche der Akustiksignal-Verarbeitungseinrichtung, die in 4A gezeigt ist, gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung zeigt.
• 4C ist eine schematische Ansicht, die ein Richtungserfassungselement der Akustiksignal-Verarbeitungseinrichtung gemäß einer oder mehreren Ausführungsformen der vorliegenden Offenbarung zeigt.
• 4D zeigt ein Verfahren zur Ausführung einer ersten Art einer Signalverarbeitungstechnik gemäß einer Ausführungsform der vorliegenden Offenbarung.
• 4E zeigt ein Verfahren zur Ermittlung einer gewünschten Richtung, aus der ein akustisches Signal empfangen wird, gemäß einer Ausführungsform der vorliegenden Offenbarung.
• 4F zeigt ein Verfahren zur Ausführung einer zweiten Art einer Signalverarbeitungstechnik gemäß einer Ausführungsform der vorliegenden Offenbarung.
• 5A ist eine Draufsicht eines Akustiksignal-Verarbeitungssystems mit drei Mikrofonen gemäß einer Ausführungsform der hierin bereitgestellten Offenbarung.
• 5B ist eine Draufsicht des in 5A dargestellten Akustiksignal-Verarbeitungssystems, das mehrere Erfassungsabschnitte aufweist, die verwendet werden können, um die Richtung einer akustischen Quelle gemäß einer Ausführungsform der hierin bereitgestellten Offenbarung zu ermitteln.
• 5C ist ein Graph, der Gewichtungskoeffizienten darstellt, die verwendet werden können, um die Richtung einer akustischen Quelle gemäß einer Ausführungsform der vorliegenden Offenbarung zu ermitteln.
• 6A-6C zeigen Beispiele von Herzkurvenmustern, die bei unterschiedlichen Frequenzen durch die Verwendung einer ersten Signalverarbeitungstechnik gemäß einer Ausführungsform der hierin bereitgestellten Offenbarung erzeugt werden.
• 6D ist eine Draufsicht eines Akustiksignal-Verarbeitungssystems mit sieben Mikrofonen gemäß einer Ausführungsform der hierin bereitgestellten Offenbarung.
• 7A-7C zeigen Beispiele von Strahlformungsmustern, die bei unterschiedlichen Frequenzen durch die Verwendung einer zweiten Signalverarbeitungstechnik gemäß einer Ausführungsform der hierin bereitgestellten Offenbarung erzeugt werden.
• 8A-8C zeigen Beispiele von Mustern, die bei unterschiedlichen Frequenzen durch Verwendung einer ersten Signalverarbeitungstechnik und einer zweiten Signalverarbeitungstechnik gemäß einer Ausführungsform der hierin bereitgestellten Offenbarung erzeugt werden.

In order that the features of the invention set forth above may be understood in detail, a more specific description of the invention, briefly summarized above, is given by reference to embodiments, some of which are illustrated in the accompanying drawings. However, it should be noted that the attached drawings represent only typical embodiments of this invention and are not intended to limit the scope, the invention allowing other equally effective embodiments.

• 1A 12 is a schematic view illustrating an example of an acoustic signal processing system, in accordance with one or more embodiments of the present disclosure.
• 1B is a plan view of the in 1A acoustic signal processing system shown without a center microphone shown, in accordance with one or more embodiments of the present disclosure.
• 1C Fig. 12 is a schematic block diagram of assembly components included in Fig 1A acoustic signal processing system shown are present, in accordance with one or more embodiments of the present disclosure.
• 2A 12 is a view showing a top view of an acoustic signal processing system and an acoustic source, in accordance with one or more embodiments of the present disclosure.
• 2 B is a graph depicting the time delay experienced for each of the in 2A microphones shown results when an acoustic signal from the acoustic source is supplied to each of the microphones.
• 3A FIG. 14 is a graph showing a composite acoustic signal that may be received by a microphone in the acoustic signal processing system in accordance with one or more embodiments of the present disclosure.
• 3B contains graphs representing the various components present in the composite acoustic signal emitted in 3A is shown.
• 4A 12 is a schematic view showing an example of an acoustic signal processing device included in the acoustic signal processing system according to one or more embodiments of the present disclosure.
• 4B Fig. 12 is a schematic view showing portions of the acoustic signal processing device shown in Fig 4A is shown, in accordance with one or more embodiments of the present disclosure.
• 4C 12 is a schematic view showing a direction detection element of the acoustic signal processing device according to one or more embodiments of the present disclosure.
• 4D FIG. 1 shows a method for performing a first type of signal processing technique, according to an embodiment of the present disclosure.
• 4E FIG. 12 shows a method for determining a desired direction from which an acoustic signal is received, according to an embodiment of the present disclosure.
• 4F 10 illustrates a method of performing a second type of signal processing technique, in accordance with an embodiment of the present disclosure.
• 5A 12 is a top view of a three microphone acoustic signal processing system according to an embodiment of the disclosure provided herein.
• 5B is a plan view of the in 5A 1 shows an acoustic signal processing system having a plurality of acquisition sections that can be used to determine the direction of an acoustic source according to an embodiment of the disclosure provided herein.
• 5C FIG. 14 is a graph depicting weighting coefficients that may be used to determine the direction of an acoustic source, according to an embodiment of the present disclosure.
• 6A-6C 12 show examples of heart wave patterns generated at different frequencies through the use of a first signal processing technique according to an embodiment of the disclosure provided herein.
• 6D 12 is a top view of a seven microphone acoustic signal processing system according to an embodiment of the disclosure provided herein.
• 7A-7C 12 show examples of beamforming patterns generated at different frequencies through the use of a second signal processing technique according to an embodiment of the disclosure provided herein.
• 8A-8C 12 show examples of patterns generated at different frequencies by using a first signal processing technique and a second signal processing technique according to an embodiment of the disclosure provided herein.

To facilitate understanding, the same reference numbers are used, where possible, to designate identical elements that appear together in the figures. It should be noted that elements disclosed in one embodiment may be used to advantage in other embodiments without specific reference. The drawings referred to herein should not be understood to be to scale unless otherwise indicated. Furthermore, the drawings are often simplified, and details or components are omitted for ease of illustration and explanation. The drawings and explanation serve to explain principles explained below, where like designations indicate like elements.

DETAILED DESCRIPTION

In the following description, numerous specific details are set forth to provide a more thorough understanding of embodiments of the present invention. However, those skilled in the art will recognize that one or more embodiments of the present disclosure may be practiced without one or more of these specific details. In other instances, well-known features are not described to avoid obscuring one or more embodiments of the present disclosure.

Embodiments of the disclosure generally include a method and apparatus for receiving and separating unwanted external noise or noise from an acoustic input received from an acoustic source. Embodiments of the disclosure may include an acoustic signal processing system that includes a plurality of acoustic signal detection devices (e.g., microphones) arranged and configured to detect an acoustic signal that generates and communicates with the acoustic signal processing system from any location or location angle within three-dimensional (3-D) space. The acoustic signal processing system is configured to process the acoustic signals received by each of the plurality of acoustic signal detectors using a first signal processing technique capable of separating unwanted low frequency noise from the detected acoustic signals, and using a second signal processing technique capable of separating unwanted noise in a higher frequency range from the detected acoustic signals. The acoustic signal processing system is then configured to combine the signals processed by the first and second signal processing techniques to produce a desired acoustic signal having a high signal-to-noise ratio within a desired frequency range, such as the full speech range .

In some configurations, the acoustic signal processing system is designed to be portable and thus operates from an energy source that has a limited amount of energy stored therein (e.g., battery). Therefore, in some cases, the acoustic signal processing system may be further configured to receive and separate the unwanted external noise from a received input acoustic signal with an efficiency such that the service life of the portable acoustic signal processing system is extended. The acoustic signal processing system may be further configured to receive an acoustic signal from an external source, efficiently remove or separate unwanted noise from the received acoustic signal, and then provide the processed acoustic signal to a software application configured to perform further processing and/or a desired action based on the processed acoustic signal. The acoustic signal processing system can also be designed to supply the processed acoustic signal to a further electronic device which is designed to receive and process the received information in such a way that the second device can carry out a certain desired activity.

1A 1 is a schematic view showing an example of an acoustic signal processing system 100 according to one or more embodiments of the present disclosure. 1B FIG. 12 is a view showing a top view of the acoustic signal processing system 100 shown in FIG 1A is shown. The acoustic signal processing system 100 generally includes an electronic device 102. The electronic device 102 may be a computing device usable as a stand-alone electronic device or may be an electronic device that may be used in conjunction with other electronic devices. In an example, as in 1A As shown, the electronic device 102 is capable of communicating with a separate second electronic device 195 via a wired or wireless communication link 194 . Alternatively, in another example, the electronic device 102 is a component in the second electronic device 195. In either case, the electronic device 102 and/or the second electronic device 195 can be a wireless speaker, a video camera device that includes a CCD camera , keyboard, smart phone, intercom, home automation device, or other suitable electronic device. In one example, electronic device 102 or second electronic device 195 may be an Ultimate Ears Boom speaker, a Harmony universal remote control, or a Logitech Connect or Logitech BCC 950 video conferencing device manufactured by Logitech USA, Newark CA, or Logitech Europa SA of Lausanne, Switzerland. The electronic device 102 or the second electronic device 195 may also be an iPod, iPhone, iPad, Android phone, Samsung Galaxy, squeeze box, Microsoft Service, portable calculator, or similar device. Although the discussion below generally describes or provides examples of an acoustic signal processing system 100 that is a self-contained portable electronic device, this configuration should not be construed as limiting the scope of the disclosure provided herein.

The electronic device 102 includes a plurality of acoustic signal detection devices arranged in a geometric array across one or more surfaces of the electronic device 102 . In some embodiments, the geometric arrangement of the acoustic signal detection devices, or hereinafter referred to as microphones 101, can be provided in a two-dimensional (2-D) arrangement of microphones 101 or a three-dimensional (3-D) arrangement of microphones, the microphones 101 and may have one or more microphones 121 .

Electronic device 102 may have any desired form, such as that shown in FIG 1A shown cylindrical shape, and may have one or more outer surfaces on which one or more of the microphones 101, 121 can be arranged, such as a top surface 106, a side surface 108 and a support surface or holding surface 107. The holding surface 107 is a surface on of the entire electronic device 102 during of normal operation. Although further 1A-1B , 2A , 5A-5B and 6D Showing the microphones 101, 121 arranged on or extending from a surface of the electronic device 102, this arrangement is not intended to limit the scope of the disclosure provided herein, as the microphones 101, 121 can also be arranged so that the External surface of the microphones is flush or recessed with the surface on which they are placed. Although for the purpose of simplifying the explanation, the following disclosure essentially includes an explanation in terms of a geometric arrangement containing a two-dimensional (2-D) array of microphones, this arrangement is not intended to exceed the scope of the disclosure provided herein be limited, as three-dimensional (3-D) microphone arrays may also be employed in one or more of the embodiments disclosed herein. A geometric arrangement of microphones can have at least three microphones or, for example, five microphones 101, as in 1B is shown, or may even have seven microphones as in 6D is shown. However, in some cases it may be desirable to have only two microphones 101 . The microphones 101 can be any type of electrical device capable of converting barometric pressure fluctuations of a sound wave into an electrical signal, and thus, without being limited to a dynamic microphone, they can be a condenser microphone, a piezoelectric microphone, a fiber optic Include microphone, ribbon microphone, MEMS microphone or other similar device. In some embodiments, microphones 101 are omnidirectional microphones capable of detecting acoustic signals from all directions.

In some embodiments, the microphones 101 are arranged in a two-dimensional (2-D) geometric array across the top surface 106 and/or the side surface 108 of the electronic device 102 . In an example, as in 1A and 1B is shown, the geometric arrangement of microphones 101 is evenly distributed over the side surface 108 . As in 1B 1, five microphones 101 are evenly distributed in the geometric array in a circular array across side face 108 such that each microphone 101 is spaced a distance 105 from a center point 103 and has an equal angular spacing 104 therebetween (e.g., 72° ). In general, the microphones 101 are arranged in the electronic device 102 such that all of the microphones 101 are arranged in a geometric arrangement (e.g. a 2-D array or 3-D array) such that the electronic device 102 generates acoustic signals that arriving from any direction, and avoiding error in the detected direction of the received acoustic signal data, which is common in conventional rectilinear microphone configurations. Thus, the electronic device 102 is capable of receiving an acoustic signal "A" provided by an acoustic source 150 located at a distance in three-dimensional space with respect to the acoustic signal processing system 100, and then the to process received acoustic signals such that other undesired acoustic signals provided by other sources located at other positions with respect to the acoustic signal processing system 100 are preferably excluded.

1C 13 is a schematic view showing an electronic assembly 135 in the acoustic signal processing system 100 according to an embodiment of the present disclosure. In general, the electronic assembly 135 includes a processor 118, non-volatile memory 122, a power source 130, and an acoustic signal processing device 400. During operation, the electronic device 102 is configured to generate a plurality of acoustic signals, including an acoustic signal "A". from multiple microphones 101 to receive. The acoustic signal processor 400 and processor 118 then receive the detected acoustic signals from each of the microphones 101 and process the detected input signals to remove or separate the unwanted external noise or sounds from the desired acoustic signal. The processed acoustic signal can then be used to perform another task by the acoustic signal processing system 100 or by other downstream equipment.

The acoustic signal processor 400 generally includes electrical components that can efficiently separate a desired portion of an acoustic signal from received noise using a low frequency signal processing technique and a higher frequency signal processing technique. The processes performed by the acoustic signal processor 400 are believed to reduce the error rate encountered when the processed acoustic signal is used in a subsequent voice recognition, voice communication, voice-activated electronic device control, and/or voice recognition process, in comparison to acoustic signals generated by conventional noise cancellation or noise suppression techniques that are common today will. The processes described herein are also capable of extending the operational life of the acoustic signal processing system 100 before a recharge or replacement of the power source 130 is required. Although the power source 130 described herein may include a battery, the electronic device 102 may temporarily receive power from a wired connection to an outlet, wired charger, or similar device without departing from the basic scope of the disclosure provided herein.

Electronic assembly 135 may include processor 118 coupled to input/output (I/O) devices 116, power source 130, and non-volatile memory unit 122. Storage unit 122 may contain one or more software applications 124, such as the controlling software program described below. Storage unit 122 may also include stored media data 126 used by processor 118 to perform various portions of the methods described herein. Processor 118 may be a hardware device or combination of hardware devices capable of executing software applications and processing data. In some configurations, processor 118 includes a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), and/or a combination of such units. The processor 118 is generally configured to execute the one or more software applications 124 and to process the stored media data 126, each of which may reside in the storage unit 122.

I/O devices 116 are coupled to memory unit 122 and processor 118 and may include devices capable of receiving input and/or devices capable of providing output . The I/O devices 116 include the audio processing device 117 which receives battery power and an input signal 104 and generates the output signal 106 which can be received by the speaker system 111 and then output. The I/O devices 116 may further include one or more wireless transceivers 120 configured to establish one or more different types of wired or wireless communication links to other transceivers residing in other computing devices. For example, a given transceiver in I/O devices 116 could be a Wi-Fi communication link, a short-range (NFC) communication link, or a Bluetooth communication link (e.g., BTLE, Bluetooth Classic) from a variety of communication link types with similar components in the second electronic device 195. In some embodiments, electronic components in the I/O device 116 are configured to transmit signals processed by the acoustic signal processing device 400 to other internal electronic components residing in the acoustic signal processing system 100 and/or to electronics devices external to the acoustic signal processing system 100, as discussed below.

Storage unit 122 may be any technically feasible type of hardware unit configured to store data. For example, storage device 122 may be a hard drive, a random access memory (RAM) module, a flash memory device, or a combination of different hardware devices configured to store data. Software application 124 stored on storage unit 122 may include program code executable by processor 118 to perform various functions associated with electronic device 102 . The stored media data 126 may include any type of information pertaining to a desired control parameter, quasi-directional information, calculated time delay information, noise signal RMS information, user data, electronic device configuration data, device control rules, or other pertinent information , which is explained below. Stored media data 126 may include information provided to and/or received from source 150 or another electronic device, such as second electronic device 195. Stored media data 126 may represent various data files, settings, and/or parameters associated with the acoustic signal processing device control environment and/or a desired behavior of the electronic device 102 .

As previously explained, during operation the electronic device 102 is configured to detect an acoustic signal "A" (e.g. a voice command, an acoustic signal) by using a plurality of microphones 101 and subsequently received acoustic signals using the acoustic signal processing device 400 in such a way to process that the processed acoustic signals can be used to perform a desired task or activity perform acoustic signal processing using acoustic signal processing system 100 or other electronic device, such as voice recognition, voice communication, voice-activated control of an electronic device, and/or any other useful acoustic signal-activated task or activity. However, depending on the position of the acoustic source 150 in relation to the microphones 101 in the electronic device 102, there is a delay in terms of time when each microphone receives the same acoustic signal. In general, voice communication techniques include any type of 2-way communication process, such as audio chat, video chat, voice call, or other similar communication technique.

2A Figure 12 shows a configuration in which the acoustic source 150 is arranged at a first distance 201A from a first microphone 101A, at a second distance 201B from a second microphone 101B and at a third distance 201C from a third microphone 101C. Based on an assumption of sound wave propagation in the far field, the time delay experienced by microphone 101B and microphone 101C compared to the first microphone 101A located closest to sound source 150 is equal to the distance 202A between the first Microphone 101A and the second microphone 101B in the direction of the received acoustic signal and the distance 202B between the first microphone 101A and the third microphone 101C in the direction of the received acoustic signal. 2 B Figure 12 shows the delays seen by microphones 101A-101C when they detect the same acoustic signals 210A-210C produced by source 150, respectively. In addition to the acoustic signals received by microphones 101A-101C, acoustic signals from other unwanted sources 155 are also received at various different times due to the relative position of each microphone with respect to the various unwanted sources 155. The signals from these unwanted sources 155 can prevent or interfere with the electronic device 102 from detecting the desired information present in the acoustic signal detected from the source 150 .

It can be seen that the delay experienced by one microphone relative to another microphone is equal to the difference in each microphone's distance from the source and the speed of sound (e.g. 340.3 m/s at sea level). As in 2 B is the acoustic signal 210A received by the microphone 101A at time t _A , and thus the delay that the microphone 101B has when it receives the acoustic signal 210B at the time when the microphone 101A hears the acoustic signal 210A receives, equals t _B - t _A . The delay that microphone 101C has with respect to microphone 101A due to the timing of its receipt of acoustic signal 210C compared to when microphone 101A received acoustic signal 210A is t _C -t _A . Thus, the time delay experienced by each microphone compared to other microphones within the geometric array of microphones depends on the relative orientation and position of the acoustic source with respect to each of the microphones and their relative spacing with respect to one another. During the processing of the received acoustic signals by the acoustic signal processing system 100, some additional time delays caused by the signal processing, such as delays in the sample rate, may be introduced.

3H FIG. 12 shows a composite acoustic signal 301 received by a single microphone in the array of microphones present in the electronic device 102. FIG. Composite acoustic signal 301 typically includes a desired acoustic signal provided by a desired acoustic source and multiple other acoustic signals received from other undesired sources. 3B Figure 12 shows one possible configuration of a composite acoustic signal 301, which includes a constant noise acoustic signal 311 originating from a first noise source (e.g. unwanted sources 155), a desired acoustic signal 312 originating from a desired acoustic source (e.g. source 150) is provided, and includes a second acoustic noise signal 313 originating from a second noise source (e.g., unwanted sources 155). Therefore, the composite acoustic signal 301 includes multiple acoustic signals generated in multiple acoustic sources and includes the desired acoustic signal 312 from a desired acoustic source. In an example like this in 3A As shown, the desired acoustic signal 312 is received by a microphone between times t ₁ and t ₅ and has a varying intensity at different frequencies over time. Separately, the constant acoustic noise signal 311 can originate from various common noise sources found in industrial environments, office environments, or even a conference room environment, such as fans, lighting, HVAC units, or other common acoustic sources, and these may, in some cases, contain a low-frequency acoustic signal. The second acoustic noise signal 313 can be acous contain table signals generated by other acoustic sources, such as voices from other people in a room, music played through a local speaker, or other unwanted noise sources. The second acoustic noise signal 313 typically contains acoustic signals that can occur at any frequency at any time. In an example, as in 3B As illustrated, the second acoustic noise signal 313 may range from time t ₂ to t ₄ and may vary in frequency and intensity at different times.

It will be appreciated that the time at which each component of the composite acoustic signal 301 reaches each microphone will vary in at least one property depending on the spacing of the various sources compared to each of the microphones within the array of microphones used in the electronic assembly 102 is present differs. In other words, the time at which the second acoustic noise signal 313 and the desired acoustic signal 312 overlap in time and which is detected by each microphone differs, and thus the phase relationship and the delay for each type of acoustic signal component received by microphone differs to microphone different from each other.

Therefore, a goal of the acoustic signal processing device 400 in the electronic device 102 is to remove as much as possible of the acoustic signal received from the first and second types of noise sources so that the desired acoustic signal 312 can be separated therefrom. After separation, the desired audible signal 312 may then be provided to a software application configured to further process the desired audible signal such that some desired activity may be performed based on the receipt of the desired audible signal 312 . In some embodiments, the desired acoustic signal 312 includes a user's speech, which contains information over the full range of speech, typically extending from 100 Hz to about 8000 Hz.

The design and construction of the microphones in a geometric array in the electronic device 102 may be based on a balance between the need to have a microphone array configuration that is a desired spacing to ensure that the direction of the received acoustic signal is accurate can be determined, as will be explained further below, versus the need to ensure that the signal processing technique (e.g. cardiac waveform and/or beam shape) can preferentially cancel unwanted noise over the full speech range without affecting the signal processing technique does not fail at either the high end or the low end of the frequency range. It is believed that most conventional spatial noise reduction techniques in use today are unable or ineffective to operate at the high end of frequencies due to microphone spacing limitations or constraints, and hence most speech recognition or other similar programs are unable to use effectively the information contained in the upper end of the speech range, say between 4000 Hz and 8000 Hz.

4A-4C 12 are schematic views showing various system-level components that make up the acoustic signal processing device 400 that is configured to process the acoustic signals generated from FIG 5A acoustic signal processing system 100 illustrated. Although that's in 5A Although the acoustic signal processing system 100 illustrated and most of the following discussion describe a configuration in which three microphones 101A-101C are arranged in a planar circular array along the exterior surface 108 of the acoustic signal processing system 100, this configuration should not be considered limiting of the scope of the disclosure provided herein. Other positions, orientations, and numbers of microphones may also be used to implement one or more aspects of the disclosure provided herein.

4A FIG. 4 is a schematic representation of one embodiment of the overall system that may be used to form acoustic signal processing device 400. FIG. The acoustic signal processing device 400 contains an optional microphone amplification element 420, a direction detection element 430, a first signal processor 405, a second signal processor 407, one or more post-processing elements 451, 452 and a signal combining element 414. The signal combining element 414 then represents the processed acoustic signal (e.g. the desired acoustic signal) for a downstream element 415 ready. The downstream element 415 may comprise a software application or other electronic device that uses the processed acoustic signal to perform some desired activity. The various elements and/or components described herein in connection with the acoustic signal processing device 400 may, through the uses tion of various analog and digital electrical components used in combination with a controlling software program or programs executed through the use of processor 118, I/O devices 116 and memory unit 122. In some embodiments, the controlling software program or programs and various components in the acoustic signal processing device 400, such as the direction detection element 430, the first signal processor 405 and the second signal processor 407, through the use of a physical or acoustic command, the received from a user can be placed in an active state. In one embodiment, the controlling software program(s) and the various components within the acoustic signal processing device 400 are in a power saving state until a physical or acoustic command is received from a user. In one embodiment, the controlling software program(s) and various components within acoustic signal processing device 400 are in a power saving state until an acoustic signal having a desired acoustic signal level is received from a user. According to 3A in one example, the one or more controlling software programs compare a received composite acoustic signal 301 with a specified acoustic signal level 321 stored as a parameter in memory to determine whether the received composite acoustic signal 301 exceeds the acoustic signal level 321, and thus whether a desired acoustic signal has been received from a user at one or more points in time. In other embodiments, the one or more controlling software programs are always active in the background when the electronic device 102 is powered on.

4B Figure 12 is a schematic representation of the various circuit elements found in the first signal processor 405 and the second signal processor 407, as discussed below. 4C 12 is a schematic representation of the various circuit elements found in the direction sensing element 430 of the acoustic signal processing device 400, which are also discussed further below.

The optional microphone gain element 420 typically includes microphone signal gain adjustment elements configured to adjust the signal level of an input signal received from each of the microphones in the microphone array. As in 4A As shown, the microphone signal gain adjustment elements 401A-401C are configured to independently adjust the signal level of the input signal received from the microphones 101A-101C, respectively, such that the incoming signal levels provided by the respective microphone are similar when received from the various elements are processed in the acoustic signal processing device 400. A microphone signal gain adjustment element 401 contains analog and digital electrical circuit components which are designed and adjusted to provide a desired gain for an upstream microphone 101 .

According to 4A and 5A the acoustic signal processor 400 is configured to process the acoustic signals received from the plurality of microphones 101A-101C using a first signal processing technique performed by the first signal processor 405 and a second signal processing technique performed by the second signal processor 407. The acoustic signal processor 400 is then configured to combine the signals processed by the first and second signal processors 405 and 407 by using the signal combining element 414 to produce a desired acoustic signal having a high signal-to-noise ratio throughout desired frequency range, such as in the full speech range. A signal-to-noise ratio can be defined and/or measured as the ratio of a desired signal's power to the noise power, and this is often expressed in decibels.

Example of the first signal processor

In some embodiments, the first signal processor 405 is configured to separate unwanted low-frequency noise from the detected acoustic signals received from two or more of the microphones in the geometric microphone array, while the second signal processor 407 is configured to generally separate unwanted higher-frequency noise from the detected acoustic signals received by all microphones in the geometric array.

In general, the first signal processor 405 is configured to remove noise, which is at the lower end of the frequency range of the acoustic signal, from the received composite acoustic signal using a heart curve noise reduction technique. To perform the first signal processing technique, the first signal processor 405 includes or uses portions of the controlling software program and various analog and digital hardware components to perform the perform desired processes described herein. In some embodiments, the first signal processor 405 includes elements formed in a digital signal processor (DSP) module. In some embodiments, the first signal processor 405 is configured to remove or separate the unwanted noise from the desired acoustic signal using a heart-wave signal processing technique. The cardiac waveform signal processing technique performed by the first signal processor 405 is essentially designed to reduce noise received from an off-axis direction relative to the direction of a desired acoustic signal source using a pattern similar to longitudinal radiating cardiac waveform is. 6A-6C show the various heart-curve-based patterns generated using two microphones spaced a fixed distance apart at various different frequencies. As discussed in more detail below, the controlling software used in the acoustic signal processor 400 may utilize two or more microphones in the geometric array to remove unwanted low-frequency noise from the detected acoustic signals using a first-order heart-curve technique, a second order cardioid, hyper cardioid, super cardioid, or other cardioid technique.

In some embodiments, the first order cardiac waveform is generated by the first signal processor 405 through the use of two acoustic signal inputs lying along a direction that is aligned with the direction in which the acoustic signal is received from the acoustic signal source. In one embodiment, the first order cardiac waveform is generated using two acoustic signal inputs received from two of the microphones located in the microphone array. Additionally or alternatively, as will be explained in more detail below, the heart wave pattern is generated by averaging the input signals from two microphones to generate a virtual microphone acoustic signal and then combining the generated virtual microphone acoustic signal and an acoustic signal from one of the other microphones in the geometric array is used to generate the first-order heart waveform. Although other signal processing techniques could be employed with the generation of higher order heart curves from the first signal processor 405, it is nevertheless believed that the use of a first order heart curve for signal processing at low frequencies has advantages over these higher order heart curve signal processing techniques. In general, it is desirable from a power conservation standpoint to use fewer microphones to implement the signal processing techniques disclosed herein. Therefore, a first-order heart-wave generation technique has advantages over other signal processing techniques that require an increased number of microphones to generate the higher-order heart-wave patterns. It can be seen that a large proportion of the power consumption is accounted for by the process of reading out, comparing and then storing in a buffer the acoustic signals and performing other related mathematical pointer operations for each of the microphones in the electronic device at the high sampling rates required to perform these types of signal processing techniques. In another example, assume that the higher order heart wave signal processing techniques require additional computational power and time to process the acoustic signals received from three or more microphones. The additional amount of computing power can thus cause a significant power consumption for a battery-powered power source 130 .

In some embodiments, the first-order heart-curve noise reduction technique employed by the first signal processor 405 may be achieved by summing acoustic signals received from two microphone elements in the geometric array of microphones, with one of the acoustic signals being inverted and is delayed by a period of time with respect to the other acoustic signal before the two acoustic signals are summed. The amount of time delay introduced by the first signal processor 405 depends on the speed of sound and the effective distance between the microphones in the direction from which the desired acoustic signal is received. The first signal processing technique is able to generate a desired heart-wave pattern to reduce unwanted noise received from off-axis orientations, as long as the wavelength is significantly greater than the distance between the two microphones used to generate the heart-wave pattern will. However, the ability to reduce unwanted noise in a heart-wave pattern decreases as the wavelength approaches a proportional distance between the microphones. Therefore, the closer the microphones are to each other, the higher the frequency for which the heart-curve pattern can be maintained.

6A Figure 6 shows an example of a pattern 601 generated by the first signal processing technique using two microphones at a fixed distance from each other (e.g. 70 mm) at a first frequency of approximately 100 Hz. It will be appreciated that a desired acoustic source is preferably placed at the 90° position with respect to the orientation of the polar graph so that the maximum magnitude of the acoustic signal is picked up from that direction, while the magnitude of the acoustic signal picked up from others off-axis directions, falling to zero at the -90° direction (e.g. 180° to the direction of the desired acoustic source). 6A 12 also includes a second pattern 602 generated at a second frequency of approximately 1200 Hz using the same two microphones. It can be seen that changing the frequency from 100 Hz to about 1200 Hz increases the ability of the first signal processing technique to reduce the effect of off-axis noise sources, such as those noise sources located outside of the angular range lying between about 45° and about 135° extends, decreases by an appreciable amount as the frequency of the detected acoustic signal increases. Therefore, geometric microphone arrays that include the distance needed to produce the in 6A-6C pattern shown was ineffective at frequencies exceeding the point at which the heart curve pattern breaks down (e.g. about 1200 Hz) and therefore in this example the first signal processing technique is essentially ineffective for use in connection with acoustic signals, containing frequencies higher than 1200 Hz.

6B 12 shows an example of a third pattern 603 generated by the first signal processing technique at a third frequency of approximately 2000 Hz using the same two microphones used to generate pattern 601. FIG. In this example, the third pattern 603 has a distortion due to the frequency effect on the detected signal such that it forms two lobes 603A and 603B. The third pattern 603 differs significantly from that in 6A illustrated pattern 601, since the ability of the first signal processing technique to reduce off-axis noise sources in the 15° and 165° directions due to lobes 603A and 603B is non-existent, while the first signal processing technique tends to reduce nearly half of the acoustic signal that is received from a desired signal source that is at a 90° position.

6C 8 shows an example of a fourth pattern 604 generated at a fourth frequency of approximately 8000 Hz by the same two microphones used to generate the first pattern 601. FIG. In this example, the fourth pattern 604 of the first order microphone array is also distorted to contain seven lobes 604A-604G due to the frequency effect on the detected signal. Therefore, the ability of the first signal processing technique to reduce off-axis noise sources in sensitive directions of the lobes 604A-604G (e.g., 0°, 45°, 135°, etc.) is minimal, with the first signal processing technique tending to reject less than about 5 dB of the acoustic signal received from a desired acoustic source located at the 90° position.

Although the explanation related to the 6A-6C each relates to the effect that frequency has on the ability of the first signal processing technique to reject off-axis noise sources, similar effects can be obtained by adjusting the hardware design, such as the distance between the two microphones, used to generate the pattern based on a heart curve. For example, a similar pair of patterns 601 and 602 ( 6A ) are generated when the distance between the two microphones is reduced from a distance of 70 mm to about 35 mm using the measuring frequency of about 1100 Hz. Therefore, in some embodiments, the relative spacing between the various microphones in the acoustic signal processing system 100 can be adjusted to desirably set the upper end of the frequency range in which the first signal processing technique can operate. In some embodiments, the upper end of the frequency range of the first signal processing technique is set such that the transition to the frequency range in which the second signal processing technique can perform its desired function (ie, to remove higher frequency noise) is at least partially allowed. The process of adjusting the high end of the first signal processing technique and the low end of the second signal processing technique is described in connection with FIGS 8A-8C explained in more detail below. Alternatively, in some embodiments, the upper end of the frequency range of the first signal processing technique is set to a value at which the generated heart curve-based pattern is not significantly distorted, such as a value at which two or more lobes are generated and/or the fading of the desired acoustic signal provided in the direction of a desired acoustic source (e.g. in 6A-6C at 90°) is less than about 1 to 5 dB. For example, the upper end of the Fre frequency range can be selected so that the wavelength is of the order of the distance between the microphones used to generate the pattern generated by the first signal processing technique. Alternatively, in another embodiment, the upper end of the frequency range is between about 500 Hz and about 4000 Hz, such as in a frequency range of about 1000 Hz to about 4000 Hz. In some examples, the upper end of the frequency range is about 700 Hz, about 1000 Hz, at around 1500 Hz, at around 2000 Hz or even at around 4000 Hz.

It's up again 4B referred; the first signal processor 405 generally includes a microphone input selector 440, a signal delay element 441, a signal inverting element 442, a signal combining element 443 and a signal filtering element 411, a parametric equalizer 412 and a mixer gain value amplifier 413 in the post-processing element 452. 4D 4 shows a method 470 for performing a first type of signal processing technique using the first signal processor 405. During operation, in step 472 the first signal processor 405 receives information regarding the detected direction or direction detection information from the direction detection element 430 based on an evaluation of acoustic Signals received by the geometric array of microphones. The detection direction information then enables the process for generating a heart curve to be performed so that off-axis noise can be rejected, as described above. If according to 5A and 6A in one example, the sensed direction of the desired acoustic source is recognized as being in the 90° direction by an evaluation performed by the direction sensing element 430, then a heart-curve pattern oriented similarly to the patterns 601 and 602 is generated by the first signal processor 405.

Next, in step 473 , the microphone input selector 440 selects two microphones in the microphone array based on the detected direction information obtained from the direction detector 430 . In this example, the first signal processor 405 performs an evaluation of the acoustic signals received from the microphones 101B and 101C as they are aligned in a line 505B extending in the 90° and -90° directions.

After the desired pair of microphones is selected, steps 474 and 475 are completed, which result in the application of the acoustic signal received from the microphone furthest from the acoustic signal source (e.g., microphone 101B) to the signal delay element 441 and a signal inverting element 442 by applying the acoustic signal along the path of channel 440B of the first signal processor 405 to produce a delayed and inverted acoustic signal. The time delay applied by the signal delay element 441 is determined by the known distance between the two microphones based on the known speed of sound. The value of the time delay used can be stored in and retrieved from the storage unit 122 .

Next, at step 476, the "undelayed" acoustic signal received from the microphone closest to the acoustic signal source (e.g., the acoustic signal received from microphone 101C along path 440A) and the delayed and inverted one acoustic signal combined through the use of signal combining element 443 .

Next, in step 477 the combined signals are then optionally filtered through the use of signal filter element 411 in post-processing element 452 . The signal filter element may include a low-pass filter and/or a high-pass filter capable of removing frequencies that are higher and/or lower than the usable signal processing range of the first signal processor 405, such as frequencies at which the heart curve pattern becomes significantly distorted . For example, a low-pass filter frequency can match a frequency of about 1000 Hz to about 4000 Hz, and a frequency of the high-pass filter can be a frequency of about 100 Hz, or, for example, at a frequency of about 700 Hz, or at about 1000 Hz, or even at about 2000 Hz.

After the processes performed in the method 470 are complete, the first signal processor 405 then provides the processed acoustic signal, or further the first processed acoustic signal, to the signal combining element 414, wherein in step 498 the first processed acoustic signal and a second processed acoustic signal obtained from the second signal processing technique are combined. As will be explained in more detail below, in step 499 the combined first and second processed acoustic signals are then applied to downstream electronic inputs directions or elements in the acoustic signal processing system 100 fed.

Based on the geometric layout of microphones used in 5A 1, therefore, acoustic signal sources aligned in the directions parallel to line 505A, line 505B, or line 505C can be more accurately represented by a generated heart-curve pattern. Line 505A can be used when the acoustic signal sources are oriented in directions -30° and 150°. Line 505B can be used when the acoustic signal sources are oriented in the 90° and -90° directions. Line 505C can be used when the acoustic signal sources are oriented in directions 30° and -150°. However, acoustic sources that are not aligned to the directions represented by lines 505A-505C have less accurate noise rejection due to the angular misalignment error created when the misaligned heart curve pattern is generated by the first signal processor 405.

In an effort to reduce angular misalignment error introduced when the selected microphones are not aligned in the precise direction of the acoustic source, a virtual microphone can be used to minimize or effectively eliminate an error introduced by misalignment of the acoustic source generated heart curve pattern in relation to the acoustic source. A virtual microphone can be created by combining the acoustic signals received from two microphones in the array, creating a microphone that is effectively located at a point along a line that runs between the two microphones. The combined acoustic signal generally approximates a range of an acoustic signal that would be received by a virtual microphone positioned between the first and second microphones. The process of combining the acoustic signals may involve averaging or even weighting the two acoustic signals received by the two selected microphones in the geometric arrangement to produce a virtual microphone acoustic signal which is then similarly processed by the first signal processing technique Can be used in the same way as an audible signal from an "actual" microphone. For example, let it be on 5A referred; if the direction detection element 430 determines that the acoustic source is in a direction of approximately 60°, then the first signal processor 405 can average the acoustic signals received by the first microphone 101A and the third microphone 101C to a virtual microphone 520 generate located at the midpoint of the area of line 505A that runs between the two microphones. The first signal processor 405 may then use the generated virtual microphone 520 and a second microphone (e.g., microphone 101B) to complete the process of generating a first-order cardiac waveform. In this example, the first signal processor 405 performs an evaluation of the acoustic signals effectively received by the virtual microphone 520 and the microphone 101B, since these microphones are aligned along the direction 522, which is in the direction 60° and -120° extends. After the desired pair of microphones is selected (e.g., at step 473), the process of generating the heart waveform then involves applying the acoustic signal received from microphone 101B to signal delay element 441 (e.g., at step 474) and to a signal inverting element 442 (e.g., in step 475) by applying the acoustic signal along path 440B to produce a delayed and inverted acoustic signal. The value of the time delay applied by the signal delay element 441 is given by the distance between these two microphones, which in this case corresponds to the distance 521, and by the known speed of sound. The value of the time delay that is used can be stored in and retrieved from the storage unit 122 . It can be seen that the distance 521 is smaller than the distances 510A-510C and thus has an effect on the frequency limit for the first signal processing technique and can therefore be taken into account by the first signal processor 405. Next, then, are the "undelayed" acoustic signal received from the microphone closest to the acoustic signal source (e.g., virtual microphone 520) and the delayed and inverted acoustic signal received from the second microphone 101B , combined together (e.g., at step 476) through the use of the signal combining element 443, and then filtering occurs through the use of the signal filter element 411 in the post-processing element 452. After the processes in the method 470 are performed, the first signal processor 405 can then process the processed acoustic signal including the virtual microphone to the signal combining element 414, at step 498 the first processed acoustic signal and a second processed acoustic signal obtained from the second signal processing technique are then combined. As explained in more detail below, in step 499 the first and second processed acoustic signals can then be combined with other downstream electronic devices genes or elements in the acoustic signal processing system 100 are transmitted.

Therefore, by creating virtual microphones arranged along a line running between two microphones, the maximum angular deviation error produced when the selected microphones are not aligned in the exact direction of the acoustic source can be reduced by at least half the angle formed between the microphones used to create the virtual microphone. For example, the virtual microphone in the previous example would thus effectively have a microphone spaced every 60° from the actual 120° spacing between all three actual microphones 101A-101C. Thus, by using a microphone signal detection technique to generate the virtual microphone, the first signal processing technique can be used to detect acoustic signals present in at least 12 different directions using only 3 microphones. Alternatively, for virtual microphone generation techniques in which a weighted sum of the acoustic signals received from two or more microphones is used, the first signal processing technique could determine an infinite number of possible directions to place the virtual microphone along the line connecting between between the microphones, using only 3 microphones. The weighting values used to generate the weighted sum of the acoustic signals may be based on a comparison of the sensed direction obtained from the direction sensing element 430 and a relationship (e.g. relative angles) to one of the six directions parallel to the Lines are 505A, 505B or 505C. As previously indicated, therefore, by using the virtual microphone technique, fewer microphones are required to perform the first signal processing technique and thus less power is consumed by the electronic equipment in performing that signal processing technique.

In some embodiments for creating a virtual microphone, it is desirable to select two microphones arranged along a line that is substantially perpendicular to the direction in which an acoustic signal is received in order to create the virtual microphone. Furthermore, by increasing the number of actual microphones present in the geometric array of microphones, the need to create virtual microphones can be reduced as the error is reduced.

Alternative example of the first signal processing

Alternatively, in electronic device configurations that are not constrained by electrical power constraints (e.g., electrical power is obtained from an electrical outlet or a large battery) and/or processor speed and other processor-related resources, it may It may be desirable to remove or separate a noise contained in the frequency domain of the acoustic signal by employing a heart curve noise reduction technique in which a higher order than first order heart curve is used based on a heart curve signal processing technique. In some embodiments, the first signal processor 405 is configured to remove noise using a signal processing technique based on a second or higher order cardiac waveform. In one configuration, the cardioid signal processing technique performed by the first signal processor 405 is adapted to reduce noise received from an off-axis direction relative to the direction of the acoustic signal using three or more microphones within the geometric arrangement will. The controlling software used in the acoustic signal processor 400 can thus use three or more microphones and/or generated virtual microphones oriented in a desired direction to remove unwanted noise from the detected acoustic signals using a heart curve pattern to separate

In some embodiments, the first signal processing technique is configured to generate higher order heart curve patterns through the use of three or more acoustic signal inputs arranged along a direction that is on the same line as the direction in which the acoustic signal originates from the acoustic signal source is received. In one embodiment, the higher order cardiac waveform is generated using the three or more acoustic signal inputs received from the microphones located in the microphone array. Additionally or alternatively, the higher-order heart-curve patterns may be generated by combining (e.g., averaging or weighting) the inputs from two or more microphones to generate a virtual microphone acoustic signal by combining the generated virtual microphone acoustic signal and a acoustic signal from one of the other microphones or other virtually generated microphones can be used to generate the higher order cardiac waveform.

6B FIG. 12 shows an example of a configuration of a geometric arrangement that can be employed by the first signal processing technique to generate a desired cardioid pattern. In this example, the acoustic signal processing system 100 includes seven microphones 601A-601G arranged in a geometric array that includes six microphones arranged along the outer surface 108 (e.g., microphones 601A-601G) and one at the midpoint 103 of the microphone arranged on the upper surface 106 (e.g. the microphone 601G) of the electronic device 102 . However, those skilled in the art will recognize that other configurations of a geometric arrangement can be used in the first signal processing technique to generate higher order cardioid patterns. In one example, the geometric arrangement may include five microphones arranged in a configuration that includes four microphones lying along outer surface 108 and one microphone at midpoint 103 . In some embodiments, the geometric array includes multiple microphones arranged only along the outer surface 108 of the electronic device 102, such as an array containing only six microphones 601A-601G ( 6D ) or other convenient odd or even number of microphones in the geometric array.

During operation, the first signal processor 405 receives information on the direction of the acoustic signal from the direction detection element 430, announcing that acoustic signals are being received from a direction 605 (i.e. 0° direction). The first signal processor 405 then determines, using the controlling software and the microphone input selector 440, that the microphone 601G, a virtual microphone 620A and a virtual microphone 620B are required to perform the first signal processing technique, since these microphones are aligned along the direction 605 . In this example, the process of generating the desired heart curve pattern can be completed once the controlling software has generated virtual microphone 620A by averaging the acoustic signals received by microphones 601A and 601F and has generated virtual microphone 620B by averaging the acoustic signals received by microphones 601C and 601D. In this example, virtual microphone 620A is at the midpoint of line 621A and virtual microphone 620B is at the midpoint of line 621B.

Subsequently, the first signal processor 405 then uses a cascaded heart-wave generation process to perform the first signal processing technique. The cascarded heart wave generation process begins by applying the acoustic signal received from microphone 601G to signal delay element 441 (e.g. step 474) and signal inversion element 442 (e.g. step 475) to generate a first delayed and inverted acoustic signal . The value of the time delay used can be stored in and retrieved from memory unit 122 for this known microphone configuration. The "undelayed" acoustic signal received from the virtual microphone 620A and the first delayed and inverted acoustic signal received from the second microphone 601G are then combined (e.g. step 476) using the signal combining element 443, to generate a first combined cascarded acoustic signal that is stored in memory.

The cascarded heart wave generation process then proceeds to generate a second combined cascarded acoustic signal by applying the acoustic signal received from the second virtual microphone 620B to the signal delay element 441 (e.g. step 474) and the signal inversion element 442 (e.g. step 475), to generate a second delayed and inverted acoustic signal. The "undelayed" acoustic signal received from microphone 601G and the second delayed and inverted acoustic signal are then combined (e.g., step 476) using signal combining element 443 to produce a second combined cascarded acoustic signal that is stored in memory.

The cascarded heart wave generation process then directs the second combined cascarded acoustic signal to signal delay element 441 (e.g. step 474) and signal inversion element 442 (e.g. step 475) to generate a first combined delayed and inverted acoustic signal. The value of the time delay that is used can be stored in and retrieved from the storage unit 122 . The first combined cascarded acoustic signal and the first combined delayed and inverted acoustic signal are then combined (e.g., step 476) using signal combining element 443, and filtering is then performed using signal filter element 411 to extract the full combined cascarded generate an acoustic signal.

After generating a complete combined cascaded acoustic signal, the first signal processor 405 then performs the complete combined cascarded acoustic signal to the signal combining element 414 where in step 498 the complete combined cascarded acoustic signal and a second processed acoustic signal obtained from the second signal processing technique are combined. Then, at step 499, the combined signals are transmitted to other downstream electronic devices or elements in the acoustic signal processing system 100, as discussed in more detail below.

In cases where the acoustic signal source is oriented in a direction that contains three "actual" microphones, such as direction 606 in 6D , the cascarded heart wave generation process can be performed using only three microphones and thus without virtual microphones. In this example, the first signal processor 405 then determines, using the controlling software and the microphone input selector 440, that microphone 601F, microphone 601G and microphone 601C are needed to perform the first signal processing technique. Then, in a manner similar to that previously discussed, the first signal processing technique uses the three microphones and process steps previously described to similarly generate the complete combined cascarded acoustic signal.

In electronic device configurations that include an array of microphones disposed substantially along the outer surface 108 of the electronic device 102, such as configurations in which there are no centrally located microphones, a virtual microphone may be used in place of a centrally located microphone by averaging the acoustic signals received by two or more microphones of the geometric arrangement. According to 6D In one example, a virtual microphone can be created at midpoint 103 by averaging the acoustic signals received from microphones 601B and 601E. Therefore, in some cases, a cascaded heart wave generation process can be performed using only two "real" microphones and one virtual microphone. Therefore, compared to the previous example, the first signal processor 405 can determine, using the controlling software and the microphone input selector 440, that the microphone 601F, the microphone 601C, the microphone 601B and the microphone 601E are needed to perform the first signal processing technique. In this example, microphones 601B and 601E are used to create the center virtual microphone, and then the created virtual microphone and microphones 601F and 601C are used to perform the first signal processing technique. Then, similarly as previously discussed, the first signal processing technique uses these "real" and virtual microphones and the process steps previously described to similarly generate the complete combined cascarded acoustic signal.

Direction acquisition process and elements

Let it be 4A and 4C referred; During operation of the acoustic signal processing device 100, the direction detection element 430 is designed to determine information about the detected direction and to forward it to the first signal processor 405 and the second signal processor 407. 4B Figure 4 shows a method 480 for performing a technique for direction detection using direction detection element 430. Direction detection element 430 generally includes or uses portions of the controlling software program and various analog and digital hardware components to perform one or more of the processes described herein. In some embodiments, direction sensing element 430 includes elements formed in a digital signal processor (DSP) module. In some embodiments, the direction detection element 430 includes an equalizer element 460, an RMS processing element 467, a delay evaluation element 462, a direction determination element 463, a coefficient generation element 464, a direction selection element 465 and a direction supply element 466.

During operation, at step 482, after receiving the microphone inputs from the optional microphone gain element 420, during step 481, the acoustic signals received from the respective microphones 101A-101C are respectively pre-processed. The pre-processing steps may include processing the acoustic signal using an equalization element 460 such that some desired frequency range associated with the speech range can be extracted or separated from other undesired frequency ranges prior to processing by subsequent direction detection elements. In some configurations, the equalization element 460 may include parametric equalizers 461A-461C, each configured to appropriately process the acoustic signal received from the respective microphone 101A-101C. In some embodiments, the equalization element 460 is configured, preferably passing frequencies within the full speech range and routing to a or to enable multiple downstream components.

Next, at step 483, at least one of the inputs from a microphone in the array of microphones is provided to a root mean square (RMS) processing element 467 which is used to detect constant acoustic noise signals 311 and in some cases to remove contained in the composite acoustic signals 301 received by the microphones, respectively. The RMS processing element 467 may employ an RMS threshold evaluation element 468 that contains and/or monitors and stores information related to the time history of the level of the constant acoustic noise signal 311 . The RMS processing element 467 can be configured to detect the current level of the received constant acoustic noise signal 311 and further to receive an input regarding historical constant acoustic noise signal levels from the RMS threshold value evaluation element 468, so that an unwanted background noise is not used in the direction detection algorithm and thus not affecting the results of the direction detection element. In some embodiments, the RMS threshold evaluator 468 uses the measured constant acoustic noise level and compares it to a received acoustic signal to determine whether the incoming acoustic data should be used in determining its direction of arrival, and thus the received acoustic data not part of the noise in the environment 110 ( 1A ). The RMS processing element 467 can thus be used to reduce the risk of background music or crowd noise affecting the directional evaluation employed by the first and second signal processing techniques.

Next, in step 484, the delay evaluation element 462 receives the acoustic signals processed by the equalization element 460 and analyzes the acoustic signals to determine the relative delays in receiving the acoustic signal experienced by each of the microphones. In some embodiments, the delay evaluator 462 evaluates each of the received acoustic signals as a function of time to determine which attributes in each of the received acoustic signals are common to each of the received acoustic signals by the relative delay that occurs in each microphone , to investigate. In a configuration, as in 2 B As shown, the delay evaluator 462 may sense a change in sound level or signal pattern as a function of time to determine the delay experienced by each microphone. In one example, the one or more controlling software programs may not use a cross-correlation technique to determine the delay in receiving the acoustic signal that occurs at the respective microphone, although the received acoustic signal may be part of a composite acoustic signal generated by received by the microphones. In another example with back reference to 3A the one or more controlling software programs can compare the received composite acoustic signals 301 from the microphones in the geometric arrangement with the predetermined acoustic signal level 321 to determine whether at least one area related to the portion of the received composite acoustic signal 301 exceeds the acoustic signal level 321 for at least one of the microphones, and thus a desired acoustic signal received by a user at one or more points in time is extracted based on its previous signal level (e.g. dB level), in some cases additionally employing a cross-correlation technique becomes. In some cases, to reduce the power consumption of the device and/or to reduce the number of tasks to be performed by the processes, it may be desirable to only use the cross-correlation technique when the sound level of the received composite acoustic signal 301 exceeds the acoustic signal level 321 . It can be seen that the microphone that is closest to the acoustic signal source has zero time delay and that the other microphones have a time delay that depends on the distance from the microphone to the closest microphone in the direction in which the acoustic signal is received from the acoustic source. In some cases, the controlling software program ensures that the delay times are all positive and stores the relative delays in the electronic device's memory.

At step 485, after determining the relative delay of each of the microphones, the direction determining element 463 then determines the direction from which the acoustic signal is being received using one or more portions of the controlling software program. Although it is possible to perform various sophisticated mathematical analysis techniques to determine the exact position of the acoustic source with respect to the electronic device 102, these highly analytical direction detection processes have been found to require a significant amount of computational power and time, and thus a significant burden for the performance source 130 can cause. Incorporating these types of highly analytical direction sensing processes would greatly increase the cost and complexity of a consumer electronic device capable of performing these types of direction sensing processes.

Therefore, in some embodiments, the direction determination element 463 uses a less analytically intensive and less power intensive statistical grouping approach to determine the direction of the acoustic source. Although in some cases the direction determination element 463 is able to detect the exact direction of the acoustic source from the electronic device 102, in most cases the direction determined by the direction determination element 463 has an error which is at its maximum Size depends on the sampling rate of the analyzer program, the spacing of the microphones and also the size of the direction intervals selected and used by the one or more controlling software programs to determine the closest direction with respect to the direction of the acoustic source. Therefore, an acoustic signal direction determined by the statistical evaluation performed by the direction determining element 463 is referred to herein as a “quasi-direction”.

During the direction determination process, the direction determination element 463 uses the relative delay times obtained from the delay evaluation element 462 to determine the direction of the received acoustic signal. The controlling software program and the direction determination element 463 can break up the pattern of the geometric array of microphones into subdivided areas. The number of subdivided areas or grading areas is typically dependent on the number of microphones included in the geometric array of microphones and the minimum width of the beam. In an example, the in 1B illustrated electronic device 102 can be divided into five sectors or five angular regions. 5B 1 is a top view of the acoustic signal processing system 100 shown in FIG 5A 1, which further includes a plurality of direction-sensing regions that can be used to determine the direction of an acoustic source. In one example, the electronic device 102 may be divided into three sectors 531A, 531B, and 531C, each containing a center region that is 120° angularly spaced when measured from a corner point 536. FIG. In one configuration, the vertex 536 is at the geometric center of the geometric array of microphones. However, vertex 536 may be at other desired geometric locations with respect to the geometric array of microphones to simplify the evaluation performed by the controlling software program. Next, the direction determination element 463 determines from which area of the geometric arrangement of microphones the acoustic source is currently the smallest distance. In one example, the direction determination element 463 is capable of when the acoustic source is at an angle of approximately 35° with respect to the in 5B Illustrated electronic device 102 is arranged to first determine that the acoustic source is within sector 531A, which extends from 0° to 120°, since delay evaluation element 462 has determined that microphone 101A is the microphone with the smallest distance to the acoustic source, and microphone 101C has a smaller delay time than microphone 101B due to their relative positions with respect to the acoustic source.

After determining the area of the geometric array of microphones to which the acoustic source is closest, the direction determination element 463 then determines to which directional section (or subdivided area) in the determined area the direction of the acoustic source is closest, so that a closest Quasi-direction can be determined. The directional bins are generated by dividing the angular domain or sector into a desired number of sub-domains that meet the specifications for the desired accuracy and computational power. For example, each of sectors 531A, 531B, and 531C4 can be divided into subdivided areas, each separated by a 30° interval. In one example, the four-section configuration of sector 531A can be subdivided such that the edges of the section have a known quasi-direction, such as directions 0°, 30°, 60°, 90°, and 120°, the edges between the four generated sections are. Thus, the angular distance generated for the defined section or grading area (e.g., the 30° section) is between a first known direction and a second known direction. In one example, the first direction is equal to the 0° direction and the second direction is equal to the 30° direction, the first direction extending from the vertex 536 through a region of a first microphone 101A (e.g., geometric center of the microphone) and the second direction extends from the vertex 536 in the 30° direction.

In one embodiment, the direction determination element 463 first calculates ratios of various time delays (e.g., first non-zero delay/second non-zero delay) obtained from the delay evaluator element 462 to determine which of the quasi-directions the direction of the acoustic source is closest to, and compares these calculated ratios to angular time delay ratio data stored in memory unit 122 . The stored data for the wrap time delay ratio includes previously calculated data generated by calculating a ratio of the expected delay times that would occur in the microphones if the acoustic source were at the edges of the sections connecting each of the quasi-directions within a particular area of electronic device 102. Therefore, using the previous example, if the acoustic source is disposed at an angle of approximately 35° with respect to the electronic device 102, then the direction determination element 463 determines based on a calculated ratio of the delay time experienced by the microphone 101C to the delay time experienced by microphone 101B that the calculated ratio is closer to a stored angular time delay ratio associated with the 30° quasi-direction than for any other stored angular time delay ratios associated with the other 60° quasi-directions, 90° or 120° are related. Alternatively, the direction determiner 463 may determine that the calculated ratio is closer to a stored angular time delay ratio associated with the 30° quasi-direction by determining that the calculated ratio is in a range equal to half the section size on each side of the Quasi-direction corresponds. In this example, the direction determination element 463 may compare the calculated ratio to the stored angular time delay ratio associated with the directions 15°, 45°, 75° and 105° and may then determine that the calculated ratio is between the stored angular time delay ratios corresponding to the 15th ° and 45° correspond. Therefore, the acoustic source is arranged in the 30° quasi-direction with high probability.

In the somewhat rare event that the direction determination element 463 detects that the calculated ratio exactly matches a stored angular time delay ratio, the controlling software program need not continue the process of determining that the calculated ratio is between the stored time delay ratios.

It can be seen that the process of determining the direction of the acoustic source is significantly simplified over the mathematically intensive iterative process of determining the exact position of the acoustic source using a more conventional analytically intensive and power intensive theorem. The greatly simplified statistical approach to determining source direction also reduces the performance requirements that processor 118 must meet to perform these tasks.

In some embodiments, direction determination element 463 determines a direction of a received acoustic signal by first determining the relative time delays experienced by each microphone and then comparing each of the relative time delays to a plurality of stored angular time delays. Each of the plurality of stored angular time delays stored in memory can be associated with a direction oriented relative to the non-linear array of microphones. Thus, for example, each quasi-direction can be assigned a stored angular time delay for each microphone, for example 0°, 30°, 60°, 90° or 120° for quasi-directions. However, it is believed that using the ratio of the expected delay times may be advantageous in certain geometry configurations. For example, using the ratio when the acoustic source is placed in a 3-D space at an angle relative to a plane containing a planar array of microphones may be useful because of the inherent comparison of the relative delays between microphones caused by the ratio are provided compared to other techniques that only compare delay times with the stored angular delay times.

However, since the accuracy of the time delay measurements determined by the delay evaluation element 462 is also limited by the number of samples that can be collected by the processor 118 within the actual time delay experienced by the individual microphones, there will be some Uncertainty in the determined time delay values. The accuracy of the time delay is thus limited by the sampling frequency and the distance between microphones. Because each of the samples is acquired sequentially in time at a desired sampling frequency, it contains a portion of a received acoustic signal. The distance between the microphones and the sampling frequency must therefore be large enough so that at least two samples can be obtained within the time in which the reception of the acoustic signal is delayed, without an increase in sampling or upsampling taking place. It can be seen that the process of increasing the sampling becomes significant drain of processor resources as well as electrical power required to perform this task. For example, if the processor samples at a frequency of 48kHz (e.g. 21µs per sample) and the microphones are 70mm apart, 10 samples can be taken by the processor within the delay time, whereas a microphone spacing of 14mm would only allow the Processor captured two samples within the delay time. The ambiguity in the determined time delay values, due to the often small number of samples and due to the noise contained in the received acoustic signals, can introduce jitter between the determined source position states that affect the ability of the direction determination element 463 to determine a probable quasi-direction to identify and determine. High rate oscillation between the detected source position states can affect the ability of the signal processing technique to perform its desired function.

In some embodiments of step 485, the controlling software program evaluates the frequencies at which the various detected directions are selected by the direction determination element 463 to determine the most likely detected direction of the acoustic source. In one example, the controlling software program compares the frequency with which various detected directions are selected by the direction determiner 463 over a period of time, and then selects the direction that has the highest frequency for that period of time as the detected direction. Determining the most likely determined direction can be performed in a moving average fashion, after which each determined direction within the sliding time period can be taken as a "vote" over which is summed to determine which direction has the most votes over the current sliding time duration. The frequency with which each particular determined direction is determined may involve evaluating two or more audio data samples taken by the processor at the data sampling rate (e.g., 48 kHz sampling rate). This process can reduce the amount of jitter that can be seen in the output of direction determination element 463 .

Let it be 4G and step 486; in an effort to minimize the ambiguity for the determined quasi-direction obtained from the direction determination element 463, the coefficient generation element 464 is used to apply various weights to a number of the directions that are close to the determined quasi-direction, determined during previous steps. 5C FIG. 12 graphically shows some of the coefficient values that can be used in our example from before to weight the determination such that an acoustic source oriented at an angle of 35° to the electronic device has a quasi-direction that is 30° amounts to. In this example, the 30° quasi-direction may be given a coefficient value of V ₁ and the neighboring 0° and 60° quasi-directions may be given a coefficient value of V ₂ and V ₃ , respectively, which correspond to the probability that the actual direction averaged over some acoustic signal data samples runs along the respective quasi-directions. The coefficient values V ₁ , V ₂ , and V ₃ can be fixed to some value that is a percentage of a full scale, and therefore can have a value that takes on real numbers between zero and one, for example. The coefficient values may be determined based on how close the determined quasi-direction is to a stored angular time delay ratio and/or by a moving average obtained over a number of samples or over a specified period of time.

Next, in step 487, the direction selector 465 then uses the determined coefficient values to determine the likely quasi-direction. The determined coefficient values are used to weight and thus attenuate perceived jitter at the output of the direction determination element 463 when determining and subsequently refining the determined quasi-directions for all received acoustic signal data samples or some acoustic signal data samples. In some embodiments, the determined coefficient values for each of the likely directions are summed over the sample period or delay period, and the quasi-direction that has the highest total sum over the period is selected as the likely quasi-direction. In some configurations, it may also be useful to give all coefficients associated with non-probable directions a zero or negative coefficient value to reduce the probability that those directions will be selected in this step.

Then, in step 488 , the direction supply element 466 supplies the direction detection information containing the determined direction or quasi-direction to the first signal processor 405 and the second signal processor 407 . The direction detection information is then processed by the first signal processor 405 and the second signal processor processor 407 respectively during steps 472 and 491 for further use or processing.

Example of the second signal processor

According to 4A and 4B For example, the second signal processor 407 is generally configured to separate a desired acoustic signal from unwanted noise that may occur at the high end of the acoustic signal frequency range by converting the acoustic signals received from the microphone array using a beamforming noise reduction technique are processed. 4F Figure 4 shows a method 490 for performing the beamforming noise reduction technique or a second signal processing technique using the second signal processor 407. The second signal processor 407 includes or uses portions of the controlling software program and various analog and digital hardware components to perform the desired processes described herein . In some embodiments, the second signal processor 407 includes elements formed in a digital signal processor (DSP) module. The beamforming processing technique performed by the second signal processor 407 is generally used to separate the noise received from an off-axis direction using an acoustic signal delay and summation technique, with acoustic input signals emerging from the geometric arrangement Microphones are obtained are used. In some embodiments, the beamforming process is accomplished by determining a delay for each microphone based on the direction or quasi-direction determined by the direction determination element 430 and the received by each of the microphones within the microphone array delayed acoustic signals are summed up so that the acoustic signals received by the microphone from the acoustic source are each in phase and thus add constructively when combined.

In some embodiments, the second signal processor 407 uses the received direction information or quasi-direction information for a first period of time that is longer than the time it takes for the direction detection element 430 to update the direction information or quasi-direction information (e.g., a second time or time of 1 second). In this case, the rate at which the time delays are updated during the beamforming process is less than the rate at which the direction detection element 430 is able to update the direction information or quasi-direction information, thereby using a significant amount of computing power, battery power and Time to be expended by the electronic device 102 is reduced. Using this process can be helpful to smooth the final processed acoustic signal, which is generally not achieved if the rate at which direction is updated is too fast.

7A Figure 7 shows an example of a first beamforming pattern 701 generated by the second signal processing technique using the plurality of microphones spaced a specified distance apart (e.g. 70 mm) at a first frequency of approximately 100 Hz. It can be seen that a desired acoustic source is at the -30° position relative to the orientation of the polar graph. However, at this low frequency, the beamforming technique has no ability to separate off-axis noise from the desired acoustic source, as shown by the circularity of the beamforming pattern. 7B 12 shows an example of a second beamforming pattern 702 generated by the second signal processing technique at a second frequency of approximately 1300 Hz using the same microphones used to generate the first beamforming pattern 01. In this example, the second beamforming pattern 702, generated at a frequency of 1300 Hz, has a nearly heart-shaped waveform pointing in the direction of the desired acoustic source (e.g., -30°). 7C 12 shows an example of a third beamforming pattern 703 generated at a third frequency of approximately 5000 Hz by the same microphones used to generate the first beamforming pattern 701. FIG. In this example, the third beamforming pattern 703 is also distorted to contain eight lobes 703A-703H due to the effect of frequency on the detected signal. The ability of the second signal processing technique to reject or reduce off-axis noise sources in all but three of eight off-axis directions (for example, lobes 703D-703F) at this frequency is significant, while lobe 703A, which is aligned in the direction of the acoustic source picks up most or even all of the acoustic signal received from a desired acoustic source located at the -30° position. As previously indicated, in some embodiments the relative spacing between the various microphones in the acoustic signal processing system 100 can be adjusted such that the frequency range over which the second signal processing technique can operate can be desirably determined. at For example, the lower end of the frequency range in which the second signal processing technique is able to adequately reduce off-axis noise can be selected to be between 1000 Hz and about 4000 Hz, while on the other hand the second signal processing technique is at the upper end of the frequency range, say at about 4000-8000 Hz, is capable of forming a pronounced lobe (e.g. lobe 703A) in the direction of the acoustic source, as is similarly shown in FIG 7C is shown.

It is now open again 4A and 4B referred; the second signal processor 407 generally includes beamforming signal delay elements 431A-431C, a signal combining element 432 and a signal filtering element 408, a parametric equalizer 409, a mixing gain amplifier 410 located after the processing element 451. It's up again 4F referred; during operation, the second signal processor 407 receives direction detection information from the direction detection element 430 (e.g. step 491) based on acoustic signals received by the geometric array of microphones. The direction detection information then allows the beamforming process to proceed in the desired direction and efficiently reduces the amount of off-axis noise in the acoustic signals received after traversing the various components in direction detector 420, as previously described. The beamforming process is generally performed using the acoustic signals received from all microphones and using the direction detection information received from the direction detection element 430 .

In steps 492-493, in some embodiments, the controlling software program performs an evaluation on the received direction detection information obtained from the direction detection element 430 and then determines based on the detected direction or quasi-direction how the desired delay each of the acoustic signals already being received by the microphones, this being done on the basis of specific direction-time-delay information stored in the storage unit 122. The direction-time-delay information stored in storage unit 122 may include a table of all possible directions or quasi-directions that direction-sensing element 430 applies to the second signal processing technique, and may include all time-delay values associated with the possible directions or quasi-directions for each of the microphones are related. Therefore the table contains a time delay value for each microphone for each of the possible directions or quasi-directions. In an example, as in 5A as shown, the multiple directions include all directions that have at least one component of the actual direction within a plane that is parallel to the top surface 106 ( 1 ) from all different angles (e.g. 0° to 360°). The time delay used by the signal delay elements 431A-431C is determined by the known distance between the microphones in the direction from which the acoustic signal is received and the known speed of sound.

Next, at step 494, the delayed acoustic signals are combined together using signal combiner 432. Subsequently, in step 495 the delayed acoustic signals are optionally filtered using the signal filter element 408 in the post-processing element 451 . The signal filter element may include a high-pass filter capable of removing frequencies lower than the usable range of the second signal processor 407 . For example, the high pass filter may be configured in frequency to pass frequencies greater than about 1000 Hz. The appropriately delayed and combined acoustic signals then add constructively such that the signal-to-noise ratio of the acoustic input of the acoustic source is improved compared to the off-axis noise produced by unwanted noise sources.

As previously indicated, after the first and second signal processing processes are complete, the processed acoustic output signal from the first signal processor 405 can be further processed using post-processing elements 452 (step 477) and the processed acoustic signal output from the second signal processor 407 can be processed using of the post-processing elements 451 (step 495) are further processed before they are combined with one another by the signal combining element 414 in step 498. The post-processing elements 451, 452 may each include one or more amplifiers capable of adjusting the signal levels of the processed acoustic signals prior to their combination.

8A-8C show examples of various patterns generated at different frequencies using a first signal processing technique and a second signal processing technique will. The to generate the in 8A-8C The first and second signal processing techniques employed in the pattern shown use a geometric microphone array containing three microphones spaced at a fixed spacing of 70 mm. 8A Figure 8 shows an example of a pattern 801 generated by the first signal processing technique using a first order heart curve based signal processing technique and a pattern 802 formed by the second signal processing technique at a first frequency of approximately 100 Hz. It can be seen that the desired acoustic source is located at a 30° position relative to the polar graph orientation, such that the maximum magnitude of the acoustic signal at that frequency is picked up by the first signal processing technique, while the second signal processing technique assists in reducing all acoustic signals is inefficient in all directions at this frequency. Therefore, the noise reduction provided by the first signal processing technique can be used preferentially to "clean up" the acoustic signal for the low-end frequencies.

8B Figure 8 shows an example of a pattern 811 generated by the first signal processing technique using a first-order heart curve based signal processing technique and a pattern 812 generated by the second signal processing technique at the frequency of about 1600 Hz. It can be seen that changing the frequency from 100 Hz to approximately 1600 Hz improves the ability of the second signal processing technique to reduce off-axis noise sources, and as in 8B shown, it is similar to the heart curve pattern generated at this frequency by the first signal processing technique. Since the noise reduction results obtained by both signal processing techniques are similar at 1600 Hz, a frequency at or near 1600 Hz can be selected and used by the controlling software as a crossover point for the frequency range at which those obtained from the first signal processing technique Results at frequencies below the crossover point are used and the results obtained from the second signal processing technique are used at frequencies above the crossover point. In some embodiments, it is desirable to select the high-pass filter and low-pass filter used in the acoustic signal processing device 400 such that the upper end of the frequency range of the first signal processing technique is fixed at a point that is at or at least close to this crossover frequency. and that the lower end of the frequency range of the second signal processing technique is fixed at a point at least close to this crossover frequency.

8C Figure 8 shows an example of a pattern 821 generated by the first signal processing technique using a first order heart curve based signal processing technique and a pattern 822 generated by the second signal processing technique at a third frequency of approximately 6500 Hz. It can be seen that the desired acoustic source is at a 30° position relative to the orientation of the polar graph, such that the maximum magnitude of the acoustic signal at that frequency is picked up by the second signal processing technique (ie, pattern 822), while nearly all of acoustic signal at this frequency in directions of the acoustic source is rejected by the first signal processing technique, and the off-axis sidelobe directions (e.g. at angles -38°, -95°, 98° and 155°) are actually favored over the direction of the acoustic source. Therefore, the noise reduction provided by the second signal processing technique can be used preferentially to "clean up" the acoustic signal at the high-end frequencies.

In step 499 the signal combining element 414 then supplies the processed acoustic signal (e.g. the desired acoustic signal) to a downstream element 415 . As previously indicated, the downstream element 415 may be a software application or other electronic device that uses the processed acoustic signal to perform some desired action. The downstream element 415 may be an electronic component that is in direct communication or wireless communication with the signal combining element 414 that is located in the I/O device 116 . In one configuration, the downstream element 415 may be an electronic component that is arranged in the acoustic signal processing system 100 . In another configuration, the downstream element 415 may be an electronic component located in an electronic device external to the acoustic signal processing system 100 . Examples of an external electronic device include a wireless speaker, video camera device, keyboard, smart phone, speakerphone, home automation device, or other appropriate electronic device arranged to connect to one or multiple electronic components provided in the acoustic signal processing system 100.

One or more of the embodiments of the disclosure provided herein may be implemented as a program product for use in connection with a computer system. The one or more programs of the program product define functions of the embodiments (including the methods described herein) and may be embodied on a variety of computer-readable storage media. Illustrative computer-readable storage media include, but are not limited to: (i) non-writable storage media (e.g., read-only memory devices in a computer, such as CD-ROM disks readable by a CD-ROM drive, a flash -Memory, ROM chips or any type of non-volatile semiconductor memory) on which information can be permanently stored; and (ii) writable storage media (e.g., floppy disks in a floppy disk drive or a hard disk drive, or any type of random access semiconductor memory) on which alterable information is stored.

The invention has been described above with reference to specific embodiments. However, one skilled in the art will recognize that various modifications and changes can be made therein without departing from the broader spirit and scope of the invention as set forth in the appended claims. The foregoing description and drawings are, therefore, to be regarded in an illustrative rather than a restrictive manner.

Claims (10)

An electronic device (102) comprising: a non-linear array of microphones (101, 121) including at least three microphones; a direction detection element (430) configured to determine a direction of a first acoustic signal (A) received by the microphones (101, 121) in the non-linear array of microphones from an external acoustic source (150). by using a controlling software program executed by a processor (118) in the electronic device (102); a first signal processor (405) configured to remove a first portion of an acoustic signal (313) generated in a second acoustic source (155) from a composite acoustic signal (301) by using the controlling software program, the first Signal processor (405) has: a microphone selection element (440) configured to select a first microphone and a second microphone from the non-linear array of microphones (101, 121) based on the determined direction received from the direction detection element (430); a signal delay element (441) configured to receive the composite acoustic signal (301) from the second microphone and to generate a delayed acoustic signal after a predetermined delay time has elapsed; a signal inversion element (442) configured to receive the delayed acoustic signal and to generate an inverted acoustic signal based on the delayed acoustic signal; and a first signal combiner (414) configured to combine the inverted acoustic signal and the composite acoustic signal received from the first microphone and generate a first combined acoustic signal; a second signal processor (407) configured to remove a second portion of the acoustic signal (313) generated in the second acoustic source (155) from the composite acoustic signal (310) by use of the controlling software program, the second Signal processor has: a signal delay element (431A-431C) configured to receive the composite acoustic signal from at least two of the microphones (101, 121) in the non-linear array of microphones and a second delayed acoustic signal for each of the at least two microphones (101 , 121) based on the determined direction received from the direction detection element (430); and a second signal combining element (432) configured to combine each of the second delayed acoustic signals generated by the signal delay element (431A-431C) and to generate a second combined acoustic signal; and a third signal combining element (443) configured to combine the first combined acoustic signal and the second combined acoustic signal to generate a third combined acoustic signal. The electronic device (102) after claim 1 wherein the third combined acoustic signal is transmitted to an electronic component (415), the electronic component (415) being configured to perform an acoustic signal processing activity based on the receipt of the third combined acoustic signal. The electronic device (102) after claim 2 , wherein the electronic component (415) in a wireless speaker, a video camera device, a keyboard, an intelligent ten telephone, a loudspeaker telephone or a home automation device. The electronic device (102) after claim 2 wherein the acoustic signal processing activity performed by the electronic component (415) includes a software program configured to perform a speech recognition, a speech capture, or a speech communication process. The electronic device (102) after claim 2 , the direction determined being based on a comparison of a delay time experienced by each of the microphones (101, 121) upon receipt of the first acoustic signal. The electronic device (102) after claim 1 , wherein the first microphone comprises a virtual microphone generated by averaging an acoustic signal input received from a third microphone and a fourth microphone located in the non-linear array of microphones (101, 121). The electronic device (102) after claim 6 , wherein the acoustic signal input comprises the composite acoustic signal. The electronic device (102) after claim 1 wherein the non-linear array of microphones (101, 121), the first signal processor (405), the second signal processor (407) and the third signal combiner (443) each derive power from a battery. The electronic device (102) after claim 1 , further comprising: a first filter (411) configured to filter the first combined acoustic signal prior to combining the first combined acoustic signal and the second combined acoustic signal to generate the third combined signal, wherein filtering the first combined signal comprises : removing from the first combined acoustic signal all frequencies higher than a first frequency; and a second filter (408) configured to filter the second combined acoustic signal prior to combining the first combined acoustic signal and the second combined acoustic signal to produce the third combined acoustic signal, wherein filtering the second combined acoustic signal comprises: removing all Frequencies lower than a second frequency from the second combined acoustic signal. The electronic device (102) after claim 9 , wherein the first frequency or the second frequency is between about 1000 and 4000 Hertz (Hz).

Images