Microphone array and selecting ideal pickup mode

Abstract

The invention relates to a microphone array comprising at least two microphones each associated with a predefined microphone pickup mode, and a method of selecting a desired pickup mode for a microphone array. The desired pickup mode is selected from a list of predefined pickup modes comprising the predefined microphone pickup modes and predefined mixed pickup modes. The method comprises receiving microphone audio signals from the microphones to establish a plurality of microphone audio signals, and mixing the plurality of microphone audio signals according to a plurality of predefined mixed pickup modes to establish a plurality of mixed audio signals. Individual level characteristics of the respective individual microphone audio signals and the respective individual mixed audio signals are determined, and the desired pickup mode is selected from the list of predefined pickup modes based on the individual level characteristics.

Description

Technical Field

[0001] This invention relates to a microphone array, and to a method for selecting an ideal pickup pattern for the microphone array. Background Technology

[0002] Microphones are characterized by specific polarity / pickup patterns, which describe the microphone's sensitivity to sound arriving from different angles relative to its central axis. One of the most common pickup patterns is the cardioid pattern, which provides high sensitivity on one side of the microphone while rejecting sound from the other side. Therefore, speech can be efficiently picked up by the sensitive side of the microphone, while ambient noise and interference from other directions are rejected. Using microphone arrays, preferred beamforming techniques can be applied to direct the sensitivity of the microphone array in different directions to enhance sound acquisition in specific directions. Forming a beam that provides sensitivity to sound in the correct direction can greatly improve sound acquisition of said sound and, for example, reduce the signal-to-noise ratio. On the other hand, in certain situations, forming an unsuitable beam can lead to unsatisfactory or even unusable sound acquisition. However, beamforming techniques are very flexible; the beam can be modified according to certain parameters, for example, during sound recording. Therefore, in practice, this technique can make the sensitivity of a beamforming microphone system unpredictable for users accustomed to using conventional microphones. For example, the system can limit recording variations by continuously adjusting the sensitivity (beamforming) based on the sound, thereby reducing alterations, such as those that may occur when the voice changes position relative to the microphone array. Furthermore, beamforming technology requires significant computer processing power. Moreover, selecting the optimal beam or optimal pickup pattern for a specific situation is difficult and requires extensive knowledge of microphone pickup patterns and rich task experience. Summary of the Invention

[0003] The inventors have identified the aforementioned problems and challenges related to microphones, and subsequently derived the following inventions that provide advantageous microphone systems.

[0004] This invention relates to a microphone array and a method for selecting an ideal pickup pattern for the microphone array, as set forth in the claims.

[0005] By using a microphone array capable of selecting an ideal pickup pattern, sound can be advantageously recorded in an ideal pickup pattern that is highly sensitive to the direction of the specific sound while being less sensitive to sound emanating from other directions (ambient noise). By rejecting such ambient noise while still providing high sensitivity to sound, an ideal signal-to-noise ratio can be advantageously achieved. Therefore, users of the invention will understand that the invention advantageously makes it possible to record sound (e.g., speech) in noisy environments using an ideal pickup pattern that is highly sensitive to sound (e.g., a user's speech) while efficiently deflecting noise from other directions to minimize contamination of the recorded audio signal with said noise. Furthermore, users will readily appreciate that, in embodiments of the invention, the microphone array of the invention can automatically adapt to different recording conditions by automatically switching to an ideal pickup pattern from a predefined pickup pattern list.

[0006] Advantageously, the method for selecting the ideal pickup pattern requires very little processing power, and therefore can be implemented in many different systems and applications related to optimizing recording. Thus, systems that cannot utilize conventional complex beamforming methods due to limited processing power can advantageously utilize the microphone array of the present invention.

[0007] Furthermore, this invention provides a stable, ideal pickup pattern during recording, unlike conventional beamforming methods where the beam is constantly redirected to a specific, identified sound source. This has the advantage that, for example, when a singer or speaker uses the method of this invention to record speech, the singer / speaker, like with conventional microphones, can utilize the dynamic changes in the recorded audio level as the singer / speaker moves the associated microphone array to add dynamics to the recorded sound. For example, when, for instance, the singer / speaker intentionally moves to one side of the microphone array and / or further away from it, the recorded audio signal becomes lower. Therefore, it should be understood that the method of this invention can reduce ambient noise by utilizing an ideal pickup pattern that is sensitive to, for example, the user's voice, while providing similar dynamics and a similar user experience as conventional microphones.

[0008] Advantageously, embodiments of the invention may include an audio switcher capable of automatically switching from the active pickup mode to the selected ideal pickup mode. This advantageously has the effect of always applying the ideal pickup mode without requiring any user instruction. Furthermore, this enables users without experience recording using different pickup modes to achieve recordings of ideal quality by ensuring that the user always uses the ideal pickup mode.

[0009] Another feature of embodiments of the invention is the selection of an ideal pickup mode based on predefined pickup patterns, thereby enabling the simultaneous creation of individual audio signals associated with these individual predefined pickup patterns while still using minimal processing power. By simultaneously creating all these individual audio signals, the microphone array of the invention can seamlessly switch between pickup patterns, and, potentially, also seamlessly switch between audio signals associated with different predefined pickup patterns when an audio switcher is implemented, without the user or listener experiencing any perceptible delay. This feature can be advantageously utilized, for example, when the microphone array of the invention is used by more than one user and where these users are in different positions relative to the microphone array. An example of this could be an interview scenario, where the invention can select the ideal pickup pattern to use when the interviewer is speaking, and automatically select a different ideal pickup pattern when the interviewee is speaking. Since all these pickup patterns are predefined, the microphone array can advantageously switch between audio signals associated with two ideal pickup patterns related to two corresponding users, without the user and / or others listening to, for example, the output of the microphone array experiencing any perceptible delay. In similar situations, a bipolar pickup pattern can be chosen as the ideal pickup mode because it is sensitive to sounds from two opposite directions. Therefore, it allows recording of both the interviewer and the interviewee sitting in front of the interviewer, while advantageously minimizing ambient noise from both sides.

[0010] In a preferred embodiment of the invention, voice activity detection can be applied to identify voices in the recorded audio signal. The level of the identified voice can then be measured, and a pickup pattern providing the highest level of the identified voice can be selected as the ideal pickup pattern. Therefore, in this configuration, the microphone array can advantageously be particularly sensitive to voices while deflecting other sounds. By utilizing predicted data of voices or other sounds (including, for example, different instrument sounds), voice activity detection can be adjusted to identify specific sounds, such as voices and / or different instrument sounds, thereby allowing the selection of the ideal pickup pattern of the invention to be sensitive to these different predicted sounds, which is advantageous. Attached Figure Description

[0011] Various embodiments of the invention will be described below with reference to the accompanying drawings, in which...

[0012] Figure 1 A block diagram illustrating a microphone array according to an embodiment of the present invention;

[0013] Figure 2 This describes the predefined microphone pickup mode and the predefined hybrid pickup mode according to embodiments of the present invention;

[0014] Figure 3A block diagram illustrating a microphone array with a switching block according to the present invention;

[0015] Figure 4 This describes an audio analyzer block with voice activity detection according to an embodiment of the present invention;

[0016] Figure 5 This describes a microphone array with an audio processing unit and a user interface according to an embodiment of the present invention; and

[0017] Figure 6 The method steps according to an embodiment of the present invention are described. Detailed Implementation

[0018] In the following text, various concepts of the invention are given without reference to specific embodiments.

[0019] The microphone array according to the invention may include at least two microphones, an audio mixer block, an audio analyzer block, and predefined pickup patterns. The mixer block and the audio analyzer block are configured to determine an ideal pickup pattern among the predefined pickup patterns based on the recorded microphone audio signals. The mixer block, the audio analyzer block, and the predefined pickup patterns may be included in an audio processing unit. The audio processing unit may have a memory including the predefined pickup patterns, and a processor configured to execute instructions (actions) stored in the memory. The audio processing unit is preferably a digital signal processor. However, according to certain embodiments of the invention, the invention may also be implemented using analog audio processing. The audio processing unit may be arranged together with the microphones in a housing. However, in some embodiments, the microphones and the audio processing unit may be arranged in separate housings so that the microphones can be located away from the audio processing unit, which may be advantageous in some cases. In the case of recording in a recording studio, the audio processing unit may be located in a control room so that the user can control the microphone array of the invention from the control room, while the microphones are located, for example, in a recording studio to record the voice of an instrument or singer. Similarly, during a live performance, the microphone can be located on the stage to record sound, while the audio processing unit is located, for example, in a stage frame unit or the like, or in a mixing tower. Similarly, when implemented in an automotive audio system, for example for recording sound during a telephone call, the audio processing unit can be located in a different position from the microphone of this invention.

[0020] As will be discussed in more detail in the following paragraphs, a microphone array may include several different types of microphones, as well as microphones with different pickup patterns. Microphones may be positioned to record sound from various directions, and microphones may be physically arranged to point in different directions relative to each other. Microphones record audio and provide a microphone audio signal comprising a representation of the sound recorded by the microphone. The microphone audio signal is typically initially an analog signal provided by the microphones in the microphone array. These analog microphone audio signals can be mixed, processed, and / or analyzed in the analog domain; however, in a preferred embodiment of the invention, the microphone audio signal is converted into a digital audio signal by an analog-to-digital converter to achieve audio mixing, audio signal processing, and audio signal analysis in the digital domain. Therefore, it should be understood that when referring to a microphone audio signal, the signal may be an analog audio signal or a digital audio signal, depending on the specific embodiment of the invention. Thus, even if the specification does not describe or mention an analog-to-digital converter regarding the microphone audio signal, the audio signal may be a digital signal.

[0021] It should be understood that in some embodiments of the invention, a portion of the audio processing / audio mixing / audio analysis can be performed in the analog domain, while different portions of the audio processing / audio mixing / audio analysis can be performed in the digital domain. For example, in a particular embodiment of the invention, the mixer block can be an analog mixer block that provides analog signals to an audio switching block and an audio analyzer block, while the audio signal to the audio analyzer block is converted into a digital signal before analysis, and the audio switching block receives the analog audio signal. When the audio analyzer block has selected an ideal pickup mode based on the digitized audio signal, the audio analyzer block can instruct the switcher to select the analog audio signal associated with the selected ideal pickup mode as the audio output signal. Thus, any aliasing or noise that may occur during analog-to-digital conversion and / or digital-to-analog conversion is advantageously avoided because the audio output signal that can be provided as the audio output signal remains in the analog domain. Therefore, this analog bypass system for analog audio can advantageously provide improved sound quality.

[0022] Every microphone has a specific pickup pattern that describes its directional sensitivity. The pickup pattern is an inherent characteristic of a microphone and depends on various parameters related to its configuration and construction, including the specific microphone type (technology), the materials used, the size of the microphone components, the housing, etc. It should be noted that the pickup pattern can also be referred to as a polarity pattern. It should be understood that when referring to a predefined microphone pickup pattern, this refers to the pickup pattern of the microphones in a microphone array. Conversely, when referring to a predefined hybrid pickup pattern, this refers to a pickup pattern formed by combining / mixing two or more predefined microphone pickup patterns, and therefore numerous predefined hybrid pickup patterns can be achieved by mixing predefined microphone pickup patterns and predefined hybrid pickup patterns. Advantageously, the microphone array of the present invention can utilize these predefined hybrid pickup patterns to improve sound recording by selecting an ideal pickup pattern from among predetermined pickup patterns best suited for a given recording situation, for example, selecting a pickup pattern that is sensitive to the direction of the sound to be recorded (e.g., speech) but significantly less sensitive to sound from other directions.

[0023] An ideal pickup pattern can be a predefined pickup pattern associated with the level characteristics of a microphone audio signal or mixed audio signal that provides the highest signal-to-noise ratio (e.g., signal-to-ambient-noise ratio) under a specific recording condition. Level characteristics can include the complete audio signal or a subset of audio signals. When the level characteristics are the complete audio signal, the level is simply determined for each individual audio signal, and the signal providing the highest level determines the ideal pickup pattern as the predefined pickup pattern associated with that audio signal. The level can be measured, for example, as the root mean square of the audio signal, or as loudness, signal peak, or other level metrics. In other embodiments of the invention, an ideal pickup pattern can be selected relative to one or more level characteristics. For example, a level characteristic can thus refer to, for example, a specific predetermined frequency band of an audio signal, or it can refer to a specific sound identified in one or more microphone audio signals and / or one or more mixed audio signals, such as speech or instrumental or animal sounds. Therefore, in a first step, level characteristics are identified, and then in a subsequent step, the level of the identified level characteristics is determined. The levels of each identified level characteristic are then compared, and an ideal pickup pattern is selected as the predefined pickup pattern associated with the level characteristic having the highest level. Therefore, it is advantageous to identify a specific recorded sound (e.g., speech) as a level characteristic, then determine the level of the identified speech, and use said level to identify an ideal pickup pattern. Advantageously, the ideal pickup pattern can therefore be a pickup pattern among predefined pickup patterns that provides the highest signal-to-ambient-noise ratio, wherein the signal is considered as the identified speech and the ambient noise is considered as anything other than the identified speech.

[0024] Predefined pickup patterns are predefined and can be predefined before the microphone array of the present invention is activated and used for audio recording and before the ideal pickup pattern is determined. Predefined pickup patterns and / or representations of predefined pickup patterns (e.g., instructions on how to mix microphone pickup patterns to achieve a mixed pickup pattern) can therefore be obtained in some embodiments of the invention and stored on the microphone array, for example, in a memory implemented in the array, and thus provided as pickup patterns that can be advantageously selected as the ideal pickup pattern, depending on the sound recorded under a given recording condition. Predefined pickup patterns include predefined mixed pickup patterns. Predefined mixed pickup patterns can be obtained by combining / mixing predefined microphone pickup patterns, for example, by linear and / or nonlinear combinations of predefined microphone pickup patterns. Even additional predefined mixed pickup patterns can be established by mixing predefined mixed pickup patterns. This can advantageously provide a large number of predefined pickup patterns, which effectively provides the flexibility to design microphone arrays that can be fine-tuned for recording in a variety of different environments.

[0025] Optionally, an ideal pickup pattern can be selected by considering pre-measured audio data (including one or more speech sounds, instrument sounds, animal sounds, or other sounds of interest to the user). Thus, an ideal pickup pattern can be selected as a predetermined pickup pattern associated with the audio signal that best matches the sound characteristics or the speech characteristics of the predicted audio data. Speech characteristics can, for example, refer to the frequency characteristics and / or occurrence of speech features.

[0026] Communication connections between microphones, audio processing units, etc., can usually be implemented via wired connections, but in principle they can also be wireless, such as Bluetooth or even Wi-Fi or wireless communication and / or other wireless connections.

[0027] Optionally, the present invention can provide an audio output associated with an ideal pickup pattern including a representation of the recorded sound. This signal can optionally be transmitted to, for example, a speaker for reproduction and / or to a recording device for storing the recorded sound. Such reproduction may typically be real-time reproduction, but in principle it can also be achieved by processing a slight delay, or by recording for later reproduction as mentioned, such as podcast recording.

[0028] The main feature of this invention is the determination of an ideal sound pickup pattern. As mentioned, this can be facilitated by an audio processing unit. The audio processing unit may be, for example, a digital signal processor, or several digital signal processors, or even a central processing unit that controls multiple microphone arrays according to the invention.

[0029] The microphone array of the present invention can be advantageously implemented as a handheld, desktop, or cantilever microphone array, or it can be implemented in cellular phones, tablet computers, personal computers, conference microphone systems, live music PA systems, studio recording systems, etc., to record sound from a sound source from different positions relative to the microphone of the mentioned device. Regardless of the implementation, the present invention can advantageously select an ideal pickup pattern for sound recording to ensure an ideal signal-to-noise ratio.

[0030] It should be understood that, depending on the specific implementation, the various audio signals described with respect to this invention can be digital or analog signals. For example, it may be preferred to perform audio processing using digital audio signals, including mixing audio signals and determining an ideal pickup pattern. Therefore, as mentioned, the microphone signal can be converted from an analog signal to a digital signal using an analog-to-digital converter before mixing and / or before further audio processing / analysis. Thus, analog-to-digital conversion can be applied at different steps in the audio processing pipeline. For example, before mixing the audio signals, before analyzing the mixed audio signals to determine the ideal pickup pattern, etc.

[0031] Various embodiments of the present invention are described below with reference to the accompanying drawings.

[0032] Figure 1 A block diagram illustrating a microphone array 100 according to an embodiment of the present invention is provided. The microphone array includes two microphones 1a and 1b, a mixer block 3, an audio analyzer block 5, and a predefined pickup pattern 15.

[0033] In this embodiment, the two microphones 1a and 1b are condenser microphones, each equipped with a cardioid pickup pattern (polarity mode). Microphones 1a and 1b are arranged at a 90-degree angle to each other, thereby making them sensitive to sound from different directions. This advantageously broadens the achievable pickup patterns of the microphone array 100 compared to a microphone array where the microphones are arranged pointing in the same direction. Microphones 1a and 1b are connected to a mixer block 3, each microphone providing the mixer block with a microphone audio signal 2a and 2b representing the sound 101 recorded by each microphone. Upon receiving the microphone audio signals, the mixer block 3 mixes the two audio signals according to a predefined mixing pickup pattern among a plurality of predefined pickup patterns 15, which also includes predefined microphone pickup patterns. In this exemplary embodiment, the predefined microphone pickup pattern refers to the aforementioned cardioid pickup pattern of the two microphones. Figure 2 The text describes the cardioid predefined microphone pickup modes 14a and 14b.

[0034] For example, a predefined hybrid pickup pattern 15 of an embodiment of the invention is established by a linear combination of two cardioid microphone pickup patterns. This means that a hybrid audio signal associated with a given linear combination of pickup patterns can be obtained by a similar linear combination of microphone audio signals. In this embodiment, the first hybrid predefined pickup pattern in the predefined pickup pattern 15 is achieved by summing the two predefined microphone pickup patterns. Figure 2 This section describes an associated example of predefined microphone pickup patterns 14a and 14b and a predefined hybrid pickup pattern for two microphones 1a and 1b. A predefined hybrid pickup pattern 14c is achieved by summing the two predefined microphone pickup patterns 14a and 14b. Note that both microphone pickup patterns have a similar cardioid shape, but each microphone pickup pattern also has its own specific directivity, referring to the direction in which the microphone is most sensitive. Summing the two predefined microphone pickup patterns 14a and 14b advantageously produces a predefined hybrid pickup pattern 14c with a different directivity compared to the pickup patterns in the two individual predefined microphone pickup patterns. A second predefined hybrid pickup pattern 14d in predefined pickup pattern 15 is achieved by subtracting the two predefined microphone pickup patterns 14a and 14b. This particular predefined hybrid pickup pattern has a bidirectional pickup pattern 14d. When using the bidirectional pickup pattern, the microphone array becomes sensitive in two opposite directions, which is advantageous when it is desirable to record sound from two opposite directions while limiting the recording of sound from other directions.

[0035] To establish an audio signal associated with the two predefined mixed pickup patterns mentioned above, mixer block 3 sums the audio signals 2a and 2b from the two microphones to establish an audio signal associated with the two predefined mixed pickup patterns. Figure 2 The first mixed audio signal 4a is associated with the mixed cardioid pickup pattern 14c described herein. Simultaneously, a second mixed audio signal is created using mixer block 3, which subtracts the two microphone audio signals to create a signal associated with... Figure 2 The second mixed audio signal 4b is associated with the mixed bidirectional pickup mode 14d described herein. Therefore, four audio signals, each based on a different pickup mode and each comprising a representation of the recorded sound, are established. An audio analyzer block 5, coupled to the mixer block 3 via a communication method, receives these audio signals.

[0036] Audio analyzer block 5 receives the aforementioned audio signals from the mixer and determines the individual level characteristics of the respective individual microphone audio signals and the respective individual mixed audio signals. Next, analyzer block 5 compares the determined individual level characteristics to select an ideal pickup pattern. In this exemplary embodiment, the level characteristic is the root mean square (RMS) of the recorded audio signal. Therefore, audio analyzer block 5 calculates the RMS of each of the microphone audio signals and the mixed audio signals, and then compares the RMS of the signals to detect the audio signal with the highest RMS. The audio analyzer block then selects the pickup pattern from the predefined pickup patterns associated with the audio signal with the highest RMS as the ideal pickup pattern. In this example, since sound 101 is emitted from a position between the two microphones 1a and 1b at an angle of approximately + / -45 degrees relative to the center of the front of each microphone, it is... Figure 2 The hybrid cardioid pickup pattern 14c described herein has the highest RMS value for the audio signal obtained by summing the audio signals from the two microphones. Therefore, the audio analyzer block 5 selects this pickup pattern as the ideal pickup pattern among the predetermined pickup patterns 15. Note that the directivity of the selected ideal pickup pattern is directed towards the direction of the sound 101.

[0037] Optionally, the audio analyzer block 5 provides an ideal pickup pattern signal 6, which includes information about the selected ideal pickup pattern.

[0038] Optionally, additional predefined hybrid pickup patterns can be created, for example, by adding and dividing predefined microphone pickup patterns (e.g., by a non-linear combination of predefined microphone pickup patterns).

[0039] Optionally, the mixed audio signal may also include filtering using different audio filters having filter coefficients determined based on a predetermined pickup pattern. The filters may include, for example, finite impulse response (FIR) filters and infinite impulse response (IIR) filters.

[0040] Optionally, some embodiments of the present invention may be implemented using different types of level characteristics (including, for example, the loudness of an audio signal, signal peak, voice activity, etc.).

[0041] Figure 2Examples of predefined microphone pickup patterns and predefined hybrid pickup patterns according to embodiments of the present invention are described. It should be understood that although two or more predefined pickup patterns may have the same shape (also referred to as a pattern), such as a heart shape, these pickup patterns can also be considered different if they have different directivity, where directivity refers to the direction in which the pickup pattern points. In the context of the present invention, the directivity of a pickup pattern refers to the direction in which the pickup pattern is most sensitive. If the pickup pattern is omnidirectional, then the pickup pattern is not directive. If the pickup pattern is bidirectional, then this pickup pattern therefore has two directivities pointing in opposite directions, such as by... Figure 2 The predefined hybrid pickup pattern 14d is described.

[0042] In an optional embodiment of the invention, each predefined pickup pattern is associated with a shape and a directivity, and at least two predefined pickup patterns in the predefined pickup pattern list include the same shape (e.g., heart-shaped) but different directivity; and / or wherein at least two predefined pickup patterns in the predefined pickup pattern list include the same directivity but different shapes. This is advantageous because multiple different pickup patterns, each adapted to record sound from a variety of different directions (and deflect the sound), can be provided, thereby providing a flexible microphone array that can record sound from many different directions relative to the microphone array with a high signal-to-ambient-noise ratio while keeping the required processing to a minimum.

[0043] Although Figure 2 The microphone pickup modes and hybrid pickup modes have been discussed. Figure 1 The embodiments described herein are illustrated, but it should be understood that the microphones in the microphone array according to the invention may be characterized by being mixed with other different predefined pickup patterns to establish even other different predefined mixed pickup patterns. Furthermore, the microphone array of the invention may include microphones, each with a different microphone pickup pattern. For example, one microphone may have a cardioid pickup pattern, while another microphone in the microphone array may, for example, have a supercardioid pickup pattern. Moreover, depending on the embodiment of the invention, the microphone array may include various numbers of microphones. For example, three, four, five, or five to ten microphones, or even more microphones. Therefore, the microphone array according to the invention can utilize the same microphone with the same pickup pattern, or a large number of microphones with different microphone pickup patterns. Non-limiting examples of microphone pickup patterns that can be implemented in the microphone array according to the invention include omnidirectional pickup patterns, bidirectional pickup patterns, cardioid pickup patterns, supercardioid pickup patterns, super-supercardioid pickup patterns, and extremely supercardioid pickup patterns, to name just a few.

[0044] Different types of microphones can be implemented in the microphone array, which is advantageous, for example, to adjust the frequency response and / or sensitivity according to different specific use cases of the microphone array. Non-limiting examples of microphone types that can be implemented according to the invention include condenser microphones, dynamic microphones, ribbon microphones, fiber optic microphones, carbon microphones, liquid microphones, electret microphones, laser microphones, and crystal microphones.

[0045] The microphones of the microphone array according to the invention can be positioned relative to each other in various ways. In some embodiments of the invention, all microphones can be positioned such that their front faces point in the same direction, while in other embodiments, the microphones can be configured at an angle to each other. In yet another embodiment of the invention, the microphones are not located in a linear array, but rather, for example, alternatively, the microphones can be located in a circular array, which can advantageously provide a flexible microphone array capable of adjusting the pickup pattern of the array even more precisely for sound from any direction.

[0046] Figure 3 A block diagram illustrating a microphone array 100 according to an embodiment of the present invention. Figure 3 The embodiments can be interpreted as Figure 1 The embodiments described herein are extended to include additional optional features and microphones positioned differently relative to each other. Therefore, it should be understood that in some embodiments of the invention, [the following is related to...]. Figure 1 The optional features associated with the described embodiments may optionally be implemented as follows: Figure 3 Optional features of the embodiments described herein, and vice versa. As in another embodiment of the invention, Figure 3 The embodiments are able to select the ideal pickup pattern of the microphone array. However, in this particular embodiment of the invention, audio processing is performed by the audio processing unit 18 included in the microphone array. This is advantageous because it has the following effect: the microphone array can perform the selection of the ideal pickup pattern of the microphone array without any communication with one or more external audio processing units, thereby making the array installation less complex and less prone to errors.

[0047] Except for about Figure 1 In addition to the two microphones 1a and 1b, the mixer block 3, the audio analyzer block 5, and the predefined pickup pattern 15, the described embodiments also include Figure 3 The embodiment described also includes an analog-to-digital converter (ADC) 16 and an audio switching block 7, and the audio analyzer block 5 includes a level characteristic determiner block 11 and a level characteristic comparator block 13. The ADC 16, mixer block 3, audio analyzer block 5 and switching block 7 are all included in the audio processing unit 18.

[0048] Microphones 1a and 1b are oriented in opposite directions to enable recording of sound from these opposite directions. The microphones are connected to an ADC that receives microphone audio signals from the microphones and converts them into digital microphone audio signals, including digital representations of the microphone audio signals. The digital microphone signals are then provided to a mixer unit, which, as in the previously described embodiment, mixes the two digital microphone audio signals according to a predefined mixing pickup pattern among a plurality of predefined pickup patterns 15, which also include predefined microphone pickup patterns. As previously described (see, for example...) Figure 1 (As described above), the audio mixer provides microphone audio signals and mixed audio signals to an audio switching block and an audio analyzer block 5, which is communicatively coupled to the mixer block 3. The level characteristic determiner block 11 of the audio analyzer block 5 then identifies the characteristics of the received audio signals, in this case, identifying the filtering of the received audio signals using a bandpass filter to create new audio signals including a specific frequency band associated with human speech, such as from 400Hz to 2000Hz, from 200Hz to 4000Hz, or from 400Hz to 6000Hz. The levels of these filtered audio signals are then determined. Here, the level is measured as loudness. The loudness measurement is then received by a level characteristic comparator block 13, which compares the levels of the filtered audio signals and determines the ideal pickup pattern of the microphone array 100 as a predefined pickup pattern in a predefined pickup pattern 15, which is associated with the filtered audio signal having the highest level. The audio analyzer 5 then outputs an ideal pickup mode signal 6, which includes information about the ideal pickup mode. The ideal pickup mode signal 6 is received by the audio switching block 7. Based on the received ideal pickup mode signal 6, the audio switching block 7 then switches from a given activated audio signal to the audio signal associated with the selected ideal pickup mode from the received microphone audio signal and the mixed audio signal, as the output audio 8. Therefore, the ideal pickup mode becomes the activated pickup mode. Thus, in this embodiment, the process of selecting the ideal pickup mode and providing the audio output associated with the ideal pickup mode is completely automatic.

[0049] In this embodiment, if a new ideal pickup pattern different from the activated pickup pattern is identified, the activated pickup pattern can be updated according to the ideal pickup pattern during sound recording. This advantageously ensures that the ideal pickup pattern is always selected as the activated pickup pattern.

[0050] Figure 4 A block diagram illustrating an audio analyzer block 5 according to an embodiment of the present invention includes a level characteristic determiner block 11 and a level characteristic comparator block 13, wherein the level characteristic determiner block includes voice activity detector blocks 9a-n and level determiner blocks 20a-n. Figure 4 The audio analyzer block 5 can be implemented as a feature in various other embodiments of the invention to determine the ideal pickup pattern of the microphone array using speech activity detection. Therefore, Figure 4 The specific features of the audio analyzer block 5 described herein can be understood as optional features of the present invention, and these optional features may preferably be implemented, for example, when the microphone array is used to record speech / voice.

[0051] In this exemplary embodiment, the analyzer block receives multiple microphone audio signals 2a-n and multiple mixed audio signals 4a-n from the mixer block (not shown). All these audio signals are received by voice activity detector blocks 9a-n, which analyze the audio signals using voice activity detection to determine the voice activity audio signal for each of the received audio signals 2a-n and 4a-n. The voice activity detector blocks essentially extract sounds belonging to speech (e.g., human speech or even human singing) from the received audio signals and then output the voice activity audio signal including the extracted speech. Advantageously, this reduces the noise content in the voice activity audio signal. Voice activity can be detected in various ways. For example, by filtering the received audio signal by extracting one or more frequency bands belonging to the frequency band including human speech. Voice activity detection can also be performed by comparing the received audio signal with predicted audio signals of different speech. The comparison may include correlation of the signals or may include machine learning-based speech recognition, such as using supervised or possibly unsupervised, semi-supervised, or reinforcement learning. However, methods using minimal processing power are preferred. Alternatively, a supervised learning algorithm requiring significant processing power can be trained on an external system using the predicted data, and then the trained classifier can be implemented on the microphone array of this invention to provide voice activity detection. These methods can include Bayesian classification methods, including Naive Bayes, Support Vector Machines, different types of neural networks, decision trees, and even unsupervised methods including different types of clustering or autoencoder networks.

[0052] In this exemplary embodiment, the voice activity detector blocks 9a-n include a noise reduction stage for denoising the received audio signal, a feature extraction stage for extracting quantities as features from a portion of the received denoised audio signal, and a classification stage for classifying the portion as voice or non-voice when one or more extracted features exceed a threshold. The noise reduction stage includes spectral subtraction, wherein a noise spectrum is estimated during a speech pause, and then the noise spectrum is subtracted from the noisy speech spectrum to estimate noise-free speech.

[0053] The individual voice activity detector block outputs an individual voice activity audio signal including the detected voices. Individual voice audio signals are received by individual level determiner blocks 20a-n, which determine the voice activity signal level of each of the received signals. As previously described, the voice activity signal level of an individual signal can be determined in various ways; however, in this example, the level is determined as RMS. The voice activity signal level (RMS) of the individual signal is received by level characteristic comparator 13, which compares all voice activity signal levels to determine the voice activity audio signal 12 with the highest level. The level characteristic comparator then selects an ideal pickup pattern as the pickup pattern associated with the voice activity signal having the highest level (RMS), and finally outputs an ideal pickup pattern signal 6. The ideal pickup pattern signal 6 includes information about the ideal pickup pattern selected by the audio analyzer block 5. In an embodiment, the information about the ideal pickup pattern in signal 6 may simply be a channel number or other reference to an audio channel that includes the audio signal that produces the highest level.

[0054] The decision rules used to determine speech activity may optionally be based on, for example, a frame-by-frame basis using an instantaneous metric of the divergence distance between speech and noise. Different metrics may be used according to the invention, including spectral slope, correlation coefficient, log-likelihood ratio, cepstrum, weighted cepstrum, and modified distance metrics.

[0055] Optionally, voice activity detection can provide feedback, whereby voice activity detection decisions are used to improve noise estimation in the denoising stage or to adaptively change thresholds in the classification stage. Advantageously, these feedback operations improve voice activity detection performance, for example, in unstable noise (i.e., when the noise varies considerably).

[0056] In an optional preferred phase, a feedback mechanism incorporating user feedback is used as feedback to the reinforcement learning algorithm to optimize voice activity detection in a particularly preferred embodiment, wherein voice activity detection includes reinforcement learning.

[0057] Optionally, the voice activity detector block can be calibrated based on pre-recorded audio data including voice recordings from different individuals. Preferably, these individuals have different genders and represent different voices. For example, voices with different pitches, levels, and changes in pitch and level over time (phrasing).

[0058] Optionally, the calibration of the voice activity detector blocks 9a-n may include pre-recorded audio data, which includes recordings by users intending to use the microphone array of the present invention. Advantageously, this calibration ensures that the voice activity detector can learn characteristics and / or representations of speech similar to those of the user, thereby improving the accuracy and / or sensitivity of voice activity detection when the user is using the microphone.

[0059] Optionally, voice activity detection can be performed using a bandpass filter whose bandwidth and center frequency correspond to the desired characteristics of human voice. For example, a band within the 300Hz to 3500Hz range, or alternatively, a band within the 200Hz to 6000Hz range. The band can be selected to an even narrower range to eliminate contaminating noise.

[0060] Optionally, the method includes the step of calibrating the microphone array based on the average voice level of the predicted voice recording data.

[0061] Optionally, voice activity detection is calibrated based on the average voice characteristic frequency response of the predicted voice recording data.

[0062] In a preferred alternative embodiment of the invention, voice activity detection can be configured to identify other sounds, such as sounds from musical instruments, sounds from the audience, music, etc.

[0063] Those skilled in the art of speech recognition also know of several other speech activity detection methods, which can also be implemented using this invention.

[0064] The feature of voice activity detection is generally advantageous because it ensures that the ideal pickup pattern is selected based on voice activity identified by voice activity detection, rather than on, for example, irrelevant ambient noise (which would produce less desirable results). Voice activity can refer to any relevant voice recorded by a microphone and identified by voice activity detection. For example, relevant voice activity can be human-generated speech. Advantageously, the microphone array can thus determine the ideal pickup pattern based on voice activity, thereby improving the quality of the sound recorded by the microphone array compared to situations where ambient noise would have a greater impact on the microphone audio signal, thus reducing the quality of detection of the ideal pickup pattern. For example, by utilizing voice activity detection, the signal-to-ambient-noise ratio can be improved, which is advantageous.

[0065] Figure 5 A block diagram illustrating a microphone array 100 according to an embodiment of the present invention is provided. The invention includes features that can be implemented in conjunction with features of other embodiments of the invention. Therefore, this embodiment can be considered as, for example... Figure 1 , Figure 2 , Figure 3 and Figure 4 Optional extensions to the previously described embodiments (including optional features described with respect to these embodiments).

[0066] Figure 5An embodiment includes five microphones 1a-e, an audio processing unit 18, and a user interface 17. The user interface is communicatively coupled to the audio processing unit 18 via a data communication link 19, which facilitates bidirectional communication between the user interface 17 and the audio processing unit 18. The five microphones are oriented in different directions so that the microphone array 100 is sensitive to sound from any direction pointed to by the microphones and from directions in between, according to a predefined mixed pickup pattern. The five microphones are connected to the audio processing unit 18, each microphone providing a microphone audio signal 2a-e to the audio processing unit. As previously described with respect to other embodiments of the invention, the audio processing unit mixes the microphone audio signals according to a predefined mixed pickup pattern and determines an ideal pickup pattern based on a comparison of the different level characteristics of the microphone audio signals and the mixed audio signals. The audio processing unit provides information to the user interface having the selected ideal pickup pattern via the data communication link.

[0067] Simultaneously, the audio processing unit provides an audio output 8 associated with the activated pickup mode. The activated pickup mode can be automatically selected by the audio processing unit 18 as the audio signal corresponding to the selected ideal pickup mode. However, the activated pickup mode can also be selected by the user via a user interface.

[0068] Based on this information, the user interface can indicate to the user whether the ideal pickup pattern is active. This can be done via a light source on the user interface; when the ideal pickup pattern matches the active pickup pattern, the light source illuminates in a specific color. If this is not the case, the light source may change color, begin flashing, or otherwise indicate that the ideal pickup pattern is not active. Other ways to alert the user when the ideal pickup pattern is not active include through sound, through the user interface screen, through haptic feedback to the user touching the microphone, or through vibration of the microphone array.

[0069] It should be understood that the activated pickup mode based on the ideal pickup mode can refer to the pickup mode selected by the user based on their understanding of the selected ideal pickup mode. Therefore, in some embodiments of the present invention, the activated pickup mode may correspond to the selected ideal pickup mode. Alternatively, the activated pickup mode may correspond to the pickup mode selected by the user in the user interface based on the ideal pickup mode.

[0070] As mentioned, the user interface can be configured to allow the user to select a predefined pickup pattern as the active pickup pattern. Advantageously, the ideal pickup pattern can be communicated to the user via the user interface, so the user can select the ideal pickup pattern based on this information or select a different predefined pickup pattern (if the user has such a need).

[0071] Optionally, the user interface is configured to indicate the ideal pickup pattern and a subset of different recommended pickup patterns from the predefined pickup patterns, based on a list of predefined pickup patterns and additional recommended pickup patterns. The user can then try different pickup patterns based on the recommended pickup patterns, which are predefined pickup patterns associated with level characteristics similar to those associated with the selected ideal pickup pattern. The recommended pickup patterns may be, for example, second, third, and fourth most ideal pickup patterns determined based on level characteristics.

[0072] Optionally, users can cancel the selected ideal pickup mode via the user interface. This advantageously allows users to control the pickup mode according to their personal preferences.

[0073] Optionally, an ideal pickup pattern can be selected based on different level characteristics, which can be selected via the user interface. Level characteristics may include, for example, the voice activity level, the RMS of the audio signal, and the loudness of the audio signal.

[0074] It should be understood that the user interface is communicatively coupled to the audio processing unit 18, and therefore to the components and / or blocks included by the audio processing unit. Thus, the user interface is configured to control switching blocks based on the pickup mode selected by the user via the user interface.

[0075] The user interface may include the layout of buttons and indicator lights. However, the user interface may also be a screen, such as a touchscreen, in which the functionality of the user interface is operated by the user via the touchscreen.

[0076] Figure 6 The method steps according to an embodiment of the present invention are described. The described method steps are steps of a method for selecting an ideal pickup pattern for a microphone array including at least two microphones, each of which is associated with a predefined microphone pickup pattern, and wherein the ideal pickup pattern is selected from a list of predefined pickup patterns including predefined microphone pickup patterns and predefined hybrid pickup patterns, and wherein the method includes at least the following four steps.

[0077] In the initial method step S1, microphone audio signals are received from each of at least two microphones to establish a plurality of microphone audio signals.

[0078] In the next step S2, multiple microphone audio signals are mixed according to multiple predefined mixed pickup patterns to create multiple mixed audio signals.

[0079] In another step S3, the individual level characteristics of the respective individual microphone audio signal among the multiple microphone audio signals and the respective individual mixed audio signal among the multiple mixed audio signals are determined.

[0080] In the additional step S4, an ideal pickup pattern is selected from a list of predefined pickup patterns based on individual level characteristics.

[0081] Optionally, a mixed audio signal can be established based on a linear combination of at least two microphone audio signals from a plurality of microphone audio signals. This has the advantage of providing audio signals similar to various different pickup patterns. Another advantage is that the microphone array can provide audio signals based on pickup patterns associated with the linear combination of audio signals. Thus, a flexible microphone array is provided that can provide a variety of different audio signals based on various different pickup patterns and an ideal pickup pattern.

[0082] Choose any location Figure 6 The method described may include an optional step of alerting the user when the ideal pickup mode is not active. This is advantageous because the user can subsequently become aware that the ideal pickup mode is not activated. In an optional embodiment of the invention, the user may advantageously use this information to select a different pickup mode, such as the ideal pickup mode, via a user interface. The step of alerting the user when the ideal pickup mode is not used may employ an indicator, such as a light source (e.g., an LED), to indicate that the ideal pickup mode is not active. The light source may begin to flash in a certain pattern, and / or the light may change from one color to another to indicate that the ideal pickup mode is not active. According to embodiments of the invention, many different indicators may be used, including, for example, sound from a speaker, a combination of sound and light, and indications via a user interface, such as text and / or images displayed on the screen of the user interface.

[0083] Optionally, the described method may include a step of indicating the selected ideal pickup pattern in the user interface. This is advantageous because it provides information about the ideal pickup pattern to the user of the microphone array.

[0084] Optionally, embodiments of the method of the present invention may include the step of switching the active pickup mode to the ideal pickup mode based on input from the user interface. This is advantageous because the user can select the ideal pickup mode if it is not already active.

[0085] Optionally, embodiments of the method of the present invention may include the step of automatically switching the active pickup mode to the ideal pickup mode. This is advantageous because it ensures that the active pickup mode is the ideal pickup mode. Therefore, if the microphone is on and the ideal pickup mode is not selected, the microphone array will automatically switch to the ideal pickup mode, which is advantageous.

[0086] Optionally, the method embodiments may include an additional step of providing audio output based on a subset of audio signals from multiple microphones according to an ideal pickup pattern. This is advantageous because the method can provide audio output including a representation of the recorded sound based on the ideal pickup pattern. Therefore, compared to, for example, using a subideal pickup pattern, the method can, for example, provide audio output with a higher signal-to-ambient-noise ratio, which is advantageous.

[0087] Optionally, selecting an ideal pickup pattern from a predefined list of pickup patterns may include associating individual level characteristics with one or more predicted level characteristics. Thus, the ideal pickup pattern is selected based on the level characteristics of the sound that matches the predicted sound. This advantageously has the effect that the selected ideal pickup pattern is ideal for a particular sound, rather than ideal for different dominant sounds that the user might not want to record. In the context of this invention, correlation should be understood in the broadest sense as a comparison determining a degree of similarity or equality. Therefore, it should be understood that correlation does not necessarily simply refer to statistical correlation and different correlation coefficients. The predicted sound can be human speech, or it can be different instrument sounds or other different sounds that the user of the microphone array wants to record with the microphone array. Therefore, predicted data can be acquired to calibrate the microphone array based on the predicted data, so that the microphone array can be calibrated to be sensitive to the predicted sound and thus select the ideal pickup pattern associated with said predicted sound.

[0088] It should be understood that the optional features described with respect to specific embodiments of the invention may also be implemented optionally in any other embodiments of the invention. For example, Figure 3 , Figure 4 as well as Figure 5 and Figure 6 Features and optional features of the embodiments can be found in relation to Figure 1 The described embodiments are implemented, and vice versa. For example, Figure 1 The embodiments may optionally be implemented using additional microphones located in various directions, for example, regarding Figure 5 As described in the embodiments. Similarly, regarding Figure 4 The described audio analyzer block can be implemented in any embodiment of the invention. Furthermore, regarding... Figure 5 The user interface of the described embodiments can be implemented in any other implementation of the invention.

[0089] The invention has been illustrated above by way of example, with reference to specific examples of methods and embodiments, for purposes of explanation and not limitation. Details of specific methods and system structures have been provided to facilitate understanding of embodiments of the invention. It should be noted that detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted to avoid obscuring the description of the invention with unnecessary detail. It should be understood that the invention is not limited to the specific examples described above, and those skilled in the art can implement the invention in other embodiments without these specific details. Therefore, the invention can be designed and modified in various variations within the scope of the invention as specified in the claims.

[0090] List of reference numerals in the attached diagram:

[0091] 1 microphone

[0092] 2a-n microphone audio signal

[0093] 3 Mixer Blocks

[0094] 4a-n mixed audio signal

[0095] 5 Audio Analyzer Block

[0096] 6 Ideal pickup pattern signal

[0097] 7 Audio Switching Block

[0098] 8 Audio Output

[0099] 9a-n Voice Activity Detector Block

[0100] 10. Voice activity audio signal

[0101] 11 Level Characteristic Determiner Block

[0102] 12. Voice activity signal level

[0103] 13 Level Characteristic Comparator Block

[0104] 14a-b cardioid microphone pickup patterns

[0105] 14c Hybrid Cardioid Pickup Pattern

[0106] 14D Hybrid Two-Way Pickup Mode

[0107] 15 Predefined pickup patterns

[0108] 16 Analog-to-Digital Converters

[0109] 17 User Interface

[0110] 18 audio processing units

[0111] 19 Data Communication Link

[0112] 20 Level Determiner Block

[0113] 100-microphone array

[0114] 101 Sounds

[0115] S1-S4 Method Steps

Claims

1. A method for selecting an ideal pickup pattern for a microphone array, the microphone array comprising at least two microphones, characterized in that... The method includes the following steps: A list of predefined pickup patterns is provided, the list of predefined pickup patterns including predefined microphone pickup patterns and predefined hybrid pickup patterns, wherein at least two microphones are respectively associated with the predefined microphone pickup patterns; Receive microphone audio signals from each of the at least two microphones to establish a plurality of microphone audio signals; Multiple microphone audio signals are mixed according to the multiple predefined mixed pickup patterns to create multiple mixed audio signals; Determine the individual level characteristics of a corresponding individual microphone audio signal among the plurality of microphone audio signals and a corresponding individual mixed audio signal among the plurality of mixed audio signals; The ideal pickup pattern is selected from the predefined pickup pattern list based on the individual level characteristics; and The ideal pickup pattern is the pickup pattern with the highest individual level characteristic among the individual level characteristics of the predefined pickup patterns in the predefined pickup pattern list.

2. The method of claim 1, wherein the mixed audio signal is established based on a linear combination of at least two of the plurality of microphone audio signals.

3. The method of claim 1, wherein the predefined hybrid pickup pattern is established based on a linear combination of the predefined microphone pickup patterns.

4. The method of claim 1, wherein at least one of the mixed audio signals is established by summing at least two of the plurality of microphone audio signals.

5. The method of claim 1, wherein at least one of the mixed audio signals is constructed by subtracting at least two of the plurality of microphone audio signals.

6. The method of claim 1, wherein the step of selecting the ideal pickup pattern from the predefined pickup pattern list comprises associating individual said level characteristics with one or more predicted level characteristics.

7. The method of claim 1, wherein the predefined pickup pattern is associated with a pattern and a directivity, and wherein at least two predefined pickup patterns in the predefined pickup pattern list include the same pattern but different directivity; and wherein at least two predefined pickup patterns in the predefined pickup pattern list include the same directivity but different patterns.

8. The method of claim 1, wherein the method includes the step of providing an output audio signal from the plurality of microphone audio signals and the plurality of mixed audio signals based on an activated pickup pattern.

9. The method of claim 1, wherein the method includes the step of alerting the user when the ideal pickup pattern is different from the activated pickup pattern.

10. The method of claim 1, wherein the method includes the step of instructing the selection of the ideal pickup mode in a user interface.

11. The method of claim 1, wherein the method includes the step of selecting an activated pickup mode based on input from a user interface.

12. The method according to claim 1, wherein the method comprises the following steps: Automatically select the active pickup mode corresponding to the ideal pickup mode, and switch the output audio signal from the plurality of microphone audio signals and the plurality of mixed audio signals to a single audio signal based on the activated pickup mode.

13. The method of claim 1, wherein the level characteristic is a voice level characteristic based on individual voice activity detection of the respective individual microphone audio signals of the plurality of microphone audio signals and the respective individual mixed audio signals of the plurality of mixed audio signals.

14. The method of claim 1, wherein the voice activity detection is calibrated based on the average voice level of the predicted voice recording data.

15. A microphone array, characterized in that... include: The memory stores a predefined list of pickup patterns, including predefined microphone pickup patterns and predefined hybrid pickup patterns. At least two microphones, each associated with a predefined microphone pickup pattern in the predefined microphone pickup pattern; An audio mixer block is communicatively coupled to the at least two microphones and configured to receive microphone audio signals from each of the at least two microphones to establish a plurality of microphone audio signals, and is further configured to mix the plurality of microphone audio signals according to the plurality of predefined mixing pickup patterns to establish a plurality of mixed audio signals; An audio analyzer block, communicatively coupled to the audio mixing block and configured to receive the microphone audio signal and the plurality of mixed audio signals; The audio analyzer block is configured to determine the individual level characteristics of the respective individual microphone audio signals of the plurality of microphone audio signals and the respective individual mixed audio signals of the plurality of mixed audio signals; The audio analyzer block is further configured to select an ideal pickup pattern from the predefined pickup pattern list based on the individual level characteristics; and The ideal pickup pattern is the pickup pattern with the highest individual level characteristic among the individual level characteristics of the predefined pickup patterns in the predefined pickup pattern list.

16. The microphone array of claim 15, wherein the microphone array includes an audio switching block configured to output an output audio signal from the plurality of microphone audio signals and the plurality of mixed audio signals based on an activated pickup mode, wherein the activated pickup mode is based on the ideal pickup mode.

17. The microphone array of claim 15, wherein the mixer block is configured to generate the plurality of mixed audio signals based on a linear combination of at least two of the plurality of microphone audio signals.

18. The microphone array of claim 15, wherein the level characteristic is a voice level characteristic, and wherein the analyzer block includes a level characteristic determiner block comprising a voice activity detector block and a level detector block, wherein the voice activity detector block is configured to perform individual voice activity detection on the respective individual microphone audio signal of the plurality of microphone audio signals and the respective individual mixed audio signal of the plurality of mixed audio signals, and wherein the level detector block is configured to establish the voice level characteristic based on the individual voice activity detection.

Images