KR102167469B1 - Feedback processing apparatus - Google Patents

Abstract

The present invention relates to a feedback processing method capable of more accurately determining and removing whether feedback occurs in an acoustic signal.
The feedback processing method according to the present invention includes the steps of receiving an acoustic signal, converting the input sound signal into a signal in the frequency domain in a frame unit, performing a VAD process and a feedback verification process, and according to a previously stored VAD performance. , A step of performing a logical operation between the voice determination result of the VAD process and the feedback occurrence determination result of the feedback check process, and the voice determination result of the VAD process and the feedback generation of the feedback check process according to the logic operation determined in the execution step of the logic operation. And processing the determination result to determine a final feedback generation result, and performing mute processing according to the final feedback generation result in the determination step.

Description

Feedback processing method {FEEDBACK PROCESSING APPARATUS}

The present invention relates to a feedback processing method, and more particularly, to a feedback processing method capable of more accurately determining and removing whether feedback occurs in an acoustic signal.

A device that amplifies the sound input through a sound sensor such as a microphone and outputs sound so that the user can hear the sound more easily has been applied to hearing aids or loudspeakers and has been widely used. In the existing sound volume control device, there is a loudspeaker that uniformly amplifies the input sound according to the set sound amplification rate. However, such a loudspeaker uniformly amplifies all input sounds, thereby amplifying unnecessary noise, causing the user to feel uncomfortable in hearing and make it difficult to clearly recognize meaningful sounds.

In addition, as a conventional sound volume control device, there is a device that processes the input sound and adjusts the sound volume to a volume desired by the user to be heard. For example, in the case of hearing aids, it does not uniformly amplify all input sounds, but analyzes the frequency of the input sound, and selectively amplifies the signal of the frequency component corresponding to the frequency band that needs compensation due to loss in hearing. Works in a way. And for more accurate volume control, the frequency band where hearing loss occurs is checked through the hearing test, and the parameters of the device are set. For these once set parameters, the prior art does not provide a technique for performing an update to improve the performance of the device.

A conventional feedback cancellation system includes a summing unit for summing the voice signal S(n) and the output signal v(n) of the feedback path block F(z), and an output signal of the summing unit. An analysis unit that performs a Fast Fourier Transform and outputs an output value (Y(k)), and an estimated value from the block (W(z)) from the output value (Y(k)) from the analysis unit (v( k)) to output the output value (E(k)) (i.e., the feedback value), and the output value (X(k)) by multiplying the output value (E(k)) from the subtraction unit by the gain. ) Outputting block (G(z)), output value (X(k)) of block (G(z)) and output value (E(k)) from subtraction unit as input values to remove feedback component The block (W(z)) (feedback estimation block) that outputs the estimated value (v(k)) and the output value (X(k)) from the block (G(z)) are converted into an Inverse Fast Fourier Transform. A synthesis unit that performs transform) to output an output signal (x(n)), and a feedback path block (F()) that receives the output signal (x(n)) and outputs the output signal (v(n)) to the summing unit. z)).

The conventional feedback removal system receives output values (E(k)) and X(k) for a certain period of time, processes them in a block (W(z)), and outputs an estimated value (V(k)) for feedback in the process. The accuracy of judgment as to whether or not it occurs is an important issue.

(Patent Document 1) Korean Patent Publication No. 2000-0067641 "Adaptive Feedback Removal Device and Method for Multi-Di Station Hearing Aids" (published on November 25, 2000)

An object of the present invention is to provide a feedback processing method capable of more accurately determining and eliminating whether feedback has occurred in an acoustic signal by making the feedback determination method different according to the performance of speech detection.

The feedback processing method of the present invention includes the steps of receiving an acoustic signal, converting the input acoustic signal into a signal in the frequency domain in frame units, and checking the VAD process and feedback for each frame of the acoustic signal converted into the frequency domain. Based on the step of performing the process and the step of performing a logical operation between the voice determination result of the VAD process and the feedback occurrence determination result of the feedback check process, and the result value at the performing step of the logic operation according to the previously stored VAD performance. Determining the final feedback generation result, and performing mute processing according to the final feedback generation result of the determination step, wherein the VAD process includes an average value (Mve) of energy in the voice frequency domain in the current frame and Calculating the average value (Mne) of the energy in the non-speech frequency domain, respectively, and calculating the current first calculated value (Vc) corresponding to the number of times when the average value (Mve) is greater than the average value (Mne) for each frame, If the average value Mve is greater than the average value Mne, the first increment value is added to the previous first calculated value Vc and stored as the first calculated value Vc in the current frame, and the average value Mve is If it is less than or equal to the average value Mne, the step of subtracting the first reduction value from the previous first calculated value Vc and storing it as the first calculated value Vc in the current frame, and the current first calculated value If (Vc) is greater than the first reference value (Vf), it is determined that there is a voice, otherwise, determining that there is no voice.

In addition, if the VAD performance is greater than or equal to the reference performance in the execution step of the logical operation, the voice determination result of the VAD process and the feedback occurrence determination result of the feedback check process are ORed in the determination step, and the VAD performance is less than the reference performance in the determination step, In the determination step, it is preferable to AND the voice determination result of the VAD process and the feedback generation determination result of the feedback confirmation process.

In addition, the feedback processing method of the present invention determines whether to perform the VAD process and the feedback check process according to the steps of receiving an audio signal, converting the input sound signal into a signal in the frequency domain in frame units, and pre-stored VAD performance. Depending on the decision step and the decision at the decision step, the VAD process and the feedback verification process are performed, or any one of the VAD process and the feedback verification process is performed first, and the remaining processes are performed according to the result of the first process. In the step and in the execution step, a final feedback generation result is determined according to one or more of the voice determination result of the VAD process and the feedback generation determination result of the feedback confirmation process, and in the determination step, mute processing is performed according to the final feedback generation result. It includes performing steps.

In addition, if the VAD performance is greater than or equal to the reference performance in the decision step, one of the VAD process and the feedback check process is first performed in the execution step, and the remaining process is performed according to the result of the first performed process, and the VAD performance is performed in the decision step. If it is less than this reference performance, it is preferable to perform the VAD process and the feedback check process in the execution step.

In addition, in the VAD process, it is preferable to determine whether there is a voice or no voice using the average value of the energy in the voice frequency domain and the average energy in the non-voice frequency domain.

In addition, in the feedback check process, it is preferable to determine the occurrence of feedback and no occurrence of feedback using the energy of the frequency band in which the feedback occurs and the energy of the frequency band other than the frequency band where the feedback occurs.

According to the present invention, the feedback determination method is different according to the performance of speech detection, so that the final occurrence of feedback in the sound signal can be more accurately determined and eliminated, and the amount of computation is reduced by simplifying the processes performed during the feedback determination. There is.

1 is a block diagram of an electric device that performs a feedback processing method according to the present invention.
2 is a flowchart of a feedback processing method according to the present invention.

In the following, the present invention will be described in detail through examples and drawings. However, this is not intended to limit the present invention to a specific embodiment, it should be understood to include various modifications, equivalents, and/or alternatives of the embodiments of the present invention. In connection with the description of the drawings, similar reference numerals may be used for similar elements.

In this document, expressions such as "have", "may have", "include", or "may contain" are the presence of corresponding features (eg, elements such as numbers, functions, actions, or parts). And does not exclude the presence of additional features.

In this document, expressions such as "A or B", "at least one of A or/and B", or "one or more of A or/and B" may include all possible combinations of items listed together. . For example, “A or B”, “at least one of A and B”, or “at least one of A or B” includes (1) at least one A, (2) at least one B, Or (3) it may refer to all cases including both at least one A and at least one B.

Expressions such as “first”, “second”, “first”, or “second” used in this document can modify various elements regardless of their order and/or importance, and It is used to distinguish it from the component, but does not limit the component. For example, a first user device and a second user device may represent different user devices regardless of order or importance. For example, without departing from the scope of the rights described in this document, a first component may be referred to as a second component, and similarly, a second component may be renamed to a first component.

Some component (eg, a first component) is "(functionally or communicatively) coupled with/to)" to another component (eg, a second component) or " When referred to as "connected to", it should be understood that the certain component may be directly connected to the other component or may be connected through another component (eg, a third component). On the other hand, when a component (eg, a first component) is referred to as being “directly connected” or “directly connected” to another component (eg, a second component), the component and the It may be understood that no other component (eg, a third component) exists between the different components.

The expression "configured to" used in this document is, for example, "suitable for", "having the capacity to" depending on the situation. It can be used interchangeably with ", "designed to", "adapted to", "made to", or "capable of". The term "configured to (or set) to" may not necessarily mean only "specifically designed to" in hardware. Instead, in some situations, the expression "a device configured to" may mean that the device "can" along with other devices or parts. For example, the phrase “a processor configured (or configured) to perform A, B, and C” means a dedicated processor (eg, an embedded processor) for performing the operation, or executing one or more software programs stored in a memory device. By doing so, it may mean a generic-purpose processor (eg, a CPU or an application processor) capable of performing corresponding operations.

Terms used in this document are only used to describe a specific embodiment, and may not be intended to limit the scope of other embodiments. Singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as commonly understood by one of ordinary skill in the technical field described in this document. Among the terms used in this document, terms defined in a general dictionary may be interpreted as having the same or similar meaning as the meaning in the context of the related technology, and unless explicitly defined in this document, they may be interpreted in an ideal or excessively formal meaning. It is not interpreted. In some cases, even terms defined in this document cannot be interpreted to exclude embodiments of this document.

1 is a block diagram of an electric device that performs a feedback processing method according to the present invention. The electric device 10 may be implemented as, for example, an acoustic conversion device such as an earphone or a headset, a hearing aid device such as a hearing aid, and the like, and FIG. 1 corresponds to an example of the electric device 10.

The electrical device 10 includes a microphone 1 that acquires external sound and converts it into an sound signal, a speaker 3 that emits sound by receiving the sound signal, and an external electrical device (not shown) and wired or wireless A communication unit 5 for performing communication, a storage unit 7 for storing a feedback processing algorithm and an acoustic signal, and an acoustic signal applied from the microphone 1 or an acoustic signal or a storage unit 7 received from the communication unit 5 A data processor 9 that performs feedback processing by performing a feedback processing algorithm on the acoustic signal stored in ). However, a power supply unit (not shown), a microphone 1, a speaker 3, and a communication unit 5 for supplying power to components requiring power are provided, but it is obvious to a person skilled in the art who is familiar with the technical field to which the present invention belongs. It corresponds to a recognized technology and its description is omitted.

The storage unit 7 is implemented as a volatile and/or non-volatile memory, and stores acoustic signals, feedback processing algorithms, voice activity detection (VAD) performance data, energy thresholds, etc., and details on feedback processing algorithms. Is described in detail in FIG. 2 below.

The VAD performance data refers to the performance of the voice detection function performed by the data processor 9, for example, if the performance of determining that the sound signal contains voice is 80% (reference performance) or more, it is high performance, If it is less than 80%, it is judged as low performance. That is, the VAD performance data includes either high performance or low performance. In addition, the data processor 9 may receive the VAD performance data statistically calculated by an external electric device through the communication unit 5 and store it in the storage unit 7.

The energy threshold is a reference value for determining whether feedback has occurred in the acoustic signal. For example, if the energy ratio for each band of the acoustic signal is greater than the energy threshold, it is determined that feedback has occurred.

The data processor 9 is a processor (eg, CPU, MICROPROCESSOR, etc.) that performs feedback processing on an acoustic signal by performing a feedback processing algorithm using an acoustic signal as an input signal.

The data processor 9 processes the sound signal in a frame unit of a preset size, but has a function of performing Fast Fourier Transform (FFT) and Inverse Fast Fourier Transform (IFFT).

In addition, in order to perform the VAD process, the data processor 9 uses a signal in the frequency domain of the frame to provide an average value Mve of energy in the voice frequency domain (for example, 0 to 2 kHz) and the non-voice frequency domain (2 kHz). Each of the average energy values (Mne) of the frequency range exceeding 8 kHz) is calculated and used. In addition, the storage unit 7 stores a first reference value Vf (for example, 100) for determining the presence or absence of a voice. Since the data processor 9 has a high probability of having a voice when the average value Mve is greater than the average value Mne, the number of times when the average value Mve is greater than the average value Mne in each frame to use this point. The first calculated value Vc corresponding to is calculated and stored. The initial value of the first calculated value Vc is stored in the storage unit 7 as 0, and the data processor 9 uses the previous first calculated value Vc when the average value Mve is greater than the average value Mne. (Previous Vc) is added to the first increment (e.g., 30) and stored as the first calculated value (Vc) (current Vc) in the current frame, and the average value (Mve) is less than the average value (Mne) In the same case, a first reduction value (eg, 5) is subtracted from the previous first calculated value Vc and stored as the first calculated value Vc in the current frame. If the current first calculated value Vc is greater than the first reference value Vf, the data processor 9 determines that there is a voice, otherwise it determines that there is no voice. This process is described in more detail below.

In addition, the data processor 9 performs a feedback check process of determining whether feedback has occurred using energy in a frequency band in which the feedback occurs. The data processor 9 includes energy (Esp) in a frequency band in which feedback occurs (for example, 5,750 to 6,500 Hz) and a frequency band other than the frequency band in which feedback occurs (for example, from 0 Hz to less than 5,750 Hz). The energy (Eot) of the band and the band from over 6,500Hz to 8kHz) is calculated and used. In addition, the storage unit 7 stores a second reference value Ff (for example, 100) for determining the presence or absence of a feedback occurrence (yes/no). Since the data processor 9 has a high probability of generating feedback when the energy (Esp) is greater than the energy (Eot), in order to use this point, when the energy (Esp) is greater than the energy (Eot) The second calculated value Fc corresponding to the number of times is calculated and stored. The initial value of the second calculated value Fc is 0 and is stored in the storage unit 7, and the data processor 9 uses the previous second calculated value Fc when the energy Esp is greater than the energy Eot. The second increment value (for example, 30) is added to and stored as the second calculated value (Fc) in the current frame, and if the energy (Esp) is less than or equal to the energy (Eot), the previous second calculation A second reduction value (eg, 1) is subtracted from the value Fc and stored as a second calculated value Fc in the current frame. If the current second calculated value Fc is greater than the second reference value Ff, the data processor 9 determines that there is a feedback, otherwise it determines that there is no feedback. This process is described in more detail below.

In the present invention, the data processor 9 processes a plurality of frames of an acoustic signal as described in the process of calculating the first and second calculated values Vc and Fc to finally determine the occurrence of feedback.

2 is a flowchart of a feedback processing method according to the present invention. The data processor 9 performs feedback processing while receiving power.

In step S1, the data processor 9 receives an acoustic signal from the microphone 1 or an acoustic signal from the communication unit 5 or an acoustic signal read from the storage unit 7. The data processor 9 feedback-processes the input sound signal in units of frames having a preset size, and performs processing on each frame of the input sound signal after step S1.

In step S3, the data processor 9 converts a signal in the time domain into a signal in the frequency domain by performing a Fast Fourier Transform (FFT) on the input frame (eg, a first frame). do. The data processor 9 performs step S3 and proceeds to step S5.

In step S5, the data processor 9 uses the signal in the frequency domain of the frame to generate an average value Mve of energy in the voice frequency domain (for example, 0 to 2 kHz) and the non-voice frequency domain (above 2 kHz and within 8 kHz). Calculate each of the average energy values (Mne) in the frequency range of. The data processor 9 performs step S5 and proceeds to step S7.

In step S7, the data processor 9 compares the average value Mve and the average value Mne, and if the average value Mve is greater than the average value Mne, proceeds to step S9, otherwise, step S11 Proceed to.

In step S9, the data processor 9 adds the first incremental value (for example, 30) to the stored first computational value Vc up to the previous frame, and then adds the first computational value Vc in the current frame. ) To the storage unit (7). The data processor 9 performs step S9 and proceeds to step S13.

In step S11, the data processor 9 subtracts the first decremented value (e.g., 5) from the stored first calculated value Vc up to the previous frame, and then the first estimated value Vc up to the current frame. ) To the storage unit (7). The data processor 9 performs step S11 and proceeds to step S13.

In step S13, the data processor 9 compares the first calculated value Vc stored up to the current frame with the first reference value Vf, and if the first calculated value Vc is greater than the first reference value Vf. If so, it proceeds to step S15, otherwise it proceeds to step S17.

In step S15, the data processor 9 sets the first result value R1 (e.g., a logical value F) as the sound signal up to the current frame has a voice and stores it in the storage unit 7 . The data processor 9 performs step S15 and proceeds to step S19.

In step S17, the data processor 9 sets the first result value R1 (e.g., a logical value T) to the sound signal up to the current frame as no voice (non-speech), and the storage unit 7 ). The data processor 9 performs step S17 and proceeds to step S19.

In step S19, the data processor 9 uses the signal in the frequency domain of the frame in step S3 to generate the energy (Esp) of the frequency band in which the feedback occurs, and the frequency band other than the frequency band where the feedback occurs. Calculate the energy (Eot) of each.

In step S21, the data processor 9 compares energy (Esp) and energy (Eot), and if the energy (Esp) is greater than the energy (Eot) proceeds to step (S23), otherwise step (S25) ).

In step S23, the data processor 9 adds the second incremental value (for example, 30) to the second computed value Fc up to the previous frame to obtain a second computed value Fc up to the current frame. Save it. The data processor 9 performs step S23 and proceeds to step S27.

In step S25, the data processor 9 subtracts a second decremented value (e.g., 1) from the second calculated value Fc up to the previous frame to obtain the second calculated value Fc up to the current frame. Save it. The data processor 9 performs step S25 and proceeds to step S27.

In step S27, the data processor 9 compares the second calculated value Fc up to the current frame with the second reference value Ff, and if the current second calculated value Fc is the second reference value Ff If it is greater than ), the process proceeds to step S29, otherwise the process proceeds to step S31.

In step S29, the data processor 9 sets the second result value R2 (for example, a logical value T) by generating feedback on the acoustic signal up to the current frame and stores it in the storage unit 7 . The data processor 9 performs step S29 and proceeds to step S33.

In step S31, the data processor 9 sets the second result value R2 (e.g., a logic value F) due to no feedback on the acoustic signal up to the current frame and stores it in the storage unit 7 do. The data processor 9 performs step S31 and proceeds to step S33.

In step S33, the data processor 9 reads the VAD performance data from the storage unit 7 to determine whether the VAD performance is equal to or greater than the reference performance. If it is more than the reference performance, the process proceeds to step S35, otherwise the process proceeds to step S37.

In step (S35), the data processor 9 is a high performance whose VAD performance is greater than or equal to the reference performance, so if it is determined as non-voice (no voice) in the VAD process or as feedback is generated in the feedback check process, it is finally determined that the feedback has been generated. In order to determine as, the first result value R1 and the second result value R2 are ORed and the result value R processed by the logical operation is stored in the storage unit 7.

In step S37, since the data processor 9 is a low-performance VAD whose performance is less than the reference performance, it is determined that it is non-voiced (no voice) in the VAD process, and if it is determined that feedback has occurred in the feedback check process, the feedback has been generated. In order to make a final decision, the first result value R1 and the second result value R2 are ANDed and the result value R processed by the logical operation is stored in the storage unit 7.

In step S39, the data processor 9 proceeds to step S41 when it is determined that the result value R is that the feedback has finally occurred (for example, a logical value T), otherwise, the data processor 9 finally returns the feedback. It is determined that this has not occurred and proceeds to step S43.

In step S41, the data processor 9 performs a mute process by setting the value of the FFT BIN of the current frame to 0 in order to remove the feedback.

In step S43, the data processor 9 performs an inverse fast Fourier transform (IFFT) on the signal in the frequency domain processed in step S41 or the signal in the frequency domain on which the processing in step S41 has not been performed. It converts to a signal in the time domain. The data processor 9 applies a sound signal, which is a signal in the time domain, to the speaker 3 to emit sound, transmits it to an external electric device through the communication unit 5, or stores it in the storage unit 7. The data processor 9 performs step S43 and proceeds to step S45.

In step S45, the data processor 9 determines whether the input of the sound signal has ended. That is, the data processor 9 terminates the feedback processing method if the input of the sound signal from the microphone 1, the sound signal from the communication unit 5, or the sound signal read from the storage unit 7 is ended, otherwise The process proceeds to step S1, and the next frame is processed.

The above-described steps (S5) to (S17) are VAD processes, and the above-described steps (S19) to (S31) correspond to the feedback confirmation process. However, the data processor 9 may perform the feedback check process prior to the VAD process, or simultaneously perform the feedback check process and the VAD process.

In the feedback check process, the data processor 9 may compare the energy ratio (Esp/Eot) and the energy threshold instead of comparing the energy (Esp) and the energy (Eot) in step S21. That is, the data processor 9 reads the energy threshold stored in the storage unit 7, and if the energy ratio (Esp/Eot) is greater than the energy threshold, it determines that feedback has occurred and proceeds to step S23, otherwise It proceeds to step S25.

In FIG. 2 described above, in order to more accurately determine whether feedback has occurred in the sound signal, the data processor 9 reflects the result values of the VAD process and the feedback check process differently according to the VAD performance (OR operation or AND Operation) is described.

As another embodiment, the data processor 9 may determine whether to perform the VAD process and the feedback check process according to the performance of the VAD. In detail, after performing steps (S1) and (S3), the data processor 9 is divided into a case where the VAD performance is greater than or equal to the reference performance and a case where the VAD performance is less than the reference performance, similar to the step S33. Is explained.

When the VAD performance is greater than or equal to the reference performance, the data processor 9 first performs either the VAD verification process or the feedback verification process, and when it is determined as the determination of non-voice or the occurrence of feedback in the first performed process, step S41 Otherwise, if it is determined that voice or feedback has not occurred in the first process, the remaining process is performed, and if it is determined that non-speech or feedback is generated, the process proceeds to step S41. If the determination of non-speech or feedback generation is determined in the process performed first through such control, the data processor 9 proceeds to step S41 without performing the remaining process, so that the amount of data computation can be reduced.

When the VAD performance is less than the reference performance, the data processor 9 may perform a VAD check process and a feedback check process, respectively, perform an AND operation in step S37, and proceed to step S39.

At least a part of a device (eg, a processor or its functions) or a method (eg, operations) according to various embodiments is, for example, in a computer-readable storage media in the form of a program module. It can be implemented with stored instructions. When the command is executed by a processor, the one or more processors may perform a function corresponding to the command. The computer-readable storage medium may be, for example, a memory.

Computer-readable recording media include hard disks, floppy disks, magnetic media (e.g. magnetic tape), optical media (e.g. CD-ROM, DVD (Digital Versatile Disc), magnetic- It may include optical media (eg floptical disk), hardware devices (eg ROM, RAM, flash memory, etc.), etc. In addition, program instructions, such as those made by a compiler, may be included. It may include not only machine code but also high-level language code that can be executed by a computer using an interpreter, etc. The above-described hardware device may be configured to operate as one or more software modules to perform operations of various embodiments. The same goes for vice versa.

A processor or functions performed by the processor according to various embodiments of the present disclosure may include at least one or more of the above-described components, some of the above-described components may be omitted, or additional other components may be further included. Operations performed by a module, a program module, or other components according to various embodiments may be executed sequentially, in parallel, repeatedly, or in a heuristic manner. Also, some operations may be executed in a different order, omitted, or other operations may be added.

As described above, the present invention is not limited to the specific preferred embodiment described above, and anyone with ordinary knowledge in the technical field to which the present invention belongs without departing from the gist of the present invention claimed in the claims It goes without saying that modifications can be made, and such changes are within the scope of the claims.

1: microphone 3: speaker
5: communication unit 7: storage unit

Claims (6)

Receiving an audio signal;
Converting the input sound signal into a frequency domain signal in a frame unit;
Performing a VAD process and a feedback check process for each frame of the sound signal converted to the frequency domain;
Performing a logical operation between the voice determination result of the VAD process and the feedback generation determination result of the feedback verification process according to the previously stored VAD performance;
Determining a result of generating a final feedback based on a result value in the step of performing a logical operation;
Including the step of performing mute processing according to the result of the final feedback generation in the determination step,
The steps to perform the VAD process are:
Calculating an average value (Mve) of energy in the voice frequency domain and an average value (Mne) of energy in the non-speech frequency domain in the current frame;
For each frame, the current first calculated value (Vc) corresponding to the number of times when the average value (Mve) is greater than the average value (Mne) is calculated, but if the average value (Mve) is greater than the average value (Mne), the previous first calculation The first increment value is added to the value Vc and stored as the first calculated value Vc in the current frame. If the average value Mve is less than or equal to the average value Mne, the previous first calculated value Vc Subtracting the first reduction value from) and storing it as a first calculated value Vc in the current frame;
And determining that there is a voice if the current first calculated value Vc is greater than the first reference value Vf, and if not, determining that there is no voice. The method of claim 1,
If the VAD performance is greater than or equal to the reference performance in the execution step of the logical operation, the voice determination result of the VAD process and the feedback occurrence determination result of the feedback check process are ORed in the determination step
When the VAD performance is less than the reference performance in the performing step of the logical operation, the voice determination result of the VAD process and the feedback generation determination result of the feedback check step are ANDed in the determination step. The method of claim 1,
The steps to perform the feedback verification process are:
Calculating energy (Esp) in a frequency band where feedback occurs in the current frame and energy (Eot) in a frequency band other than the frequency band in which feedback occurs, respectively;
For each frame, a second calculated value Fc corresponding to the number of times when the energy Esp is greater than the energy Eot is calculated,
If the energy (Esp) is greater than the energy (Eot), the second increase value is added to the previous second calculated value (Fc) and stored as a second calculated value (Fc), and the energy (Esp) is the energy (Eot). If it is less than or equal to, subtracting the second reduction value from the previous second calculated value Fc and storing it as a second calculated value Fc;
And determining that there is no feedback if the second calculated value Fc is larger than the second reference value Ff, and if not, determining that there is no feedback. The method of claim 3,
Feedback processing method, characterized in that the frequency band in which the feedback occurs is 5,750 ~ 6,500Hz. delete delete

Images